KR20010042037A - Methods for making character strings, polynucleotides and polypeptides having desired characteristics - Google Patents

Abstract

It is to provide a related integrated system using genetically working genes and libraries produced by in silico nucleic acid recombination methods and in silico shuffling methods.

Description

METHODS FOR MAKING CHARACTER STRINGS, POLYNUCLEOTIDES AND POLYPEPTIDES HAVING DESIRED CHARACTERISTICS}

[Reference to Related Applications]

This application is a USSN 60 / 116,447 filed January 19, 1999 by Selifonov and Stemmer, which is not a preliminary pseudonym application of "METHODS FOR MAKING CHARACTER STRINGS, POLYNUCLEOTIDES AND POLYPEPTIDES HAVING DESIRED CHARACTERISTICS", and also Selifonov and Stemmer USSN 60 / 118,854, filed February 15, 1999, filed October 12, 1999 by Selifonov et al., Of non-specific applicant, "METHODS FOR MAKING CHARACTER STRINGS, POLYNUCLEOTIDES AND POLYPEPTIDES HAVING DESIRED CHARACTERISTICS" US Ser. No. 09 / 416,375, filed in part, entitled "METHODS FOR MAKING CHARACTER STRINGS, POLYNUCLEOTIDES AND POLYPEPTIDES HAVING DESIRED CHARACTERISTICS."

This application is also part of USOL 09 / 408,392, filed September 28, 1999 by Crameri et al., And is a continuing application of "OLIGONUCLEOTIDE MEDIATED NUCLEIC ACID RECOMBINATION," filed February 5, 1999 by Crameri et al. US Ser. No. 60 / 118,813, entitled "OLIGONUCLEOTIDE MEDIATED NUCLEIC ACID RECOMBINATION," and USOL 60 / 141,049, filed June 24, 1999 by Crameri et al., As "OLIGONUCLEOTIDE MEDIATED NUCLEIC ACID RECOMBINATION." Partial application of "OLIGONUCLEOTIDE MEDIATED NUCLEIC ACID RECOMBINATION", Attorney Docket Number 02-296-3 US, filed simultaneously by non-specific applicant, Crameri et al.

This application is also part of USMET 09 / 416,837, filed Oct. 12, 1999 by Selifonov and Stemmer, a partial continuing application of "METHODS OF POPULATING DATA STRUCTURES FOR USE IN EVOLUTIONARY SIMULATIONS" and concurrently filed by Selifonov and Stemmer. Attorney Docket Number 3271.002W00, filed by Majestic, Parsons, Siebert & Hsue, is a partial continuing application of "METHODS OF POPULATING DATA STRUCTURES FOR USE IN EVOLUTIONARY SIMULATIONS."

This application is also related to USSN 09 / 408,393, "USE OF CODON VARIED OLIGONUCLEOTIDE SYNTHESIS FOR SYNTHETIC SHUFFLING," filed September 28, 1999 by Welch et al.

This application claims 35 U.S.C. § 119 (e) and / or 35 U.S.C. In accordance with § 120, each application listed in this chapter, in legal form, shall be given priority and privilege. All such applications are hereby incorporated by reference.

[Copyright notice]

37 C.F.R. In accordance with 1.71 (e), Applicants understand that this publication contains material protected by copyright. The copyright owner has no objection to the facsimile copying by a third party of the patent document or patent publication as indicated in the Patent and Trademark Office patent file or record, but otherwise reserves the copyright for all rights.

Repetitive nucleic acid recombination (“shuffling”) rapidly evolves nucleic acids in vivo or in vitro. This rapid evolution produces encoded molecules (eg, nucleic acids and proteins) with new and / or improved properties. Proteins and nucleic acids, which play an important role in industrial, agricultural, and therapeutic purposes, can be made or improved through DNA shuffling processes.

Numerous publications published by the inventors and collaborators of the present invention describe DNA shuffling. See, eg, Stemmer et al. (1994) “Rapid Evolution of a Protein” Nature 370: 389-391; By Stemmer (1994) "DNA Shuffling by Random Fragmentation and Reassembly: in vitro Recombination for Molecular Evolution", Proc. Natl. Acad. USA 91: 10747-10751; Stemmer, US Patent No. 5,603,793 "METHODS FOR IN VITRO RECOMBINATION"; Stemmer et al., US Pat. No. 5,830,721 "DNA MUTAGENESIS BY RANDOM FRAGMENTATION AND REASSEMBLY"; And Stemmer et al., US Pat. No. 5,811,238, " METHODS FOR GENERATING POLYNUCLEOTIDES HAVING DESIRED CHARACTERISTICS BY INTERATIVE SELECTION AND RECOMBINATION, " for example, describe various shuffling techniques.

Many methods of application of DNA shuffling techniques have been developed by the inventors and their collaborators. In addition to the above documents, US Pat. No. 5,837,458 by Minshull et al., "METHODS AND COMPOSITIONS FOR CELLULAR AND METABOLIC ENGINEERING," provides a method for the evolution of new metabolic pathways and the improvement of biological processing through iterative shuffling techniques. Crameri et al. (1996), “Construction And Evolution Of Antibody-Phage Libraries BY DNA Shuffling” Nature Medicine 2 (1): 100-103, for example, describe antibody shuffling against antibody phage libraries. Additional descriptions regarding DNA Shuffling are described, for example, in WO95 / 22625, WO97 / 20078, WO96 / 33207, WO97 / 33957, WO98 / 27230, WO97 / 35966, WO98 / 31837, WO98 / 13487, WO98 / 13485 and WO98 / 42832 It can also be found in many published applications, such as.

Many of the literature of the present inventors and their co-workers and other researchers in the art also promote DNA shuffling by providing methods for reassembling oligonucleotides encoding, for example, genes or gene fragments obtained from small fragments of genes. The technique of letting is described. In addition to the above documents, US Pat. No. 5,834,252 to "END COMPLEMENTARY POLYMERASE REACTION" by Stemmer et al. (1998) describes amplification and detection of target sequences (eg, in a mixture of nucleic acids) as well as from fragments. A process for assembling large polynucleotides is described.

Looking at the above documents, it can be seen that DNA shuffling is an important new technology that is actually used a lot. Thus, new technologies that promote DNA shuffling are highly desirable. In particular, techniques for reducing the number of physical operations required for the shuffling process are particularly useful. The present invention provides novel techniques and other aspects of DNA shuffling that will become apparent upon a complete review of the specification.

[Overview of invention]

The present invention provides a novel “in silico” DNA shuffling technique in which part or all of the DNA shuffling process is performed or modeled in a computer system, without the need for physical manipulation (some or all) of the nucleic acids. do. This approach is collectively called "Genetic Algorithm Guided G ene Synthesis "or" GAGGS ".

It is a first aspect of the present invention to provide a method for obtaining a "chimeric" or "recombinant" polynucleotide or polypeptide (or other biopolymer) having the desired characteristics. In this method, two or more parent character strings are provided that encode sequence information for one or more polypeptides and / or one or more single- or double-stranded polynucleotides. All or some of the sequences (ie, one or more subsequences) comprise an area of identity and an area of heterology. A set of strings of predetermined or selected lengths that encode single-stranded oligonucleotide sequences is provided that includes at least a portion of each source string and / or overlapping sequence fragments of a portion of a polynucleotide chain complementary to the source string.

In a group of embodiments, the invention provides a method of making a biological polymer library. The method involves generating various groups of strings in a computer, where the strings are generated through alterations (recombination, mutagenesis, etc.) of existing strings. Strings of various groups are then synthesized to form a library of biological polymers (nucleic acids, polypeptides, peptide nucleic acids, etc.). Typically, members of a biological polymer library are selected for one or more activities. In a second aspect of the invention, additional sets of additional libraries or strings having members of a biological polymer library exhibiting activity below the intended threshold are filtered by subtracting them. In a third aspect of the invention, additional libraries or additional sets of strings having members of a biological polymer library exhibiting activity above the intended threshold are filtered by biasing them.

A set of single-stranded oligonucleotides prepared according to the set of sequences defined in the string is provided. Some or all of the single-stranded nucleotides prepared are pooled under denaturing or annealing conditions, where two or more single-stranded oligonucleotides represent a portion of two different source sequences. The resulting single stranded oligonucleotide population is incubated with a polymerase that anneals the single stranded fragment of the identity zone to form the annealed fragment. This identity zone is sufficient for either of the paired chains to prime the replication of the other to increase the length of the oligonucleotide. The resulting single and double stranded oligonucleotide mixtures are denatured into single stranded fragments. The above steps are performed repeatedly so that a portion of the mixture of at least single-stranded chimeric and mutated polynucleotides is used in the continuous circulation step. Recombinant polynucleotides that have been evolved to have the desired properties are selected or screened.

In a fourth aspect of the invention, the invention provides for the use of a gene effector (eg, in a computer). In such a method, the sequence column corresponding to the oligonucleotide is selected by a computer from a sequence column corresponding to one or more of the following sets of single-stranded oligonucleotides.

a) oligonucleotides synthesized to include certain mutations, randomly or non-randomly, in the source sequence, according to a variant sequence, including replacing one or more letters with another letter, or deleting or inserting one or more letters ;

b) oligonucleotide sequences synthesized to include degenerate, mixed or non-natural nucleotides at random positions, randomly or non-randomly; And

c) A chimeric oligonucleotide synthesized according to the artificial sequence of the character subcolumn designed to include some joined sequence of two or more source sequences.

In any embodiment, the set of (c) oligonucleotides comprises one or more mutagenic or degenerate positions as defined in (a) and (b). (c) the set of oligonucleotides is a crossover point selected according to a method that enables the identification of multiple character substrings showing pair identity (homology) between all or a portion of a column pair comprising sequences having different source strings. Is any chimeric nucleotide having

The crossover points for preparing chimeric oligonucleotide sequences are randomly or nonrandomly selected in some or intermediate regions of the identified pair identity regions (homology) or by other sets of selection samples.

In a fifth aspect of the invention, one or more crossover points for one or more chimeric oligonucleotide sequences are selected outside the detected identity region.

In a sixth aspect of the invention, a mixture of the foregoing single-stranded oligonucleotides comprises a region of part and / or all of any source sequence provided, and / or a region of heterogeneity and identity with any region of the provided source strings. Pooled one or more times with an additional set of polynucleotides comprising one or more double- or single-stranded polynucleotides encoded by other strings.

Polynucleotides obtained from an additional set of polynucleotides can be used to synthesize oligonucleotides corresponding to any source string (or homologs thereof) or to random fragmentation (e.g., DNases of polynucleotides). Cleavage, or chemical cleavage) and / or restriction enzyme fragmentation of polynucleotides encoded by heterologous and any other strings comprising the same region as described above and / or any region of the source string provided. That is, any nucleic acid synthesized by GAGGS can be further modified by practical methods to obtain additionally modified nucleic acids. Moreover, any mutated nucleic acid can be used as a substrate for further circulation of GAGGS.

The method comprises synthetic oligonucleotides of a wide range of lengths (e.g., 10-20 nucleotides or more, 20-40 nucleotides or more, 40-60 nucleotides or more, 60-100 nucleotides or more, 100-150 nucleotides Or more), various forms of source sequences (e.g., therapeutic proteins such as EPO, insulin, growth hormones, antibodies, etc .; plant hormones, disease resistance factors, herbicide resistance factors (e.g. p450s) Agricultural proteins; industrial proteins (e.g., proteins involved in bacterial desulfurization of oils, polymer synthesis, detoxification and complexation of proteins, fermentation, etc.) and multiple selection / screening cycles (e.g., one or more cycles, two or more) Cycles, 3-4 cycles, 10 cycles, 10-50 cycles, 50-100 cycles, or 100 cycles. Rounds may be altered by physical nucleic acid shuffling and / or selection assays in a variety of formats (in vivo or in vitro): Selected nucleic acids (ie, nucleic acids with intended properties) may be sequenced or restriction enzyme assays, real-time PCR assays, etc. Other processes such as deconvolution, for example, may not physically manipulate DNA obtained from previous GAGGS rounds, so that the process can be started using sequence information to induce gene synthesis. .

In the above method, polynucleotides obtained from single-stranded oligonucleotides are usually synthesized by assembly PCR. Other options at the time of nucleic acid preparation include ligation reactions, cloning and the like.

In a typical embodiment, a set of strings encoding a single stranded oligonucleotide comprising a source column fragment and comprising chimeric and mutant / degenerate fragments of predetermined length are processed such as a computer with software for sequence column manipulation. Synthesized using a device containing arguments.

In a seventh aspect of the invention, the invention provides a single source GAGGS. The methods are described in more detail in the Examples herein.

The present invention relates to the application of genetic algorithms to genetic algorithms and nucleic acid shuffling methods.

1 is a flow chart illustrating some of the evolution induced by GAGGS.

2 is a flow chart illustrating some of the evolution induced by GAGGS. The flowchart of FIG. 2 may be arbitrarily connected with FIG.

3 is a flow chart illustrating part of the evolution induced by GAGGS. The flowchart of FIG. 3 can be arbitrarily connected with FIG.

4 is a flow chart illustrating some of the evolution induced by GAGGS. The flowchart of FIG. 4 may be arbitrarily connected with FIG. 3.

5 is a relational tree showing the percent similarity for different subtilisin (standard shuffling target).

FIG. 6 is a dot-plot arrangement showing homology zones for different subtilisin.

FIG. 7 is a dot plot arrangement showing homology to seven different source subtilisins.

In Figure 8, Panels A-C are parallel histograms showing conditions that determine the possibility of crossover selection capable of independently controlling any region and source chain pair of selected gene length.

9 is a chart illustrating the introduction of retrieved crossover markers into each source sequence.

10 shows a process for assembling oligonucleotides to prepare nucleic acids.

11 is a continuation of FIG. 10 illustrating an oligonucleotide assembly process.

Figure 12 is a difference plot and correlation tree for naphthalene deoxygenase shuffling.

Figure 13 is a schematic diagram of the digital system of the present invention.

14 is a schematic diagram showing geometric relationships between nucleotides.

15 is a schematic diagram of an HMM matrix.

In the present invention, "genetic" or "evolutionary" algorithms are used to prepare sequence sequences that can be converted into physical molecules that have been shuffled and tested to have the desired properties. This strongly facilitates the evolutionary process, as the ability to preselect shuffling substrates in the shuffling method reduces substantial physical manipulation of the nucleic acid. Furthermore, the use of strings as "virtual substrates" in shuffling methods reduces the need to obtain physical source molecules encoding genes when performed in conjunction with gene reconstruction methods.

Genetic Algorithms (GAs) are used in various fields to solve the problem that they are not fully characterized or are too complex to be fully characterized and only analytical evaluation is possible. That is, GA is used to solve problems that can be assessed by quantifiable measures of the relative value of a solution (ie, the minimum value of the relative value of a comparison of one potential solution with another solution). The basic concept of genetic algorithms is to encode potential solutions to problems as a set of parameters. A single set of parameter values is treated as the "genomic" of each solution, ie the genetic material solution. The result is a large cluster of candidate solutions. The solution means that the principle of survival of the fittest, ie the likelihood that each solution will repeat several parameter values for the final solution set, is directly related to the optimal state of the individual (ie the solution within the cluster for the selected parameter). How to maintain a good relationship with this other entity).

Breeding is accomplished by using operators such as crossovers and mutations similar to basic biological recombination. Simple use of the operator in combination with the theoretical selection mechanism can produce noticeable results for a wide range of problems.

Introduction to genetic algorithms is described by David E. Goldberg (1989) Genetic Algorithms in Search, Optimization and Machine Learning, Addison-Wesley Pub Co; IBSN: 0201157675 and Timothy Masters (1993) Practical Neural Network Recipes in C ++ (Book & Disk edition) Academic Pr ; ISBN: 0124790402. Many more recent references discuss the use of genetic algorithms for solving many difficult problems. See for example http://garage.cse.msu.edu/papers/papers-index.html and references cited therein; http://gaslab.cs.unr.edu and references cited therein; http://www.aic.nrl.navy.mil/ and references cited therein; http://www.cs.gmu.edu/research/gag/ and references cited here and http://www.cs.gmu.edu/research/gag/pubs.html and references cited here See references.

In the present invention, a genetic algorithm (GA) is used to provide string-based samples for processes that result in mutations in biopolymers (provided clusters of strings such as gene sequences, for example one or more gene effectors). (E.g., evolving strings on a computer by applying them to a source library).

Samples of GA-generating string clusters (or "derivative libraries") may be used for polynucleotide synthesis (eg, non-error prone synthesis, error-prone synthesis, parallelism). It is used as a sequence guide set in a form suitable for controlling parallel synthesis, pooled synthesis, chemical synthesis, chemical enzymatic synthesis (including synthetic PCR of synthetic oligonucleotides). Polynucleotide synthesis is performed with sequences encoded by strings in the derived library. This provides a physical representation (polynucleotide library) of the "gene" (or other string) diversity generated by the operation.

Physical selection of the polynucleotides having the intended characteristics is also optionally (and usually) performed. This selection is based on the results of physical assays on the properties of the polynucleotides or polypeptides, ie whether they are translated in vitro or expressed in vivo.

Polynucleotide sequences known to have the desired characteristics are simplified (eg, by recording the location of the polynucleotide if sequencing or position information is available). This is done by DNA sequencing, localization, real-time PCR (eg, TaqMan), restriction enzyme cleavage, or other methods mentioned herein or currently used.

The steps can be optionally repeated, for example for one to four or more cycles, wherein simplified sequences are optionally used as a source of information for synthesizing a new set of modified strings used to begin each process each time. . Of course, any nucleic acid sequence synthesized in silico can be synthesized and shuffled by any known DNA shuffling method, including methods taught in the references by the inventors and co-workers mentioned herein. DNA thus synthesized may also be mutated or otherwise modified according to existing techniques.

In summary, GAGGS applies information manipulation steps (genetic algorithms to strings representing biopolymers such as nucleic acids or proteins) that create a set of specific informational factors (eg strings) that are physically used as templates for synthesizing nucleic acids. Evolutionary process). The information factor may be inserted into the database or otherwise manipulated in silico, for example, by repeatedly applying GA to the synthesized sequence. Corresponding physical nucleic acids may be recombined / selected into simplified (eg, sequenced or analyzed) nucleic acids or may cause other diversity and repeat the process to obtain the desired nucleic acid.

Typical Benefits of GAGGS

Compared with the prior art, GAGGS has a number of advantages. Since sequence information is used in oligonucleotide design and selection, one of them is that there is no need for physical access to a gene / organism, for example. Many common databases provide additional sequence information, including, for example, Genbank ^™ and the aforementioned information. Additional sequence databases are commercially available based on contracts from a number of companies that specialize in the generation and storage of genomic information.

Similarly, sequences from organisms that cannot be obtained and cultured can be used for GAGGS. For example, sequences obtained from pathogenic organisms can be used without actually dealing with the pathogen. All forms of sequence suitable for physical DNA shuffling, including damaged and incomplete genes (eg, pseudo genes), are easy to apply to GAGGS.

All gene operators, including different forms of mutagenesis and crossover, that can prevent human errors and mutations from occurring in physical experiments involving DNA manipulation, are completely and independently controllable in a reproducible form. . GAGGS can impart self-learning capabilities by artificial intelligence (optimization of algorithm output parameter behavior based on feedback on yield, success rate and failure rate of physical screening).

In the GAGGS process, sequences in which frame-shift mutations (generally undesirable) occur are removed or modified (removed from a feature set or repaired in silico). Similarly, portions that are prematurely terminated and portions of which sequence features are known to be important for exhibiting the intended properties (eg, ligands conserved in metal bonds) are removed or repaired.

Moreover, wild type parents do not contaminate derived libraries with multiple overlapping parent molecules, and in one preferred embodiment, only the already modified genes are physically shuffled and / or screened (in some cases according to the available assays) May be expensive or low throughput or otherwise not ideal).

Moreover, since no physical physical recombination method is required, protein sequences can be shuffled in the same manner as nucleic acid sequences in silico and reverse translation of the resulting shuffled sequences alleviates problems with codon usage and It can be used to minimize the number of oligonucleotides needed to construct one or more libraries. In this regard, protein sequences are shuffled in silico using gene operators based on structural domain and folding motif recognition rather than bound by annealing based homologous representations of DNA sequences or simple homology of AA sequences. Can be ring. In addition, biases based on theoretical structures are readily coupled to library constructs when such information is useful.

The significant operating cost of performing GAGGS is the cost of synthesizing a huge library of genes represented in silico. Synthetic assembly of genes can be performed, for example, by assembly PCR using 40-60bp oligonucleotides, which can be synthesized inexpensively with current techniques.

GAGGS Induced Evolution

All changes that occur to any DNA sequence during any evolutionary process can be described as a limited number of events, i.e. events resulting from the action of minimal gene effectors. In the subspace of any given source sequence, such changes can be accurately described and simulated in the physical representation of the evolutionary process intended to bring diversity to the sequence when physically screened for the desired properties. Physical double-stranded polynucleotides are not needed to initiate the GAGGS process: instead, they are synthesized for the purpose of physical screening and / or selection and / or by performing an initial GAGGS process as a result of the screening or selection. There is no need to produce a very large library used for screening / selection.

Genetic algorithm (GA)

String: In general, a string can be any representation of an array of characters (eg, a linear array of characters provides a “command” when a non-linear array is used as a cipher that produces a linear array of features). In performing GAGGS, the string is preferably any encoded sequence that can be explicitly converted into a string representing a monomer or multimeric sequence (whether composed of natural monomers or artificial monomers) in a polynucleotide, polypeptide, etc., Encoding a polynucleotide or polypeptide string directly or indirectly, including the placement of an image or object.

Genetic Algorithms: Genetic algorithms generally simulate the process of evolution. Genetic Algorithms (GAs) are used in various fields to solve the problem that they are not fully characterized or are too complex to be fully characterized and only analytical evaluation is possible. That is, GA is used to solve problems that can be assessed by quantifiable measures of the relative value of a solution (ie, the minimum value of the relative value of a comparison of one potential solution with another solution). Within the detailed description of the present invention, genetic algorithms are the process of selecting or manipulating strings in a computer, which typically can correspond to one or more biopolymers (eg, nucleic acids, proteins, PNAs, etc.). A biological polymer is any polymer that shares some structural properties with naturally occurring polymers such as RNA, DNA, and polypeptides, including RNA, RNA analogs, DNA, DNA analogs, polypeptides, polypeptide analogs, peptide nucleic acids, and the like.

Induced Evolution of a String or Object:

The process of artificially altering strings by artificial selection is of various genetic diversity (2), with several beings (1), some of which are determined as a result of selection for fitness (expected characteristics). Reproductive success) occurs in different reproductive communities. The reproductive community may be, for example, a physical community within a computer or a virtual community.

Gene Operators (GOs): A user-defined operation or set of operations, each of which contains a set of logical instructions for string manipulation. The gene effector is adapted to cause a change in the population to launch the purpose (useful) zone of the search space (population with the intended intent characteristics) as a predetermined means of selection. Predetermined (or partially scheduled) means of selection include computational tools (operators that include logic steps derived by analyzing information describing a library of strings), where the physical tool for physical characterization of a physical object is a string library It can be windowed from data to physically represent the information describing it. In a preferred embodiment, some or all of the logical operations are performed in a computer.

Gene effector

Any change that occurs in any cluster with any form of string (i.e. any physical property of the physical object encoded by such a column) randomizes a limited set of logical mathematical functions, including various types of gene operators. And / or the result of the intended application.

In mathematical terms, the instructions are not abstract axioms by assumption. In fact, the command is a proof derived from a strict formal proof that is easily obtained from Wiles' proof of Fermat's last theorem. The fundamental application of Wiles' Proof in evolutionary molecular biology is a process of proof of speculation that states that all elliptic curves are fundamentally modular. Specifically, all the diversity and evolution of living matter in the universe (ie, multiple objects with properties that can be represented by a limited number of elliptic curves) can be described as five basic arithmetic operating languages: addition, Deduction, manipulation, division, and modular forms (ie, the evolution of life can be effectively described by restrictively changing a limited cluster of strings, eg, a simple combination of information in all DNA in the universe). Not only can we determine the language of a form based on the nucleic acid of an organism, but we can also define all the basic forms of gene effectors that apply to nucleic acids by evolutionary selection.

Mathematical modeling for any gene effector is described, for example, in Sun (1999) "Modeling DNA Shuffling" Journal of Computational Biology 6 (1): 77-90; Kelly et al. (1994) “A test of the Markovian model of DNA evolution” Biometrics 50 (3): 653-64; Boehnke et al. (1991) "Statististical method for multipoint radiation hybrid mapping" Am. J. Hum. Genet. 49: 1174-1188; Irvine et al. (1991) "SELEXION: systematic evolution of ligands by exponential enrichment with integrated optimization by non-linear analysis" J. Mol. Biol. 222: 739-761; Lander and Waterman (1988) Genomic mapping by fingerprinting random clones: a mathematical analysis "Genomics 2: 231-239; Large (1997) Mathematical and Statistical Methods for Genetic Analysis Springer Verlag, NY; Sun and Waterman (1996)" A mathmatical analysis of in vitro molecular section-amplification "J. Mol. Biol. 258: 650-660; Waterman (1995) Introduction to Computational Biology Chapman and Hall, London, UK.

The following describes any basic genetic operation applicable to the present invention.

Proliferation (including redundancy and duplication) is a form of string reproduction, which produces an additional copy of the string that includes the origin cluster / library of the stream. Proliferation behavior has a number of variables. They can be applied to each row or to groups of identical or unequal columns. Selecting ten groups for propagation can be random or biased.

Mutations: All mutation forms in each member of a set of rows can be described by several simple features that can be reduced to elements that include replacing one set of features with another set of features. One or more features may be mutated in a single action. If one or more features are mutated, the feature set may or may not be continuous over the entire row length (a feature useful for simulating intimate clustering by any chemical mutant). Single point mutation An operator can substitute a single feature for another single feature. The characteristics of the novel features may vary and they may be obtained from the same set of features that make up the source row, or from different sets of features (e.g., representing degenerate nucleic acid bases, unnatural nucleic acid bases or amino acids, etc.). have. Deletion mutations are more complex operators that remove one or more features from the heat. Each single conjugation in a nucleic acid stabilizing column may be undesirable for manipulating a row representing a polynucleotide sequence; 3x clustered (continuous or distributed) deletions ("ternary deletion frame shift") may be acceptable. However, single point deletions are acceptable and useful in the evolutionary operations of the sequences encoding polypeptides. Insertion mutations are similar to deletions except that one or more new features are inserted. The added features are optionally varied and they can be obtained from the same set of features constituting the source row, or from different sets of features (e.g., representing degenerate nucleic acid bases, unnatural nucleic acid bases or amino acids, etc.). have. Death can simply be defined as a mutation of a deletion operator. This occurs when, as a result of the application of a gene operator (or a combination thereof), deletions occur in each string or in the entire string (sub) population. Death can also be defined as a variation of the operator of elitism tendency propagation (proliferation of values that limit the level of redundancy of one or more rows to zero). Death also transfers operations by indexing and classifying the behavior of a column's indexed library without performing non-selective actions within the operator that selects a subgroup of columns (all untransferred columns are considered dead or nonexistent in subsequent operations). Can be).

FRAGMENTATION OF STRINGS is an important class of any non-elementary (complex) operator that may have the advantage of stimulating the evolution of heat in various ways of DNA shuffling. In operation, fragmentation may be described as a formal variation of a combination of deletion or propagation effectors. However, those skilled in the art will appreciate that there are many other simple algorithmic operations that can fragment any given string to produce shorter columns of descendants. Fragmentation operations can be random or biased. The size of different ranges of fragments can be predetermined. Thermal fragmentation may remain in the same community with the source row, or may be transferred to different populations. Thermal fragments from various thermal clusters may be collected to form new clusters.

Crossover (Recombination): The operator formally combines a continuous portion of one row with a continuous portion of another row in such a way that one or two hybrid rows (chimeras) are formed, wherein each of the chimeras is two different It comprises two or more contiguous row sections joined comprising at least part of the sequence of the recombinant row. The zones / points from which sequence features are obtained from different source columns are called crossovers / recombination zones / points. Crossover operation may be associated with mutational operations that affect one or more features of the recombination rows adjacent to connecting crossover regions / points. When applied repeatedly to clusters of strings, complex chimeras can be formed that are linked to some sequence of two or more source rows.

LIGATION is essentially a mutation of an insertion mutant operator in which the entire portion of one row is combined in such a way that the entire portion of the other row and the last feature of one row follow the first feature of the other row. The ligation operation can be combined with a mutation operation that affects one or more features of the ligation columns adjacent to the linking point. The ligation can also be regarded as a means of chimera formation.

ELITISM is a concept that provides a useful form of bias, including distinguishing the standard used as any gene operator, and various forms of positive and negative bias to which assignment and function can be assigned. . The proper designation of an elite operator is based on the concept of conformity. Suitability has been obtained from physical selection for the physical representation of strings and / or recognizing various sequence specific features (GC content, frame shift, termination, sequence length, specific subcolumns, homology, ligand binding, folding motifs, etc.) Determined interrelation parameters (enzyme activity, stability, ligand binding, etc.) are determined using thermal analysis tools. It is understood that different elitism expressions can be individually applied to the gene actuators, or combinations of actuators described above. In addition, if the input / output parameters of each similar operator are controllable independently (or interdependently), it may be possible to use the ellipsis with several operators of the same type in the same evolutionary operation. Different elitism representations can be used to control the changes caused by the action of each operator in a group of thermal features.

Sequence homology or sequence similarity is particularly important for sequence specific elitism useful for controlling changes in string community caused by crossover / recombination in genetic algorithms used to evolve strings encoding polynucleotide and polypeptide sequences. Form.

Various approaches, ie methods and algorithms known in the art, can be used to detect homology or similarity between different strings. Optimal alignment states of the comparative sequences are described, for example, in Smith & Waterman, Adv. App. Math. Zone homology algorithm of 2: 482 (1981), Needleman & Wunsch, J. Mol. Biol. 48: 443 (1970) homology alignment algorithm, Pearson & Lipman, Proc. Nat'l. Acad. Sci. By searching the similarity method of USA 85: 2444 (1988), by the algorithm's computed method (Wisconsin Genetics Software Package, Genetics Computer Group, 575 Science Dr., Madison, WI's GAP, BESTFIT, FASTA and TFASTA) or This can also be done through visual observation (usually Ausubel et al., Hereinafter).

One standard algorithm suitable for determining sequence identity and percent sequence similarity is the BLAST algorithm, which is described by Altschul et al., J. Mol. Biol. 215: 403-410 (1990). Software for performing BLAST assays is commercially available through the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/). The algorithm involves first identifying a High scoring Sequenc Pair (HSP) by identifying a short length instruction named W in the query sequence, where some amount of the sequence, if arranged as an instruction of equal length in the database sequence, Satisfies or matches the threshold score T of the value. T is called as the adjacent instruction score threshold (Altschul et al., Homologous). These initial contiguous instruction hits serve as seeds for initiating a search for longer length HSPs containing them. The instruction hit is then expanded in both directions along each sequence by which the cumulative alignment score can be increased. The cumulative score is calculated using the parameters M (response score for matched residue pairs; always greater than 0) and N (penalty score for mismatched residues; always less than 0). For amino acid sequences, score matrices are used to calculate cumulative scores. The cumulative alignment score decreases from its maximum value to an X value; When the cumulative score decreases to zero or less by accumulating one or more negative score residue alignments; Or when reaching the end of either sequence, the instruction hit no longer extends in each direction. The BLAST algorithm parameters, W, T and X, determine the sensitivity and speed of the alignment. The BLASTN program (for nucleotides) uses the instruction length (W) 11, expected value (E) 10, cutoff 100, M = 5, N = -4 and comparison values of both chains as default values. For amino acid sequences, the BLASTP program uses instruction length (W) 3, expected value (E) 10, and BLOSUM62 score matrix as default values (Henikoff & Henikoff (1989) Proc. Natl. Acad. Sci. USA 89: 10915).

In addition to calculating the percent sequence identity, the BLAST algorithm also statistically analyzes the similarities between the two chains (see Karlin & Altschul (1993) Proc. Nat'l. Acad. Sci. USA 90: 5873-5787). ). One measure of similarity provided by the BLAST algorithm is the smallest sum of probabilities (P (N)), which indicates the probability of a match between two nucleotide or amino acid sequences. For example, a nucleic acid is considered to be similar (homologous) to a reference sequence if the minimum total sum of probabilities when the test nucleic acid is compared to the reference nucleic acid is less than about 0.1 or less than 0.01, even less than about 0.001.

A further example of a useful sequence alignment algorithm is PILEUP. PILEUP generates multiple sequence alignments from groups of related sequences using progressive or parallel alignments. It is also possible to plot a tree representing the clustering relationship used to create the sort. PILEUP is Feng & Doolittle, J.Mol.Evol. 35: 351-360 (1987) using the progressive alignment method. The method used is similar to the method described in Higgins & Sharp, CABIOS 5: 151-153 (1989). The program can, for example, align up to 300 sequences of up to 5,000 characters in length. The overlap alignment process begins by aligning the two most similar sequences in parallel, producing an aligned two sequence cluster. The cluster can then be aligned to the next most relevant sequence or aligned sequence cluster. The two sequence clusters can be aligned by simply expanding the parallel alignment of the two respective sequences. Final alignment is performed by a series of progressive, parallel alignments. The program can also be used to plot a dendogram or tree representing a clustering relationship. The program proceeds by designating specific sequences for the comparative sequence regions and cooperation of their amino acids or nucleotides.

Therefore, different forms of similarity with varying levels of identity and length can be detected and recognized. For example, many homology determination methods have been designed for the analysis of biopolymer sequences, spell checking during command processing, and retrieval of data from various databases. By understanding the double-stranded parallel complementarity response between four major nucleic acid bases in natural polynucleotides, a model that simulates the annealing of complementary homologous polynucleotide sequences can also be used as the basis of sequence specific elitism useful for controlling crossover effectors. have.

Thus, the homology-based elitism of the crossover operator can be used to find a suitable thermal recombination pair in a column cluster, or to find / sect a region / point that is particularly suitable / intentioned for recombination of the length of a string selected for recombination. It may also be used.

Setting the severity of a predetermined form and similarity / homology as the condition under which crossover occurs is a form of elixism that controls chimeric formation between varying degrees of representative homology strings.

Repetitive use of gene effectors for string evolution. All gene operators described can be applied in repetitive rooming, and the specific parameters for each application incidence can remain the same or can be systematically or randomly varied.

Randomness in gene effectors for string evolution. Each gene effector may be applied to randomly selected rows and / or randomly selected locations at randomly selected randomly selected incidences within a range over one or more column lengths.

Array of GOS in GAS. Applying order crystallization at each GO to the derivative library product of a string may depend on the configuration of a particular set for each GO selected to perform GAGGS in different and different ways. The order may be linear, circular, parallel or a combination of the three forms and may typically be represented graphically. Multiple GO sequences can be used to simulate natural sexual and mutant processes that generate genetic diversity, or artificial processes such as DNA shuffling from a single source or family. However, the purpose of GA is not limited to simulating some known physical DNA manipulation methods. GA's ultimate goal is to provide a formal and intelligent tool based on an understanding of natural and human evolutionary processes that can provide efficient benefits to the methods currently implemented and to create and optimize the evolutionary processes in practice. To provide.

Gene synthesis

The physical synthesis of genes encoded by derivative libraries of strings obtained by operating genetic algorithms creates the physical representations of substances that conform to physical assays for their intended properties and produces substrates that are further involved in the process of generating physical diversity. Means. Thus, a first aspect of the invention relates to a method of synthesizing a gene with a selected sequence by undergoing one or more computer shuffling procedures presented herein.

For GAGGS, a technique that is effective in time and resources, gene synthesis techniques are commonly used to construct gene libraries in close fashion and in close proximity to sequence expressions produced by GA manipulation. Typically GAGGS is 10 ⁴ -10 ⁹ uses a gene synthesis method in "gene" to quickly configure the mutant library. This is typically suitable for the screening / selection method, because larger libraries are more difficult to manufacture and maintain and often cannot be fully sampled by physical assays or selection methods. For example, existing physical assays in the art (including, for example, "survival and death" selection methods) generally sample about 10 ⁹ or less variants by specific screening of specific libraries and multiple assays. Is effective for sampling 10 ⁴ -105 members. Therefore, it is a good idea to make some small libraries, because it is not easy to fully screen large libraries. However, larger libraries can also be prepared using, for example, high throughput screening methods.

Gene synthesis technology

There are a number of methods that can be used to synthesize genes with well defined sequences. For purposes of clarity only, this chapter focuses on one of a number of possible and commercially available forms of known genes and trinucleotide synthesis methods.

Current technology in polynucleotide synthesis is effective for oligomer production and best expressed as well known and mature phosphoramidite compounds. Although slightly impractical, the compounds can be commonly used to synthesize oligomers that are much longer than 100 bp with longer synthetic oligomers that are generally purified prior to use, because for longer oligomers the quality of the sequence can be degraded. Because. Typically oligomers of the "40-80 bp size can be obtained in conventional and direct methods with very high purity and without substantially degrading the quality of the sequence.

For example, oligonucleotides for use in in vitro amplification / gene reconstruction methods and for use as gene probes or shuffling targets (e.g., synthetic genes or gene segments) are commonly used by Needham VanDevanter et al. (1984) Nucleic Acids Res. , Chemical according to the solid phase phosphoramidite ester method described in Beaucage and Caruthers (1981) Tetrahedron Letts., 22 (20): 1859-1862, using an automated synthesizer as described in 12: 6159-6168. Are synthesized. Oligonucleotides may also be customized and ordered from a number of commercial sources known to those skilled in the art. There are many commercial suppliers of oligomeric synthesis services, which are widely applicable mechanisms. Any nucleic acid can be found in The Midland Certified Reagent Company (mrc@oligos.com), The Great American Gene Company (http://www.genco.com), ExpressGen Inc. (www.expressgen.com), Operon Technologies Inc. ( Custom orders from a number of commercial sources such as Alameda, CA). Similarly, peptides and antibodies may be found in many forms, such as PeptidoGenic (pkim@conet.com), HTI Bio-products, Inc. (http://www.htibio.com), BMA Biomedicals Ltd (UK), Biosynthesis, Inc. You can also make custom orders from sources.

Synthesis of whole genes from small fragments that easily conform to optimization is shown in Dillon and Rosen (Biotechniques, 1990, 9 (3) 298-300). A simple and rapid PCR based assembly method can be performed that does not use ligase of genes obtained from partially overlapped single stranded oligonucleotide sets. Several groups have described how to successfully mutate the same PCR based gene assembly assays for various gene synthesiss of increased size to indicate their general applicability and their combinatorial properties to mutated gene library synthesis. Useful references include Sandhu et al. (Biotechniques, 1992, 12 (1) 15-16), (220bp gene obtained from three oligomers of 77-86bp); Prodromou and Pearl (Protein Engineering, 1992, 5 (8) 827-829 (522bp gene obtained from 10 oligomers of 54-86bp); Chen et al., 1994 (JACS, 1194 (11): 8799-8800) (779bp Genes), Hayashi et al., 1994 (Biotechniques, 1994, 17: 310-314) and the like.

More recently, Stemmer et al. (Gene, 1995, 164: 49-53) have shown, for example, that PCR based assembly methods are effectively useful for producing larger genes greater than 2.7 kb from dozens or hundreds of synthetic 40 bp oligomers. Indicates. The authors also note that among the four basal buckling steps, including PCR based gene synthesis methods (oligonucleotide synthesis, gene assembly, gene amplification and optionally cloning), the gene amplification step if 'circulating' assembly PCR is used States that it can be omitted.

Many publications by inventors and collaborators, and other researchers in the art, present techniques for facilitating DNA shuffling, for example by reassembling genes from small fragments or oligonucleotides. A second aspect of the invention relates to the ability to use over-shuffling oligonucleotides and crossover oligonucleotides as recombinant template intermediates in various DNA shuffling methods.

Indeed, many publications by inventors and collaborators, as well as other researchers in the art, present techniques for facilitating the reassembly of genes from small fragments containing oligonucleotides. In addition to the publications described above, Stemmer et al. (1998) United States Patent No. 5,834,252 END COMPLEMENTARY POLYMERASE REACTION assembles large polynucleotides from fragments and methods for amplifying and detecting target sequences (eg, in nucleic acid mixtures). It describes how. Crameri et al. (1998) Nature 391: 288-291 provide basic methods for gene reassembly, as in Crameri et al. (1998) Bio technoques 18 (2): 194-196.

More recently, a number of gene reassembly methods of simultaneously recombining and reconstructing genes have been described in USSN 60 / 118,813, filed February 5, 1999 by several applications of inventors and their co-researchers, such as Crameri et al. And USOL 60 / 141,049 filed June 24, 1999 and USOL 09 / 408,392 filed September 28, 1999, "OLIGONUCLEOTIDE MEDIATED NUCLEIC ACID RECOMBINATION," and USSN filed September 28, 1999 by Welch et al. 09 / 408,393 to "USE OF CODON-BASED OLIGONUCLEOTIDE SYNTHESIS FOR SYNTHETIC SHUFFLING". In such embodiments, a synthetic recombination method in which oligonucleotides corresponding to different homologues are synthesized and reassembled in a PCR or ligation reaction comprising oligonucleotides corresponding to one or more source nucleic acids to synthesize new recombinant nucleic acids. Used.

One advantage of oligonucleotide mediated recombination is the ability to recombine homologous nucleic acids with low sequence similarity or the ability to recombine non-homologous nucleic acids. In such low homology oligonucleotide shuffling methods, one or more fragmented nucleic acid sets are recombined with, for example, crossover familial diversity oligonucleotide sets. Each of the crossover oligonucleotides has multiple sequence diversity domains corresponding to multiple sequence diversity domains obtained from homologous nucleic acids or nonhomologous nucleic acids having low sequence homology. Fragmented oligonucleotides obtained by comparison with one or more homologous or nonhomologous nucleic acids can be hybridized to one or more crossover oligomer regions that hybridize to facilitate recombination. Such oligonucleotide sets may be selected in silico according to the methods herein.

When recombining homologous nucleic acids, sets of overlapping gene swarm shuffling oligonucleotides (obtained by comparing sets of homologous nucleic acid and oligonucleotide fragments corresponding to diversity regions and similarity regions derived from the comparisons) are hybridized and extended (e.g., For example, by assembly PCR, which provides a population of recombinant nucleic acids that can be selected for the desired characteristic or property. Typically, a set of overlapping shuffling gene swarm oligonucleotides comprises the form of a plurality of oligonucleotide members having co-continuous regions derived from a plurality of homologous target nucleic acids.

Typically, gene swarm shuffling nucleotides are provided by aligning homologous nucleic acid sequences to select conservative regions of the same sequence and conserved regions of various sequences. Multiple gene swarm shuffling oligonucleotides are synthesized corresponding to one or more sequence diversity regions (continuously or in parallel). A more detailed description of flock shuffling can be found in the aforementioned USSN 09 / 408,392.

A subset or fragment set of fragments used in an oligonucleotide shuffling approach may be used to cleave one or more homologous nucleic acids (eg, with DNase), or more generally oligonucleotides corresponding to multiple regions of one or more nucleic acids. It can be provided by synthesizing a set (typically an oligonucleotide corresponding to a full length nucleic acid is provided as part of a set of nucleic acid fragments). In the shuffling methods herein, such truncated fragments can be used in one or more recombinant reactions to produce recombinant nucleic acids with gene swarm shuffling oligonucleotides.

Assembling genes from single stranded complementary overlapping synthetic oligomers by PCR is a selection method performed in GAGGS. Used to control the stringency of reassembly and gene assembly, varying oligomer lengths to increase or decrease the number of sequence deviations during gene synthesis, number of oligomers in recombinant reactions, degree of overlap of oligonucleotides, and use in reassembly Optimization of the method, including the level and nature of sequence degeneracy, specific reaction conditions and specific polymerase enzymes can be performed.

The method may also be carried out in parallel mode, in which each member of the library comprising a plurality of genes for which continuous physical screening is scheduled is synthesized in a spatially isolated container or container arrangement or poolwise fashion, wherein the intended All or part of multiple genes are synthesized in a single container. Numerous synthetic methods for preparing other synthetic nucleotides are also known, and the particular advantages of using one and the other in performing GAGGS can be readily determined by one of ordinary skill in the art.

Unsequence

Sequence deconvolution is used to identify changes that correspond to a string (ie, corresponding to the physical sequence for a biopolymer) that results in a desirable change in the proper composition of a material (eg, polynucleotide, polypeptide, etc.). It can be carried out as polynucleotide variants known to have the intended properties.

Sequencing and other standard recombination techniques useful in the present invention, including sequencing, are described, for example, in Berger and Kimmel, Guide to Molecular Cloning Techniques, Methods in Enzymology vol. 152 Academic Press, Inc., San Diego, CA (Berger); Sambrook et al., Molecular Cloning-A Laboratory Manual (2nd Ed.), Vol. 1-3, Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1989 ("Sambrook") and Current Protocols in Molecular Biology, F.M.Ausubel et al., Eds., Current Protocols, Green Publishing Associates, Inc. And John Wiley & Sons, Inc. See Joint Venture, Aug. 1999 ("Ausubel"). In addition to sequencing GAGGS products, specific sites of subtraction can also be used to detect specific sequences. Sufficient information can also be found in Sambrook, Berger and Ausubel, (homologous), which leads the skilled person through restriction enzyme cleavage.

Methods of cloning and sequencing encoded molecules and / or characterizing cells, including plant cells and animal cells, for expression and selection, with GAGGS synthetic nucleic acids, such as methods for expressing proteins encoded by such nucleic acids, It is generally commercialized. Useful references for animal cell culture in addition to Berger, Ausubel and Sambrook include Freshney's book (Culture of Animal Cells, a Manual of Basic Techniques, third edition Wiley-Liss, New York (1994)) and references cited therein. Humanson's work (Animal Tissue Techniques, fourth edition WH Freeman and Company (1979)) and Ricciardelli et al., In Vitro Cell Dev. Biol. 25: 1016-1024 (1989). References to plant cell cloning, culture and regeneration include Payne et al. (1992) Plant Cell and Tissue Culture in Liquid Systems John Wiley & Sons, Inc. New York, NY (Payne); And Gamborg and Philips (eds) (1995) Plant Cell, Tissue and Organ Culture; Fundamental Method Springer Lab Manual, Springer-Verlag (Berlin Heidelberg New York) (Gamborg). Various cell culture media are described in Atlas and Paks (eds) The Handbook of Microbiology Media (1993) CRC Press, Boca Raton, FL (Atlas). Further information on plant cell culture can be found in the Life Science Research Cell Culture Catalog (1998) and Sigma-Aldrich, Inc. (StLouis, MO) published by Sigma-Aldrich, Inc. (StLouis, MO) (Sigma-LSRCCC). Commercially available literature such as Plant Culture Catalog of Sigma-PCCS) and Supplements (1997).

In vitro amplification methods can also be used to amplify, and the sequence GAGGS can also synthesize nucleic acids for cloning and selection. Techniques for mediating polymerase chain reaction (PCR), ligase chain reaction (LCR), Qβ replication enzymes, and other RNA polymerase mediated techniques (e.g. NASBA) include Berger, Sambrook and Ausubel (homologous) and Mullis et al. (1987) US Pat. No. 4,683,202; PCR Protocols A Guide to Method and Applications (Innis et al., Eds.) Academic Press Inc. San Diego, CA (1990) (Innis); Arnheim & Levinson (October 1, 1990) C & EN 36-47; The Journal Of NIH Reasearch (1991) 3, 81-94; Kwoh et al. (1989) Proc. Natl. Acad. Sci. USA 87, 1974; Lomell et al. (1989) J. Clin. Chem 35, 1826; Landergren et al. (1988) Science 241, 1077-1080; Van Brunt (1990) Biotechnology 8, 291-294; Wu and Wallace (1989) Gene 4, 560; Barringer et al. (1990) Gene 89, 117 and Sooknanan and Malek (1995) Biotechnology 13: 563-564. An improved method for cloning amplified nucleic acids in vitro is described in Wallace et al., US Pat. No. 5,426,039. Improved methods for amplifying large nucleic acids by PCR are summarized in Cheng et al. (1994) Nature 369: 684-685 and references therein, wherein PCR amplicons up to 40 kb are synthesized. PCR reassembly techniques are discussed as above. Those skilled in the art will understand that any essential RNA can be converted to double-stranded DNA suitable for restriction digestion, PCR amplification and sequencing using reverse transcriptase and polymerase. See literature by Ausbel, Sambrook and Berger (homologous).

If gene synthesis is essentially error-free (strict), i.e., where each member of the library is synthesized in a parallel manner in which it is synthesized in a fundamentally isolated space or container, deconvolution is performed by all library members. This is done by referring to the position encoding the index for each of the intended sequences. If the synthesis is performed in a pooled manner (ie library members are pooled through selection), one of a number of known polynucleotide sequencing techniques is used.

Iterative GAGGS Process

The repetitive nature of inductive evolution methods aimed at improving the desired properties of polynucleotides and polypeptides in a stepwise / round manner is well understood. In inducible evolution (DE) by GAGGS, a sequence of variants that results in any change in the level of the intended characteristic, where the level is optionally defined as an increase / decrease / ratio between the measurements of some of the characteristics. One or more unwrapped strings that encrypt s are used to construct a new library of strings in a new GAGGS round. Unlike conventional DNA shuffling, repetitive GAGGS does not physically manipulate polynucleotides to produce subsequent generations of gene diversity. Instead, GAGGS simply uses sequence information indicative of useful changes obtained as a basis for generating additional changes resulting in subsequent changes (ie, improvements) in the intended properties of the molecules encoded by the string. Repeated GAGGS may be used until the sequence of features evolves until a point at which the encoded polynucleotides and polypeptides have any desired properties or until no further changes in the features are obtained (eg, physical assays). Under conditions of enzymatic rotation until the theoretical diffusion rate limit is reached). Genetic algorithm parameters, namely gene synthesis methods and schemes, and physical assays and unsequencing methods, may vary in each of the different rounds / turns of inducible inheritance by repetitive GAGGS.

One particular advantage of this approach is that it can be more inducible as information about activity levels that are initially randomly or pseudorandomly accessible for library generation. For example, any heuristic teaching method or neural network method will be more effective at selecting "correct" (active) sequences. A number of studies are presented below, including principle component analysis using speech data, parameterization data, and the like.

Integration of GAGGS, DNA Shuffling, and Other Inductive Evolution Technologies

GAGGS constitutes a self-sufficient and independent technique that can be performed regardless of DNA shuffling or any useful inductive evolution method. However, one or more rounds of GAGGS can be performed with physical shuffling of nucleic acids, or site-specific mutagenesis, error-prone PCR (eg, inductive evolution processes). Or other diversity methods. The GAGGS generation library of polynucleotides can be nucleic acid shuffled and polynucleotides known to have characteristics of intent as a result of undergoing rounds of intra-silicon and / or physical shuffling are selected and sequenced to provide a string to evaluate the GAGGS process, or You can form strings for additional GAGGS operations. Therefore, GAGGS may be performed as a chain only technique, or may subsequently be subjected to shuffling, mutation, random priming PCR, and the like.

If the method of the present invention involves performing physical screening ("shuffling") and screening or selecting to evolve individual genes, whole plasmids, viruses, multiple gene clusters or even whole genes, the inventors And the skills of their collaborators are particularly useful. For example, the repetitive cycle of recombination and screening / selection may be further evolved (eg, by subsequent assembly oligonucleotide synthesis and gene synthesis / reproduction by assembly PCR) by evolving the target nucleic acid synthesized by performing GO on a string. Can be performed.

The following publications describe a number of iterative processes and / or related diversity generation methods that can be performed with the intrasilicone method of the present invention: Stemmer et al. (1999) "Molecular breeding of viruses for targeting and other clinical properties: Tumor Targeting "4: 1-4; Co-authored by Nesset et al. (1999) “Dna Shuffling of subgenomic sequences of subtilisin” Nature Biotechnology 17: 893-896; Chang et al. (1999) "Evolution of a cytokine using DNA family shuffling" Nature Biotechnology 17: 793-797; Mishull and Stemmer (1990) "Protein evolution by molecular breeding" Current Opinion in Chemicl Biology 3: 284-290; Christians et al. (1999) "Directed evolution of thymidine kinase for AZT phosphorylation using DNA family shuffling" Nature Biotechnology 17: 259-264; Crameri et al. (1998) "DNA shuffling of a family of genes from diverse directed evolution" Nature 391: 288-291; Crameri et al. (1997) "Molecular evolution of an arsenate detoxification pathway by DNA shuffling", Nature Biotechnology 15: 436-438; Zing et al. (1997) "Directed evolution of an effective fucosidase from a galactosidase by DNA shuffling and screening" Proceedings of the National Academy of Sciences, U.S.A. 94: 4504-4509; (1997) "Applications of DNA shuffling to Pharmaceuticals and Vaccines" Current Opinion in Biotechnology 8: 724-733; Crameri et al. (1996) "Constructions and evolution of antibody-phage libraries by DNA shuffling" Nature Medicine 2: 100-103; Crameri et al. (1996) "Improved green fluorescent protein by molecular evolution using DNA shuffling" Nature Biotechnology 14: 315-319; Gates et al. (1996) "Afinity selective isolation of ligands from peptide libraries through display on a lac repressor '' hesdpice dimer '"Journal of molecular Biology 255: 373-386; Stemmer (1996) "Sexual PCR and Assembly PCR" In: The Encyclopedia of Molecular Biology, VCH Publishers, New York pp447-457; Crameri and Stemmer (1995) "Combinatorial multiple cassette mutagenesis creates all the permutations of mutant and wildtype cassettes" BioTechniques 18: 194-195; Stemmer et al. (1995) "Single-step assembly of a gene and entire plasmid form large numbers of oligodeoxyribonucleotides" Gene, 164: 49-53; Stemmer (1995) "The Evolution of Molecular Computation" Science 270: 1510; Stemmer (1995) "Searching Sequence Space" Bio / Technology 13: 549-553; Stemmer (1994) "Rapid evolution of protein in vitro by DNA shuffling" Nature 370: 389-391; And Stemmer (1994) "DNA shuffling by random fragmentation and reassembly: In vitro recombination for molecular evolution" Proceedings of the National Academy of Sciences, U.S.A. 91: 10747-10751.

Further explanation of the DNA shuffling method can be found in the following United States patents filed by the inventors and their collaborators: United States Patent No. 5,605,793 to February 25, 1997, by Stemmer, "METHODS FOR IN VITRO RECOMBINATION"; United States Patent No. 5,811,238 (September 22, 1998) by Stemmer et al. "METHODS FOR GENERATING PLYNUCLEOTIDES HAVING DESIRED CHARACTERISTICS BY ITERATIVE SELECTION AND RECOMBINATION"; US Patent No. 5,830,721 to Stemmer et al., Nov. 3, 1998, "DNA MUTAGENESIS BY RANDOM FRAGMENTATION AND REASSEMBLY"; US Patent No. 5,834,252 by Stemmer et al. (November 10, 1998) "END-COMPLEMENTARY POLYMERASE REACTION" and US Patent No. 5,837,458 by Minshull et al. (November 17, 1998), "METHODS AND COMPOSITIONS FOR CELLULAR AND METABOLIC ENGINEERING ".

Moreover, details and manners for nucleic acid shuffling can be found in a number of PCT applications and foreign patent application publications: "DNA MUTAGENESIS BY RANDOM FRAGMENTATION AND REASSEMBLY" WO95 / 22625 by Stemmer and Crameri; Stemmer and Lipschutz "END COMPLEMENTARY POLYMERASE CHAIN REACTION" WO 96/33207; "METHODS FOR GENERATING POLYNUCLEOTIDES HAVING DESIRED CHARACTERISTICS BY ITERATIVE SELECTION AND RECOMBINATION" by Stemmer and Crameri WO 97/0078; "METHODS AND COMPOSITIONS FOR CELLULAR AND METABOLIC ENGINEERING" by Minshul and Stemmer WO 97/35966; "TARGETING OF GENETIC VACCINE VECTORS" by Punnonen et al. WO 99/41402; "ANTIGEN LIBRARY IMMUNIZATION" by Punnonen et al. WO 99/41383; "GENETIC VACCINE VECTOR ENGINEERING" by Punnonen et al. WO 99/41369; "OPTIMIZATION OF IMMINOMODULATORY PROPERTIES OF GENETIC VACCINES WO 9941368; by Punnonen et al." DNA MUTAGENESIS BY RANDOM FRAGMENTATION AND REASSEMBLY "by Stemmer and Crameri EP 0934999; "MODIFICATION OF VIRUS TROPISM AND HOST RANGE BY VIRAL GENOME SHUFFLING" WO9923107; "HUMAN PAPILLOMAVIRUS VECTORS" by Apt and others WO9921979; "EVOLUTION OF WHOLE CELLS AND ORGANISMS BY RECURSBIN SEQUENCE 98" by Del Cardayre et al. "METHOD AND COMPOSITIONS FOR POLYPEPTIDE ENGINEERING" by Pattern and Stemmer WO9827230; "METHODS OF OPTIMIZATION OF GENE THERAPY BY RECURSIVE SEQUENCE SHUFFLING AND SELECTION" by Stemmer et al. WO9813487.

Any of the following U.S. patent applications describe DNA shuffling and related technologies: September 29, 1998 (USSN 60 / 102,362), January 29, 1999 (USSN 60 / 117,729) by Pattern et al. ) And "SHUFFLING OF CODON ALTERED GENES", filed September 28, 1999 (USSN PCT / US99 / 22588); "EVOLUTION OF WHOLE CELLS AND ORGANISMS BY RECURSIVE SEQUENCE REOMBINATION", filed Jul. 15, 1999 by del Cardyre et al. (USSN 09 / 354,922); "OLIGONUCLEOTIDE MEDIATED NUCLEIC, filed February 5, 1999 (USSN 60 / 118,813), June 24, 1999 (USSN 60 / 141,049) and September 28, 1999 (USSN 09 / 408,392) by Crameri et al. "USE OF CODON-BASED OLIGONUCLEOTIDE SYNTHESIC FOR SYNTHETIC SHUFFLING", filed September 28, 1999 by ACID RECOMBINATION and Welch et al. (USSN 09 / 408,393).

Looking at the above publications, patents, published applications and United States patent applications, shuffling of nucleic acids (ie, "repetitive recombination") to provide novel nucleic acids with the intended properties can be performed by a number of sure methods. Any of these methods can be integrated with the methods of the present invention by binding nucleic acids corresponding to strings prepared by performing one or more GOs on one or more selected source strings. Any of these methods can be applied to the present invention to evolve GAGGS synthetic nucleic acids to produce novel nucleic acids with improved properties as discussed herein. Both methods of synthesizing such nucleic acids and nucleic acids produced by these methods correspond to aspects of the present invention.

In short, five or more different methods of recombination can generally be carried out (individually or in combination) in accordance with the present invention. First, nucleic acids prepared by synthesizing a set of nucleic acids corresponding to a string prepared by GO manipulation of a string, or useful homologs of these sets, or both, may be prepared by, for example, cleaving the nucleic acid to be recombined with DNase and then removing the nucleic acids. Recombination in vitro may be by any of a number of techniques discussed in the references, such as methods of ligation and / or PCR reassembly. Second, a set of nucleic acids corresponding to a feature strip prepared by GO manipulation of a string and / or useful homologues of such a set may be repeatedly recombined in vivo, for example by recombination between nucleic acids in a cell. Third, whole cell gene recombination methods can be used when the whole gene of a cell is recombined, optionally to a set of nucleic acids corresponding to a string produced by GO manipulation of the string or to a string prepared by a useful homologue of such a set. Spiking the genetic recombination mixture with a preferred library component, such as a corresponding set of nucleic acids. Fourth, synthetic recombination methods may be used, wherein oligonucleotides corresponding to different homologues are synthesized and reassembled in a PCR or ligation reaction comprising oligonucleotides corresponding to one or more source nucleic acids to synthesize new recombinant nucleic acids. Oligonucleotides may be prepared by standard nucleotide addition methods or may be prepared by ternary nucleotide synthesis methods. Fifth, only in silico recombination methods where GO is used in a computer to recombine sequence streams corresponding to nucleic acids or protein homologues may be useful. The resulting sequence of recombinant sequences can optionally be converted to nucleic acids by, for example, oligonucleotide synthesis / gene reassembly techniques by synthesizing nucleic acids corresponding to the recombinant sequences. Any of the above general methods of recombination can be performed independently or together repeatedly to synthesize various sets of recombinant nucleic acids.

The references provide other recombination techniques and a number of variations of these approaches in conjunction with the present invention. Regardless of the manner used above, the nucleic acids of the present invention can be recombined with each other nucleic acid and associated (or even unrelated) nucleic acids to produce a plurality of sets of recombinant nucleic acids, including homologous nucleic acids.

Other diversity generation methods can also be used to modify strings or nucleic acids. Additional variability can be introduced into the input or output nucleic acid by the method of altering each nucleotide or a continuous or discontinuous nucleotide, ie mutagenesis. Mutagenesis includes, for example, the following: recombination (PCT / US98 / 05223; Publication No. WO98 / 42727); Oligonucleotide-induced mutagenesis (Smith, Ann. Rev. Genet. 19: 423-462 (1985); Bostein and Shortle, Science 229: 1193-1201 (1985); Carter, Biochem. J. 237: 1-7 (1986); Kunkel, Nucleic acids & Molecular Biology, "The efficacy of oligonucleotide directed mutagenesis", Eckstein and Lilley, eds., Springer Verlag, Berlin (1987)). Such methods include the following methods: Oligonucleotide-induced mutagenesis (Zoller and Smith, Nucl. Acids Res. 10: 6487-6500 (1982), Methods in Enzymol. 100: 468-500 (1983) And Methods in Enzymol. 154: 329-350 (1987)); Phosphothioate modified DNA mutagenesis (Taylor et al., Nucl. Acids Res. 13: 8749-8764 (1985); Taylor et al. Nucle. Acids. Res. 13: 8765-8787 (1985); Nakayama and Eckstein , Nucl.Acids Res. 14: 9679-9698 (1986); Sayers et al., Nucl.Acids Res. 16: 803-814 (1988), mutagenesis using uracil-containing templates (Kunkel, Proc. Nat'l) Acad. Sci. USA 82: 488-492 (1985) and Kunkel et al., Methods in Enzymol. 154: 367-382); Mutagenesis using gapped double DNA (Kramer et al., Nucl. Acids Res. 12: 9441-9456 (1984); Kramer and Fritz, Methods in Enzymol. 154: 350-367 (1987); Kramer and Fritz , Nucl.Acids Res. 16: 7207 (1988); and by Fritz et al., Nucl.Acids Res. 16: 6987-6999 (1988). Further suitable methods include point mismatch repair (Fritz et al., Cell 38: 879-887 (1984)), mutagenesis using repair defective host strains (Carter et al., Nucl. Acids Res. 13: 4431-4443 (1985); Carter, Methods in Enzymol. 154: 382-403 (1987)), inducing deletion mutations (Eghtedarzadeh and Henikoff, Nucl. Acids Res. 14: 5115 (1986)), restriction selection and restriction Purification (Wells et al., Phil. Trans. R. Soc. Lond. A 317: 415-423 (1986)), inducing mutations by whole gene synthesis (Nambiar et al., Science 223: 1299-1301 1984); Sakamar and Khorana, Nicl.Acids Res. 14: 6361-6372 (1988); Wells et al., Gene 34: 315-323 (1985); and Grundstom et al., Nicl.Acids Res. 13: 3305 -3316 (1985) Mutation induction kits are commercially available (eg, Bio-Rad, Amersham International, Anglian Biotechnology).

Other methods of generating diversity are described in US Pat. No. 5,756,316; US Patent No. 5,965,408; Ostermeier et al. (1999) "A combinational approach to hybrid enzymes indepedent of DNA homology" Nature Biotech 17: 1205; US Patent No. 5,783,431; US Patent No. 5,824,485; US Patent No. 5,958,672; Jirholt et al. (1998) "Exploiting sequence space: shuffling in vivo formed complementarity determining regions into a master framework" Gene 215: 471; US Patent No. 5,939,250; WO 99/10539; WO 98/58085; WO 99/10539 et al. The method of diversity generation can be combined with each other in a silico operation, any binding method chosen by the user, or performed with a shuffling reaction to create diversity of nucleic acids, which can also be screened using any useful screening method. Can be. Any nucleic acid synthesized in the following recombination or other diversification reaction can be selected for the desired activity. In the context of the present invention, this can be identified by testing for detectable or assayable activity by a suitable assay in the art. Many related (or unrelated) properties can be assayed using any useful assay.

Thus, recombinant nucleic acids are synthesized by repeatedly recombining one or more polynucleotides of the invention (synthesized by the GAGGS method) with additional nucleic acids that are part of the invention. One or more additional nucleic acids may comprise other polynucleotides of the invention; Optionally, or in the alternative, or in addition, one or more additional nucleic acids may include, for example, nucleic acids encoding naturally occurring sequences or subsequences thereof, or any homologous sequences or subsequences thereof.

The recombination step can be performed in vivo, in vitro or in silico as described in detail above and herein. In addition, a cell included in the present invention may be a recombinant nucleic acid prepared by repeatedly recombining a nucleic acid set forth herein, a nucleic acid library, and obtained by recombining (or repeatedly recombining) a nucleic acid comprising the library or the nucleic acid provided herein with another nucleic acid. A group of cells, vectors, viruses, etc., or any other nucleic acid comprising any recombinant nucleic acid. Corresponding sequence sequences in a database residing on a computer system or computer readable medium are aspects of the present invention.

As an example, conventional physical recombination processes generally begin with at least two substrates that exhibit some identity to each other (ie, at least about 30%, 50%, 70%, 80%, or 90% sequence identity), but at any position Are different from each other (however, only in the silico or crossover oligonucleotide mediated form, the homology of nucleic acids is slight or hardly expressed). For example, two or more nucleic acids can be recombined herein. The difference between nucleic acids can be in the form of any mutation, such as, for example, substitutions, insertions, and deletions. Often different sections differ at about 1-20 positions. In recombinant methods that increase the diversity relative to the starting material, the starting materials differ from each other at two or more nucleotide positions. That is, if only two substrates are present, there are two or more diversification sites. For example, if three substrates are present, one substrate may be different from the second substrate at a single location, and the second substrate will also be different from the third substrate at a different single location. Of course, provided that only one initial string is provided, any GO can be used to modify the nucleic acid to produce various arrays of nucleic acids that can be screened for desired activity.

In the physical shuffling process, the starting DNA fragments can be different naturally occurring variants from one another, for example, allelic or variant variants. More typically, they are derived from one or more homologous nucleic acid sequences. The fragments can also be obtained from non-alleles that exhibit some degree of structural and generally functional correlation. Starting DNA fragments can also be mutually inducing variants. For example, one DNA fragment can be prepared by error prone PCR replication or by substitution of a mutagenic cassette. Inducible mutant strains may also be prepared by promoting one (or two) of fragments in a mutant strain. In this situation, strictly speaking, the second DNA fragment is not a single fragment but a large community of related fragments. The different segments forming the starting material are the same or almost the same in length. However, this need not be the case; For example, one segment may be an subsequence of another segment. The fragment may be provided as part of a larger molecule such as a vector or may be in isolated form. In one option, the target nucleic acid is derived from DE by GAGGS.

Codon Diversification Oligonucleotide Method

Codon diversification oligonucleotides are oligonucleotides that are similar in sequence but have one or more base mutated mutations, where one or more mutations can occur to encode a variation of an amino acid. They can be synthesized using ternary tri-nucleotides, ie, codon-based phosphoramidite coupling compounds, where ternary nucleotide phosphoramides, which represent codons for 20 amino acids, describe the entire codon as a solid phase. It is used to introduce into the oligonucleotide sequence synthesized by. Preferably all oligonucleotides of selected length (eg, about 20, 30, 40, 50, 60, 70, 80, 90 or 100 or more nucleotides) that bind the selected nucleic acid sequence are synthesized. In the present invention, codon variant oligonucleotide sequences may be based on sequences obtained from a selected set of nucleic acids synthesized by any of the methods mentioned herein. Further explanation of ternary nucleotide synthesis can be found in USSN 09 / 408,393, "USE OF CODON VARIED OLIGONUCLEOTIDE SYNTHESIS FOR SYNTHETIC SHUFFLING," filed September 28, 1999 by Welch et al. Oligonucleotides can be synthesized by standard addition methods or prepared by ternary nucleotide synthesis assays. The advantage of selecting a change corresponding to the difference in encoded amino acids is the modification by ternary codons that rarely cause frame shifts (thus similar to a few inactive library members). In addition, synthesis focused on codon modifications rather than simple base variations reduces the total number of oligomers required for the synthesis process.

Oligomer set

In general, oligomer sets can be combined for assembly in a number of different ways that correlate with genetic phenomena and operators at the physical level, and in different combinations of these ways.

As mentioned, nested sets of oligonucleotides can be hybridized and extended after synthesis to form nucleic acids of full length. Full length nucleic acid is any nucleic acid that is longer than the oligomer used by the investigator in the gene reconstruction method of interest. This may correspond to any percentage of naturally occurring full length sequence, eg, at least 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, or 90%.

Oligomer sets often have nucleotide overlapping sequences that facilitate gene reconstruction, which are often about 5, sometimes about 10, often about 15, generally about 20 or more. The oligomer set is optionally simplified for the purpose of gene reconstruction, ie, where there is an accidentally overlapping region, that is, for the reconstruction of genes in which repetitive sequence elements are present or designed into the gene sequence to be synthesized. The length of the oligomers in the set can be the same or different as the sequence regions can overlap. In order to facilitate hybridization and extension (eg, during the PCR cycle), the overlap zones are optionally designed to have similar melting temperatures.

The source sequence can be grided (conceptually or physically) and the ordinary sequence can be used to select a common sequence oligomer, thereby binding the oligomer member to one or more sets to produce full length nucleic acid. This reduces the oligomeric member required for. Similarly, oligonucleotides with some sequence similarity may be synthesized by pooling the oligomer pools being synthesized into different pools while adding heterologous bases, or optionally recombining in subsequent steps requiring the same addition reaction to the oligomers (pooling). ) May be synthesized by successive pooled synthesis and / or split synthesis. In the oligomer shuffling mode, heterologous oligomers corresponding to many different sources can be split or recombined during synthesis. In simple degenerate synthesis, one or more nucleic acid bases can be added during a single synthesis step resulting in two or more mutations in the resulting two or more oligonucleotide sequences. The relative percentages in nucleic acid base addition can control bias synthesis for one or more source sequences. Similarly, partial generation can be performed to prevent the insertion of stop codons in the synthesis of degenerate oligonucleotides. Thus, in the split and pull mode, some oligomers do not extend arbitrarily during every synthesis step (to prevent frame movement, some oligomers corresponding to one or more codons do not extend in this step).

When constructing an oligomer, the crossover oligomer may be constructed at one or more variation points between two or more source sequences (base changes or other variations are treated as points of crossover occurrence). The crossover oligomer has a region of the same sequence as the first source sequence, followed by the same region as the second source sequence at the location where the crossover point is present. For example, all naturally occurring mutations can be crossover points.

Another method of biased sequence recombination is to spike the oligonucleotide mixture into one or more source nucleic acid fragments (when one or more nucleic acids are fragmented, the resulting fragments are spiked into the recombinant mixture at different frequencies to one or more sources). To bias the recombination results. Recombination phenomena can also be manipulated simply by deleting one or more oligonucleotides corresponding to one or more sources obtained from the recombinant hybrids.

In addition to using a bunch of related oligonucleotides, diversity can be controlled by adding selected, pseudo-random or random oligomers to an extension mixture that can be used to bias the resulting full length sequence. Similarly, mutagenic or nonmutagenic conditions, which form various libraries of somewhat full length nucleic acids, can be selected for PCR extension.

In addition to mixing oligomer sets corresponding to different sources in the extension mixture, oligomer sets corresponding to only one source can be extended to reconstitute the sources. In both cases, any resulting full length sequence can be fragmented and recombined, as in the DNA shuffling method disclosed in the references cited herein.

For a number of other oligonucleotide sets and synthetic variants that may be correlated with genetic phenomena and effectors at the physical level, see February 5, 1999 (USSN 60 / 118,813) and June 1999 24 by Crameri et al. "OLIGONUCLEOTIDE MEDIATED NUCLEIC ACID RECOMBINATION", filed on USSN 60 / 141,049 and September 28, 1999 (USSN 09 / 408,392) and "OLIGONUCLEOTIDE SYNTHESIS FOR SYNTHETIC SHUFFLING, filed September 28, 1999 by Welch et al. "(USSN 09 / 408,393).

Targets for Codon Deformation and Shuffling

Essentially any nucleic acid can be shuffled using the GAGGS method herein. Attempts to identify hundreds of thousands of known nucleic acids are not made herein. Known sequence reservoirs for proteins generally include GenBank EMBL, DDBJ and NCBI. Other repositories can be easily found by searching the Internet.

A preferred group of targets for the GAGGS method include erythropotin (EPO), therapeutic proteins such as insulin, peptide hormones such as human growth hormone; Growth factor and epithelial neutrophil active peptide-78, GROα / MGSA, GROβ, GROγ, MIP-1α, MIP-1δ, MCP-1, epidermal growth factor, fibroblast growth factor, hepatocyte growth factor, insulin-like growth factor, interferon , Interleukin, keratin synthetic cell growth factor, leukemia inhibitory factor, oncostatin M, PD-ESCF, PDGF, fluorotropin, SCF, c-kit ligand, VEGEF, G-CSF and the like. These proteins and their encoding nucleic acids are commercially available (see, for example, the Sigma BioSciences 1997 catalog and price list) and in any case, the corresponding genes are well known.

Other groups of preferred targets for GAGGS are transcriptional or expression activators. Standard transcriptional and expression activators include genes or proteins that regulate cell growth, differentiation, modulation, and the like. Expression and transcriptional activators can be found in eukaryotic cells, including animals including prokaryotic cells, viruses and fungi, plants and mammals that provide a wide range of therapeutic targets. Expression activators and transcriptional activators, for example, binding to receptors, stimulation of signal transduction cascades, regulation of transcription factor expression, binding to promoters and enhancers, binding to proteins that bind to promoters and enhancers, It will be appreciated that transcription is regulated by a number of mechanisms such as DNA loosening, splicing of precursor-mRNA, polyadenylation of RNA and RNA degradation. Expression activators include cytokines, inflammatory molecules, growth factors such as interleukin (e.g., IL-1, IL-2, IL-8, etc.), interferon, FGF, IGF-I, IGF-II, FGF, Tumor genes such as PDGF, TNF, TGF-α, TGF-β, EGF, KGF, SCF / c-Kit, CD40L / CD40, VLA-4 / VCAM-1, ICAM-1 / LFA-1 and Hyalurin / CD44 Products and their receptors; For example, signal transduction molecules such as Mos, Ras, Raf and Met and their tumor gene products; And receptors for transcriptional activators and inhibitors, such as, for example, p53, Tat, Fos, Myc, Jun, Myb, Rel and steroid hormones such as, for example, estrogen, progesterone, testosterone, aldosterone, LDL receptor ligands, and Contains Corticosterone

Similarly, proteins obtained from infectious infectious organisms include, for example, Aspergillus, Candida; Bacteria, specifically E. Coli and Staphylococcus (e.g., aureus), Streptococcus (e.g., used as pathogenic bacterial models) , Pneumoniae), Clostridia (e.g. perfringens), Neisseria (e.g. gonoroeae), Enterobacteriaceae (e.g. coli), Helicobacter (e.g. pylori), Vibrio (e.g. cholerae) )), Capylobacter (e.g. jejuni), Pseudomonas (e.g. aruginosa), Haemophilus ( For example, influenza, Bordetella (eg, Pertusis), Mycoplasma ( For example, pneumoniae), Ureaplasma (e.g. urealyticum), Legionella (e.g. pneumophila), Spirochetes (e.g. Treponema, Leptospira and Borrelia), Mycobacteria (e.g. tuberculosis) ), Smegmatis), Actinomyces (e.g. israelii), Norcardia (e.g. asteroides), chlamydia Clamydia (for example, trachomatis, Rickettsia, Coxiella, Ehrilichia, Rochalimaea, Brucella) , Genus Yersinia, Fracisella and Pasteurella; Sporeworms (eg Plasmodia, Rhizopod) (eg Entamoeba) and flagella (Trypanosoma, Leishmania, Trichomonas) Protozoa, such as Trichomonas, Giardia, etc .; (+) RNA viruses (e.g., Poxvirus, such as vaccinia; e.g., polio species and Picornaviruses such as; Togaviruses such as, for example, rubella; Flaviviruses, such as, for example, HCV; and Coronaviruses. , (-) RNA virus (e.g., Rhabidovirus, such as VSV; Paramyxovirus, such as, for example, RSV; Orthomyxovirus, such as, for example, influenza) Bunyavirus and arenavirusers Genus (Arenavirus), dsDNA virus (eg Reovirus), RNA-> DNA virus, ie, retroviruses such as, for example, HIV and HTLV, and any DNA- such as hepatitis B virus. It is described in more detail below, including medically important bacteria such as RNA viruses.

For example, other nucleic acids encoding proteins suitable for non-medical use, such as toxins or transcriptional inhibitors such as insect pests, fungi and the like and weeds such as weeds, are also preferred as targets of GAGGS. Industrially useful enzymes such as monooxygenases, proteases, nucleases and lipases are also preferred as targets. As an example, subtilisin can be evolved by the shuffling selection of the subtilisin gene (von der Osten et al., J. Biotechnol. 28: 55-68 (1993), to provide a subtilisin encoding nucleic acid. ). Proteins that aid in folding, such as chaperin, are also desirable.

Known preferred genes suitable for codon alteration and shuffling also include the following: alpha-1 antitrypsin, agiostatin, antihemostatic factors, apolipoproteins, apoproteins, atrial sodium diuretic factor, atria Sexual natriuretic polypeptides, atrial peptides, CXC chemokines (eg, T39765, NAP-2, ENA-78, Gro-a, Gro-b, Gro-c, IP-10, GCP-2, NAP- 4, SDF-1, PF4, MIG), calcitonin, CC chemokines (eg, monocyte chemoattractant protein-1, monocyte chemoattractant protein-2, monocyte chemoattractable protein-3, Monocyte inflammatory protein-1 alpha, monocyte inflammatory protein-1 beta, RANTES, I309, R83915, R91733, HCC1, T58847, D31065, T64262), CD40 ligand, collagen, colony stimulating factor (CSF), complement factor 5a, complement inhibitor , Complement receptor 1, factor IX, factor VII, factor VIII, factor X, fibrinogen, fibronectin, gluco Lebrosidase, gonadotropin, hedgehog protein (e.g., sonic, indian, desert), hemoglobin (for blood substitution; for radioactivity detection), hirudin, human serum albumin, lactoferrin, luciferase, neuturin , Neutrophil inhibitor (NIF), bone morphogenetic protein, parathyroid hormone, protein A, protein G, relaxin, lenin, salmon calcitonin, salmon growth hormone, soluble complement receptor I, soluble I-CAM 1, soluble interleukin receptor ( IL-1, IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-9, IL-10, IL-11, IL-12, IL-13, IL- 14, IP-15), soluble TNF receptor, somatomedin, somatostatin, somatotropin, streptokinase, superantigen, ie, staphylococcus endotoxin (SEA, SEB, SEC1, SEC2, SEC3, SED , SEE), toxin shock syndrome toxin (TSST-1), ectotoxic toxins A and B, pyrogen exotoxins A, B and C and M. M.arthritides mitogen, superoxide dismutase, thymosin alpha 1, tissue plasminogen activator, tumor necrosis factor beta (TNF beta), tumor overkill factor receptor (TNFR), tumor Necrosis factor alpha (TNF alpha) and urokinase.

Other genes preferred for shuffling include p450 (the enzyme represents a very diverse set of naturally occurring diversity); See, for example, Ortiz de Montellano (ed.) (1995) Cytochrome P450 Structure Mechanism and Biochemistry, Second Edition Plenum Press (New York and London) and references cited therein, namely Cytochrome P450. Other monooxygenases and deoxygenases, acyl transferases (cis-diol), halogenated hydrocarbon dehalogenases, methyl transferases, terpene synthetases and the like can be shuffled.

Use of sympathetic genes in inducible evolution, including "diplomacy"

One of the factors associated with the standard shuffling cluster of source genes is the degree of identity of the physically recombined genes. Recombination of limited identical genes without crossovering oligonucleotides is very difficult and often produces shuffled libraries with unacceptable killing rates, or no chimeras are formed, inactive, and functional libraries are not created. Do not. In one aspect, the present invention provides for the design of " diplomat " sequences in silico with moderate homology to each of the sequences to be recombined, thereby facilitating crossover between the sequences. This difficulty is overcome by promoting chimeric formation. This diplomat sequence aligns sequences that select sympathetic sequences and modifies the codons to optimize similarity between the various nucleic acids, thereby imparting moderate sequence similarity within the diplomat sequence relative to the sequence to be recombined. May be a string produced by any of a number of GOs.

As mentioned, one method of designing a diplomat sequence is to select a sympathetic sequence, for example, using any of the methods herein. The sympathetic sequences are synthesized by comparing and line-up / pile-up gene clusters (DNA sympathetic), or by line-up / fileup (aa sympathetic) amino acid sequences. In the latter case, the amino acid sympathy sequence is further enhanced for homology or enhanced expression of the host organism or selected for use of alternative codons in order to access an alternative set of amino acid codons. It is arbitrarily reversed using the preferred codons. The other is to synthesize sympathetic sequences using different sets of gene clusters.

Moreover, the sympathetic sequence itself can encode an improved enzyme. International Coference held on "Development of Heat Stable Phytase" -a consensus phytase had by "Enzyme Opportunities in the Next Millenium", Chicago, IL, May 5-7 (presenter: Dr. Luis Pasamonted, Roche Vitamins, Inc.). See an increase of 16 degrees C in thermostability. Another example of a sympathetic protein with improved properties is sympathetic interferon (IFN-con1).

Thus, when performed independently or together with any of the methods herein, the Diplomat sequence can be designed using selected GO specimens, and can optionally be physically synthesized and shuffled using any of the techniques herein. have.

Standard process steps for back translation and oligomer design

An automated process step (eg, performed in a digital system as described herein) that performs the following functions facilitates oligonucleotide selection in the synthetic shuffling techniques herein.

For example, the system may include a set of instructions capable of entering the amino acid sequence of the protein cluster of interest.

The sequences are translated back into any desired codon usage parameter, for example to be used for expression or to facilitate recombination, or to optimize sequence alignment, or to optimize usage parameters for one or more organisms for both cases. do. For example, codon usage can be selected for multiplexed expression hosts such as, for example, E. coli or S. cerevisiae. In some cases, simply optimizing the use of codons for expression in host cells, such as sequences lose their naturally occurring species codon bias, can produce more similar homologous sequences.

The symmetric sequences can be prepared by aligning the sequences, optionally degenerate codons can be used.

Oligonucleotides are designed for one or more synthetic constructs corresponding to synthetic nucleic acids for shuffling. Input parameters for oligonucleotide design include minimum and maximum lengths, minimum lengths of terminal identical sequences, maximum degeneracy per oligonucleotide, oligomeric overlap lengths, and the like.

As mentioned, an alternative reverse translation to complete the use of optimal codons for expression in a particular organism is a reverse translation sequence for optimizing nucleotide homology between members of the herd. For example, amino acid sequences are aligned. All possible codons for each amino acid are determined and codons are selected at each location that minimize the difference between the cluster of aligned sequences.

Use of shuffling herds to identify structural motifs related to specific protein properties

It is often interesting to identify protein regions that impart specific properties and to facilitate the functional manipulation and design of related proteins. Such identification is typically performed by using structural information obtained by biophysical techniques such as X-ray crystallography. The present invention provides an alternative method of obtaining variants and analyzing them for specific properties related to sequence motifs.

Naturally occurring enzyme sequences that catalyze similar or identical reactions can vary widely: the sequences can only be about 50% or less identical. Each of these enzyme groups can catalyze essentially the same reaction, and the different properties of the enzymes can be very different. These properties include stability to temperature and organic solvents, optimum pH, solubility, ability to maintain activity when modified, ease of expression in different host systems, and the like. These also include catalytic properties including activity (k _cat and K _m ), acceptable range of substrates and compounds used.

The methods described herein also provide for non-catalytic proteins (i.e. ligands such as cytokines) and nucleic acid sequences (which can be derived by a number of different ligands) whenever the multi-functional dimension is encoded by a cluster of homologous sequences. Such as a promoter).

Due to the diversity between enzymes with similar catalytic functions, it is generally not possible to correlate specific properties with each amino acid present at any position because there are significant differences between these amino acids. However, variant libraries can be prepared from a cluster of homologous naturally occurring sequences by DNA bunch shuffling. These libraries can include the diversity of the original set of sequences in a number of different combinations. Individuals obtained from the library can then be tested under specific sets of property conditions to determine the optimal combination of sequences obtained from acupuncture sets for these conditions.

If a parameter of one of the assay conditions is changed, different individuals obtained from the library will be identified as best performers. Since the screening conditions are very similar, most amino acids can be conserved between two sets of best attendants. Comparing the sequences under two different conditions (eg in silico) will identify differences in the sequences that give the difference in performance. The main analysis of a member is a powerful tool used to identify sequences that impart specific properties. For example, Partek Incorported (St. Peters, Missouri; www.partek.com) is software for pattern recognition that can be applied to genetic algorithms for multivariate data analysis, responsive representation, various selection methods, and neurological & statistical modeling. (E.g. supplied by Partek Pro 2000 Pattern Recognition Software). Relationships can be analyzed, for example, by scatter plots mapped by Principal Components Analysis (PCA), biplots, scatter plots mapped by Multi-Dimensional Scaling (MDS), Starplot, and the like.

Once sequence motifs are identified, proteins are manipulated, for example, in any number of ways. For example, the identified changes can be arbitrarily and deliberately introduced into other sequence backgrounds. Sequences that confer different specific properties can be combined. For a more thorough study, identified sequence regions that are important for a particular function can be targeted, for example, by completing randomization using selected degenerate oligonucleotides in an in silico process.

Identification of source contributors to chimeras produced through herd shuffling

Beam experiments provide a way to identify source incentives for chimeras produced through herd shuffling.

In this method, the source gene sequence and chimeric sequence are input and each chimera is compared with each source sequence. A sequence map and a graphical map are then generated for each chimera representing the source source of each chimeric fragment. Their correlation with functional data facilitates the selection of sources for novel libraries that can be prepared by any of the methods mentioned herein and can be screened for any desired functionality by identifying sources imparting specific properties. You can.

In one embodiment, herd 3 and herd 4 genes confer activity, for example at pH 5.5, while herd 1 and herd 2 genes were more active at pH 10. Therefore, when applied at low pH, the source composition from which the library is made is biased at 3 and 4, whereas at high pH, libraries biased to superior 1 and 2 will be suitable. Thus, GO will be implemented to select oligonucleotides to reconstruct genes significantly from herds 3 and 4. Further details regarding the gene blending method used for oligonucleotide shuffling are described by Crameri et al. In the 5 February 1999 application (USSN 60 / 118,813) and the June 24, 1999 application (USSN 60 / 141,049). "OLIGONUCLEOTIDE MEDIATED NUCLEIC ACID RECOMBINATION", filed September 28, 1999 (USSN 09 / 408,392).

Gene formulation is similar to principal component analysis (PCA) for identifying specific sequence motifs. Principal Component Analysis (PCA) is the transformation of the number of correlation (possibly) variables into the number (or smaller) of uncorrelated variables called "major components". The first major component corresponds to the amount of variation in the data, and each successive component corresponds to the amount of variation retention. Typically, the principal component analysis involves a square symmetric SSCP (net-to-square sum and multiply) matrix, a covariance (scale-to-scale sum of squares and multiply) matrix, or correlation (square and multiply from standardized data). Performed on a matrix. As a result of the analysis, it was found that the objects of the form SSCP and covariance are similar. Correlation objects are used when the variance of each variant is substantially different, or when the units of measure of each variation are different. The purpose of the principal component analysis involves, for example, finding or reducing the dimensions of the data set to identify new meaningful variable values.

The main difference is that when the PCA identifies a particular motif by practicing the present invention, which source is used in the mixture (and thereby constitutes a novel naturally occurring gene based library) and also changes in derived or randomized methods By identifying possible abstract areas, one can obtain information on how well they are adapted to more comprehensive synthetic shuffling.

Partek (PCA) software, discussed above, has an "experimental design" component that identifies variables that appear to affect specific functions. When this is applied in this example, it is useful in an iterative process in which library clusters are constructed and screened and the resulting chimeras are analyzed for functional correlation with sequence variation. This is used to predict regions of the sequence that exhibit a particular function and the library is selected in silico by any desired GO inducible change in the region correlated with functionality. Focused libraries exhibiting diversity in the region are constructed by, for example, oligonucleotide synthesis methods as described herein. The resulting library members (chimeras) are analyzed for functional correlation with sequence variations. This method of research focuses on searching for variations in the sequence space of the most appropriate region of proteins or other related molecules.

After unpacking the active sequence, the resulting sequence information is used to derive further predictions for intrasilico operation, eg, neural net training approach.

For example, neural network research can be performed in conjunction with performing genetic algorithmic programs. For example, NNUGA (Neural Network Using Genetic Algorithm) is a useful program that links neural networks with genetic algorithms (http://www.cs.bgu.ac.il/~omri/NNUGA). /). Introduction to neural networks is described, for example, by Kevin Gurney (1999) An Introduction to Neural Networks, UCL Press, 1 Gunpowder Square, London EC4A 3DE, UK and http://www.shef.ac.uk/psycology/gurney/ notes / index.html. References to further useful neural networks can be found in the literature cited above with respect to genetic algorithms, for example Christopher M. Bishop (1995) Neural Networks for Pattern Recognition Oxford Univ Press; ISBN: 0198538642; Brian D. Ripley, N. L. Hjort (Contributor) (1995) Pattern Recognition and Neural Networks Cambridge Univ Pr (Short); ISBN: 0521460867.

Linking Theoretical Protein Design and Shuffling

Protein design cycles, including cycling between theory and experiment, have recently developed theoretical protein design. The organizer's approach to classifying proteins according to their local environment improved the degree of adequate energy generation. Protein design programs can be used to construct and modify proteins with an arbitrarily selected set of design samples. For example, http://www.mayo.caltech.edu/; By Gordon and Mayo (1999) "Branch-and-Terminate: A Combinatorial Optimization Algorithm for Protein Design" Structure with Folding and Design 7 (9): 1089-1098; Street and Mayo (1999) "Intrinsic β-sheet Properties Result from van der Waals Interactions Between Side Chains and the Local Backbone" Proc. Natl. Acad. Sci. USA, 96, 9074-9076; Gordon et al. (1991) "Energy Function for Protein Design" Current Opinion in Structural Biology 9 (4): 509-513; Street and Mayo (1999) "Computational Protein Design" Structure with Folding and Design 7 (5): R105-R109; By Strop and Mayo (1990) "Rubredoxin Variant Folds Without Iron" J. Am. Chem. Soc. 121 (11): 2341-2345; By Gordon and Mayo (1998) "Radical Perfprmance Enhancements for Combinatorial Optimization Algorithms based on the Dead-End Elimination Theorem" J. Com. Chem. 19: 1505-1514; By Malakauskas and Mayo (1998) "Design, Structure, and Stability of a Hyperthermophilic Protein Variant" Nature Struct. Biol. 5: 470; Street and Mayo (1998) "Pairwise Calculation of Protein Solvent-Acessible Surface Areas" Folding & Design 3: 253-258; Dahiyat and Mayo (1997) "De Novo Protein Design: Fully Automated Sequence Selection" Science 278: 82-87; Dahiyat and Mayo (1997) "Probing the Role of Packing Specificity in Protein Design" Proc. Natl. Acad. Sci. USA 94: 10172-10177; Dahiyat et al. (1997) "Automated Design of Surface Positions of Protein Helices" Prot. Sci. 6: 1333-1337; Dahiyat et al. (1997) "De Novo Protein Design: Towards Fully Automated Sequence Selection" J. Mol. Biol. 273: 789-796; And Harney et al. (1997) "Structural basis for thermostability and identification of potential active site residues for adenylate kinases from the archaeal genus Methanococcus" Proteins 28 (1): 117-30. Such design methods generally rely on energy generation to assess the quality of different amino acid sequences for the target protein structure. In any case, the designed or modified protein or string corresponding to the protein may be directly shuffled or reverse translated in silico by physical shuffling. Therefore, one aspect of the present invention is to link high throughput theoretical designs and screen genes in silico or in physical shuffling to exhibit the desired activity.

Similarly, Ornsein et al. (Http://www.emsl.pnl.gov:2080/homes/tms/bms.html; Curr Opin Rtruct Biol (1999) 9 (4): 509-13) Molecular dynamic simulations such as " theoretical " redesign the enzyme by biomolecular modeling & simulation to find novel enzyme forms that are less likely to be biologically evolved. For example, theoretically redesigning p450 cytochromes and alkanes dehalogenases is the goal of current theoretical design. Any protein designed theoretically (eg, a novel p450 homologue or novel alkaline dehydrogenase protein) can evolve by reverse translation and shuffling with other designed proteins or naturally occurring homologous enzymes associated therewith. . A detailed description of p450 can be found in Oritz de Montellano (ed.) (1995), Cytochrome P450 Structure and Mechanism and Biochemistry, Second Edition Plenum Press (New York and London).

Protein crystal structures can be compared to predict crossover points based on sequence homology rather than structural homology and crossover can be performed by oligomers that direct chimeric formation as described herein.

Modeling multiple variant sequence activities of proteins; Optimization of enzyme activity by theoretical statistics

This chapter describes how to analyze a large number of related protein sequences using recent statistical methods and how to obtain novel protein sequences with desired aspects using theoretical statistics and multiple mutation analysis.

Preliminary Multivariate Data Analysis

Multivariate data analysis and experimental design are widely used in industry, government, and research institutes. They can typically be used in areas such as gasoline production or optimization of chemical processes. In the traditional process of gasoline production, at least 25 different additives are added in different combinations by different amounts. In addition, the output of the end product is multifactorial (eg energy level, degree of contamination, stability, etc.). Using experimental design, a limited number of test formulations can be produced in a non-random manner in order to make the best use of the relevant "formulation space" of all additive amounts and presence. Different formulations are actually analyzed for proper measurements. By plotting the data points in a multivariate fashion (multidimensional fashion), the formulation space can be visualized graphically, leading to the ideal combination of additives. One of the most commonly used statistical methods in this type of analysis is Principal Component Analysis (PCA).

In this embodiment, this type of matrix is used to correlate each multi-dimensional data point with a particular output vector to identify the relationship between the matrix of dependent variable Y and the matrix of predictor variable X. A common analysis method of this type is the Partial Least Square Projection to Latent Structure (PLS). For example, the method is used in materials sciences to predict the properties of new compounds or to analyze investment experts on changing stock values. Each data point consists of hundreds of different parameters, each plotted for each in an n-dimensional hyperspace (one dimension for each parameter). Manipulation is performed within a computer system, which adds to the number of dimensions required to handle the input data. The above-mentioned methods (PCA, PLS, etc.) that help the discovery of the projection plane and accurately analyze hyperspace are mentioned above.

Preliminary sequencing

Nucleotide or amino acid sequence analysis has traditionally been focused on qualitative pattern recognition (eg, sequence classification. This includes methods for identifying sequences based primarily on similarity.) Has an excellent effect on prediction and identification in classification, but is not always correlated with quantitative values, for example, a consensus transcriptional promoter may not be the preferred promoter for a particular application, but instead is relevant. It can be a normal promoter among the ordered groups of sequences.

To determine the quantitative characteristics of the relevant biological sequence (DNA / RNA or amino acid), systematic variation (ie, systematic absence of similarity) can be analyzed from the aligned sequence with the relevant biological activity. By applying different multiple variability analysis methods (such as PLS) to protein sequences, one can predict sequences that are more active than the best sequences present in the analyzed set. Experimental data showing the methods generally applied are described below.

Preliminary Promoter Activity Multiple Variation Assay

One of the very few references that applies multiple mutation analysis data to biological sequences is to analyze a limited set of transcriptional promoters to see if they can predict a stronger promoter than any of those found in the trading set. [Jonsson et al. (1993), Nucleic Acids Res. 21: 733-739.

In this example the promoter sequence was parameterized. For simplicity, the physical-chemical differences between each nucleotide (A, C, G, T) were set equal, ie no nucleotides more closely related to any other nucleotide were set aside. These are shown as four opposite corners of the cube forming a perfect tetrahedron. By assigning the origin to the center, each corner is represented by numerical coordinates. Since the descriptors represent only an even distribution of each property, any nucleoside may be placed at any corner (see FIG. 17).

Previous studies (Brunner and Bujard (1987) EMBO J. 6, 3139-3144; Knaus and Bujard (1988) EMBO J. 7: 2919-2923; Lanzer and Bujard (1988) Proc. Natl. Acad. Sci. 85: 8973 As a result, 28 sets of promoters studied in detail in the same literature were analyzed. Such promoters included E. coli promoter, T5 phage promoter and a number of chimeric and synthetic promoters. All of them were cloned as 68 bases (-49 to +19) inserted in front of the DHFR coding region. Relative levels of transcription were determined by dot blot using the vector inducible β-lactam degrading enzyme gene as internal standard.

In this example, if 28 promoters, each consisting of 68 nucleotides, are parameterized with three descriptors as defined in FIG. 14, the result is a 28 × 204 matrix (28 promoter × 68 nucleotide × 3 parameters (volume volume, hydrophobicity and The specific sequence of each promoter can be expressed as a single point in hyperspace in 204. Thus, these 28 promoters can be aggregated to form a 28 point cluster in the space. The experimental data obtained from are repeated and the level of transcription thus obtained is plotted against a 28 point cluster formed in 204 dimensional hyperspace using PLS, then half (14) of the promoters were used to establish a statistical model and the rest. Half were used to test it. To show that the correlation between the computed strength promoter for the promoter strength observed that good.

The model formed was extrapolated to construct two new promoters, both of which were significantly better than the best promoters present between the 28 initial promoters.

Multiple Mutation Assays and Protein Sequences

The same assay method can be applied to protein sequences. Signal peptides are characterized using multivariate analysis, which shows a good correlation between location in hyperspace and final physical space (Sjostrom et al. (1987) EMBO J. 6: 823-831). The main difference between nucleotides and amino acids, instead of qualitative descriptors of nucleotides (see FIG. 14), is that quantitative descriptors should be used to parameterize amino acids. Relevant properties of amino acids can be determined (such as volumetric capacity, hydrophobicity and polarity) and can be derived from the following literature (Hellberg et al. (1986) Acta Chem. Scand. B40: 135-140; Jonsson et al. (1989) Quant. Struct.-Act.Relat. 8: 204-209.

After shuffling and characterizing the shuffled proteins, mutated proteins in the "better" state and the "worse" state were analyzed relative to the initial sequences. Statistical methods (such as PLS) can be used to extrapolate new sequences that appear to be better than the best sequences present in the analyzed set.

Modeling Protein Sequence Spaces

As mentioned above, any protein encoded by a DNA sequence can be plotted as a clear point in multidimensional space using statistical methods. A "normal" 1 kb gene may, for example, make up about 330 amino acids. Each amino acid may be described, for example, by three main physicochemical descriptors (stereoscopic capacity, hydrophobicity and polarity) for each amino acid (other descriptors for the protein are largely dependent on these three main descriptors). Jonsson et al. (1989) Quant. Struct.-Act. Relat. See 8: 204-209. The 1 kb gene is thus modeled in 330 (number of amino acids) x 20 (amino acids that may be present at each position) x 3 (three major descriptors, as described above for each amino acid) = 19,800 dimensions. Due to the scalability of the sequence space, the number of shuffled sequences is used to validate sequence activity related predictions. The closer the surrounding sequences are in space (% similarity), the higher the probability of making predictions. In contrast, the more sequence space analyzed, the more accurate the prediction. This modeling strategy can be applied to any useful sequence.

Predictive values for plotting multiple shuffled sequences are described above. Two additional methods may be used. First, for example, PLS (second-order near-squared projection of potential structure) can be used to plot the number of enzymatic activities of chimeric offspring in a library. If sufficient data are available, one can construct a function that can extrapolate the outside of the experimental data to produce sequences in silico that are more active than the best set of controls. Second, all relevant sequences can be plotted and arbitrary sequences can be grouped with those with a given related activity. By using the matrix, the resulting genes can be grouped with those with novel relevant activity that can be screened directly by a suitable activity or a subset of sequenced clones.

The whole issue of this policy is the availability of sufficient relevant sequences generated through shuffling to provide useful information. An alternative to the method of shuffling sequences is to apply the modeling method to all useful sequences, such as the GenBank database and other public sources. Although this requires a lot of computation, current technology makes this approach useful. Mapping all useful sequences can indicate the desired sequence space region. Moreover, the information can be used as a filter applied to in silico shuffling to determine which hypothetical offspring are the preferred candidates for physical methods (eg, the synthesis and / or recombination described herein).

Shuffling of Branching Intermediates

The present invention provides for shuffling of "evolutionary intermediate". Evolutionary intermediates within the context of the present invention refer to artificial constructs which mediate the feature between two or more homologous sequences, for example when the sequences are grouped in an evolutionary dendogram.

Nucleic acids are often classified into evolutionary dendograms (ie, “trees”) that represent evolutionary branches and relevance. For example, a cladistic analysis is a sequence of organisms or genetic features (including nucleic acid or polypeptide sequences) based on the assumed common ancestor's source (branched genetic features or mediators of the organism). And a classification method in which a position is determined. Although divergence analysis differs (the distinction is often made between evolutionary taxographers who deal with the extent of the difference and evolutionary taxographers who simply determine branch points in evolutionary dendograms (orthodox branching methods); however, in the present invention , The relation tree is derived by one of the above methods, and is primarily concerned with branching a relation tree (ie, "dendogram") that represents a relatedness.

Divergence mediators or other evolutionary mediators can be determined by selecting nucleic acids that are mediators in the sequence between two or more existing nucleic acids. Although the sequence cannot exist in nature, the sequence exhibits sequences similar to those selected in nature, ie, intermediates of two or more sequences represent sequences similar to the hypothetical consensus ancestry of two or more existing nucleic acids. Therefore, evolutionary intermediates are one preferred shuffling substrate, which are termed "pseudo selected" sequences, which are active and more similar to randomly selected sequences.

One advantage of using evolutionary intermediates (or oligonucleotides corresponding to these sequences) as the substrate for shuffling is that they can exhibit significant sequence diversity in very few starting substrates (ie, starting as source A and B). A single intermediate “C” is at least part of A and B above). This improves the efficiency of the method and can simplify the process of oligonucleotide synthesis in gene reconstruction / recombination methods. In addition, searching the sequence database for evolutionary intermediates increases the chance of identifying relevant nucleic acids using standard search programs such as BLAST.

The mediating sequence can also be selected between two or more synthetic sequences that are not expressed in nature by simply starting from two synthetic sequences. Such synthetic sequences may include evolutionary intermediates, suggested gene sequences, or other desired sequences associated with the sequences. Such "artificial intermediates" are also useful to reduce the complexity of genetic reconstruction methods and to improve the ability to search evolutionary databases.

Thus, in one embodiment of the present invention, a string representing an evolutionary or artificial intermediate may be determined initially using alignment and sequence relation software and then synthesized using oligonucleotide reconstruction methods. Alternatively, the intermediate may form the basis for selecting oligonucleotides used in the gene reconstruction methods herein.

Some of the sections below describe how to conduct research using the secret Markov model and other methods.

Shuffle in Silico with the Secret Markov Model

The view of synthetic shuffling lies in the assumption that each amino acid present at the source is independently present and adds (or not adds) a function by itself at a given functional level. When shuffling using a method based on DNasI, recombination is performed as a 20-200 bp fragment during assembly, thereby allowing the selective pressure of a functional unit (gene or promoter or other biological material). This problem is eliminated because each amino acid evolves between any amino acids that co-evolve in a given direction. By capturing co-variances, which are generally present in the gene family and improve the quality of the resulting library, the variants, many biologically inactive progeny generated through wild type or conventional shuffling, are removed. Artificially generated progeny may be inactivated due to structural, regulatory or other sensitive mismatches.

One way to remove the undesirable co-variance progeny is to apply statistical behavior such as Hidden Model (HMM) on the source sequence. The constructed HMM matrix (such as, for example, FIG. 15) can capture complete variants in the herd as probabilities of all possible states (ie, amino acid combinations, deletions and insertions). The matrix obtained from the analyzed herds is not similar enough to be identified by the standard BLAST algorithm of any particular sequence, but similar enough to be identified by probing using a probability distribution pattern based on the original herd. Used to search a classified database for additional members of.

The HMM matrix shown in FIG. 15 embodies a cluster of eight aminosa peptides. At each position, the peptide may be a specific amino acid (one of the 20 amino acids present in)), an insertion (o), or a deletion (o). The probability for each occurrence depends on how often it occurs between pooled sources. Through the behavior done in such a way that all possible paths are assigned to probability factors, any given source can be 'threaded'.

HMM can also be used in other ways. Instead of applying the generated behavior to identifying previously unidentified members of the herd, the HMM behavior can be used as a template for synthesizing newly members of the herd (eg, from the cladistic tree of nucleic acid). Intermediate member). For example, the HMMER program is useful (http://hmmer.wustl.edu/). The program builds HMM behavior on a finite set of members of the herd. The subprogram HMMEMIT reads the newly constructed sequence based on this behavior and this. Although the original purpose of HMMEMIT was to synthesize a positive control for the search pattern, the program can be applied to the present invention by using the output as a generated progeny in silico of HMM behavior that limits shuffling. According to the present invention, oligonucleotides corresponding to the nucleic acids are synthesized for recombination, gene reconstitution and screening.

Since the sequence relationship for each position has been described in the form of probabilities, the number of inactive offspring is significantly lower than in the shuffling reaction that simply selects these offspring. Crossovers between naturally occurring (ie, from source) genetic regulatory phenomena (structural, functional, or non-limiting) and co-evolution of point mutations or structural elements are maintained through this shuffling process.

Standard algorithm to synthesize sequence intermediates obtained from sequence alignments

The following outlines the program for synthesizing sequence intermediates obtained from the alignment of relevant nucleic acids.

Given the alignment of the sequence encoding the source and the alignment encoding the offspring

In each descendant sequence

In each source sequence

In each window

If the source sequence and progeny sequence match the window

If the window is not yet covered by the sequence in the intercept list

Amplify the windows, i.e. 5 'and 3', until too many mismatches are generated,

Add the finally amplified fragment to the list of fragments for the sequence,

In each descendant sequence

Set the position at the beginning of the sequence

Until the end of the progeny sequence is carried out as follows.

Search through intercepts to find the longest amplified section (most similar to the root section) from the point before the corresponding position

If either is added to the optimal path list, the position is set at the end of the retrieved intercept

If either does not appear to increase at the current position, display the intercept from the optimal path list.

Normalization of the Library—Use of Positive or Negative Activity Data

One aspect of the present invention is the use of positive or negative data in the method of sequence design and selection in the silico or at the stage of a physical method, or in both cases. Using positive or negative data may be included in the content of the neural network's teaching methods, or may provide a logical or physical filter by simply using positive or negative data in the design or library synthesis process. Network teaching can be described as above, which provides a convenient way of using positive or negative data that increases the chance that the resulting additional sequences will have the desired activity. Since screening often acts as a limited step in synthesizing improved genes and proteins by a forced evolution method, the ability to use negative data to reduce the screened library size provides significant advantages. Similarly, using positive data to bias the library with the desired sequence is another way to enhance the library.

For example, as mentioned, in addition to neural network teaching studies, positive or negative data can be used to provide a physical or logical “filter” for any system of interest. That is, sequences that are shown to be inactive provide useful information when closely related sequences, specifically active sequences, are also identified that will be found to be inactive. Similarly, active sequences provide useful information for when closely related sequences, particularly inactive sequences, are also identified in which closely related sequences will also prove to be active. Such active or inactive sequences can be used to provide a virtual or physical filter to bias the library to produce more active members.

For example, when using negative data, the physical subtraction method may hybridize with an inactive member under selected stringency conditions (number of libraries generated by the method herein including high stringency, homology members). Similar nucleic acids are removed from the synthesized library. Similarly, hybridization rules or other parameters are used to select for members that are considered similar to inactive sequences. For example, oligonucleotides used in gene reconstruction methods can be biased against sequences that appear to be inactive. Therefore, in any method, a library or string can be filtered by subtracting a set of strings having an initial library member of a biological polymer that is below the library or the desired activity.

When using positive data, physical amplification methods are similar, including library members synthesized by hybridization with active members under selected stringency conditions (number of libraries produced by the methods herein, including high stringency, homology members). Nucleic acids are separated. Similarly, hybridization rules or other parameters are used to select for members that are considered similar to the active sequence. For example, oligonucleotides used in gene reconstruction methods can be biased against sequences that appear to be active. Therefore, in any method, the library or string can be filtered by subtracting the set of strings with the initial library member of the biological polymer above the library or the desired activity.

Similarly, in silico assays can be used to prepare inactive sequence libraries rather than active sequences. In other words, inactive sequences can be shuffled in silico to produce clone libraries with low activity. Such inactive sequences can be physically synthesized and used to subtract the synthesized library by other methods (typically by hybridizing to library members). By subtracting this way, that is, by removing members that appear to be primarily inactive, the size of the library to be screened can be reduced.

Example--motif filtering

Selection or screening methods can often produce excessive "positive" clones for the cost-saving sequences of both positive or negative clones. However, if amplified or reduced sequence motifs are identified in positive or negative clones, these biases are used in constructing synthetic libraries that are biased with "good" motifs and biased with "baad" motifs.

If each consecutively selected region or motif (eg, if the window selected is a 20 base region) is considered to be an independent gene or genetic factor, gene generation before or after the selection or screening method. The change in frequency can be measured. Motif with increased incidence in positive clones is characterized as "good" and motif with reduced incidence in positive clones is characterized as "poor". The second generation library is selected and synthesized to amplify for good motifs and reduce to poor ones using any filtering or teaching method as presented herein.

Many methods for measuring the frequency of occurrence of motifs in a community of genes are useful. For example, one method uses a gene chip provided by Affymetrix (Santa Clara, Calif.), Other gene chip manufacturers, for example, a gene chip having an object motif encoded in a spatially addressable manner. The assay sequences can be hybridized to gene chips or other nucleic acid sequences. Similarly, hybridization to membranes that contain spatially addressable motifs and measure signal strength relative to the probe before or after the selection process is also essentially similar, such as using standard Southern blots or Northern blots. It can be done in a manner. The relative proportions of desired / desirable features identified on the chip also provide an indication of the overall library quality. Similarly, phage display or other expression libraries can be used to evaluate expression products to assess library properties.

Alternatively, real-time quantitative PCR (eg, TaqMan) can be performed to be very discriminating with respect to the properties intended by the PCR oligomers. This can be done, for example, by having a polymorphism specific to the motif present near or at the 3 'end of the oligonucleotide to effectively prime the PCR if it matches perfectly. Real-time PCR product analysis (and related real-time reverse transcription PCR) by FRET or TaqMan, for example, is a group of methods of real-time PCR monitors, a well-known technique used in numerous literature [Laurendeau et al. (1999) "TaqMan PCR-based gene dosage assay for predictive testing in individuals from a cancer family with INK4 locus haploinsufficiency "Clin Chem 45 (7): 982-6; Laurendeau et al. (1999) "Quantitation of MYC gene expression in sporadic breast tumors with a real-time reverse transcription-PCR assay" Clin Chem 59 (12): 2759-65; And Kreuzer et al. (1999) “LightCycler technology for the quantitation of bcr / abl fusion transcripts” Cancer Research 59 (13): 3171-4].

If the gene cluster of interest begins very similarly (eg, 90% sequence identity), one can simply sequence the gene groups before or after selection. If sequencing primers for several up and down genes are used, the sequences can appear in parallel on the sequencing gel. Sequence polymorphism in the vicinity of the primer can be read to indicate the relative proportion of bases at any given position. For example, if the population begins with 50% T and 50% C at a given location, and after selection begins with 90% T and 10% C, then the sequencing is initiated adjacent to the polymorphism to The ratio can be easily quantified. The method is limited because additional polymorphic parts are obtained from the primers to decipher the more polymorphic parts, and the mobility of the various sequences becomes more and more diversified and the translation proceeds together. However, as the cost of sequencing and oligomers continues to decrease, one solution to this is simply to sequence the top and bottom of the gene into a number of different oligomers.

Example: Fraction Extraction of Sequence Space

Conventional sequence spaces are very large relative to the number of sequences that can be physically cloned and characterized. The computational method describes a subset of the sequence space that is expected to amplify clones with the desired properties (homology). However, there are assumptions and computational constraints inherent in the model. Methods of fractionating the sequence space to be enriched for the molecule expected by a given model to be more or less suitable for the phenotype of interest would be useful for testing this predictive model.

A simple embodiment of how to do this is as follows. There are about 10 ²⁷ possible shuffled IFNs (common proteins in terms of size) based on naturally occurring human IFN gene clusters. This is larger than the number that can be easily screened. If the aim is to evolve shuffled active human IFNs on mouse cells, the literature shows 121 and 125 residues from human IFN-α1 that confer improved performance when transplanted onto other human IFNs such as IFN-α2α. Information can be used. Assuming that the motifs will impart improved activity in a number of different contexts, a large pool of shuffled IFN genes (typically around 10 ⁹ -10 ¹² ) are prepared and converted to ssDNA to those residues. Pass them through affinity column chromatography containing oligonucleotides complementary to human IFN-α1, wash under moderately stringent conditions to elute bound molecules, PCR amplify the eluted genes and clone the materials. Functional tests can be performed on expressed clones. In this way it is possible to physically bias the shuffled gene library to include expected motifs by a very simple model that improves the intended activity.

Ideally, one can take clusters that amplify the motif and the population that reduced the number of motifs. Both clusters are analyzed (1000 clones from each cluster). This hypothesis is "tested" by querying whether such sequence space fractions bias the average suitability in the expected manner. If so, the hypothesis is "accepted" to increase the scale of library screening. It is also possible to test a number of design algorithms as affinity-based fractions, and to accept the facts supported by these experiments and to continuously perform affinity selection to amplify clones that are design samples of the multiplicity algorithm.

In this model, shuffling such as bunch shuffling is used as the first order design algorithm. However, an additional design algorithm is an integrated downstream of shuffling that further refines the sequence space based on simple design lessons. The method can be performed at the level of nucleic acid with any design algorithm that can be translated into a nucleic acid selection scheme.

Many variations on such embodiments are useful for reducing the size of libraries produced by physical or virtual filtering methods.

For example, affinity selection of oligonucleotides encoding the motif of interest after genetic recombination / recombination (physical or in silico method) results in a population of nucleic acids synthesized by the genetic recombination / resynthesis method as mentioned herein. Reduces the variety.

Similarly, oligonucleotides encoding motifs may be selected, for example, by genetic recombination / resynthesis, by enzymatic digestion molecules that do not perfectly match the oligomer. Alternatively, genes incompletely matching oligonucleotides can be selected, for example, by binding to mutS or other DNA mismatch repair proteins.

Polymerization reactions in recombinant / genetic synthesis methods can be primed using one or more oligomers encoding the motif of interest. That is, if a mismatch occurs at or near the 3 'end of the hybridized nucleic acid, elongation is reduced or blocked. In this variation, only newly polymerized molecules will survive (used in subsequent library construction / selection steps). Such a method can be performed, for example, by priming reverse transcription of RNA and then digesting the RNA.

Another approach is to make the template specifically degraded. For example, DNA to which uracil is bound with high frequency can be synthesized. Polymerase based synthesis is primed with oligomers and amplified with dNTPs that do not contain uracil. The resulting composite is treated with uracil glycosylase and a nuclease that cleaves sites free of purine bases. Similarly, RNA nucleotides can be incorporated into the DNA chain (either synthetically or by enzymatic coalescence); These nucleotides can then be used as targets cleaved by RNA endonucleases. Many other cleavable residues are known, including any residues that are targets for enzymes or other residues and are used as cleavage portions that react to light, inferiority, and the like. If polymerases having the activity of incorporating the intended cleavage target are not practically useful, these polymerases can be prepared using shuffling methods that modify existing polymerase activity or impart new polymerase activity. have.

Localized motifs can be easily translated by affinity selection methods. However, it is often desirable to assign rules to a number of sequence features (eg, sequence features 2,3,4,5,6, etc.) in which molecules are separated in intragenic space. In this case it may be defined to select by preparing a nucleic acid template comprising all of the desired motifs separated by a flexible linker. The Tm of a molecule with all motifs can be higher than molecules with only one or two of the motifs. Thus, it is possible to amplify molecules with all motifs by selecting molecules with high Tm for the selection oligomer.

Many of these strongly bound motifs "genes" can be isolated and synthesized by flexible linkers or bases such as inosine that is accidentally formed with base pairs. Then, when selecting a nucleic acid, a gene having a high Tm may be selected. By carefully designing the selected nucleic acid template, one can amplify a gene with multiple sequence motifs that is predicted to bias genes containing multiple sequence motifs to have a desired phenotype.

The technique can also be used if there is little information on whether any given motif is expected to bias the library preferably. Motif sets can be defined or randomly selected motifs, for example, based on sequence conservation among different homologues. If the sequence space is not isotopes (equally aggregated with good members in all respects), simply fractionate the sequence space based on a set of designed or random motifs, and identify the clones in the region of the desired sequence space. Average suitability can be measured to give more weight to the most suitable zone.

In addition to simple sequence alignment methods, there are more complex research methods that are useful for identifying regions of interest, such as large molecule binding sites. For example, US Pat. No. 5,867,402, filed by Schneider et al. (1999), in " COMPUTATIONAL ANALYSIS OF NUCLEIC ACID INFORMATION DEFINES BINDING SITES, " Here's how. Substitutions within binding site sequences can be analyzed to determine whether such substitutions cause deleterious mutations or positive polymorphisms. A method of identifying new binding locations using independent amounts of information has also been proposed. Such a research method is also used in the present invention as one method for identifying a target sequence in in silico sequence manipulation.

MOTIF BREEDING

Theoretical design can be used to produce the desired motif in the sequence or in the desired sequence space. However, it is often difficult to predict whether a given designed motif will be expressed as a functional form, or whether its presence will affect other intended properties. One example of this is the design of glycosylation sites in proteins to allow them to access cellular glycosylation mechanisms and to block other proteins from binding by steric inhibition by attached polysaccharide groups. This is a way to avoid adversely affecting.

One way to address these issues is to design a motif or multiple variations of the motif at multiple candidate sites in the target gene. The sequence space is then screened for the desired phenotype and selected. Molecules that exhibit specified design criteria can be shuffled together repeatedly to optimize the desired properties.

Motifs can be constructed in any gene. Standard protein motifs include the following: N-linked glycosylation sites (ie Asn-X-Ser), o-linked glycosylation sites (ie Ser or Thr), protease sensitive sites (ie Cleavage after X position in PXGP with collagenase), Rho-dependent transcription termination position for bacteria, RNA secondary structural factors, transcription enhancing factors, transcriptional promoter factors, transcriptional suppression motifs, etc. that affect translation efficiency.

High throughput theoretical design

In addition to or in conjunction with the theoretical design approach described above, high throughput theoretical design methods are also used. In particular, high throughput theoretical design methods can be used, for example, to modify any given sequence in silico prior to performing the recombination / synthesis method. For example, Protein Design Automation (PDA) is one of the computer operating systems that designs and optimizes proteins and peptides and designs proteins and peptides.

PDAs typically design amino acids to begin with the protein backbone structure and modify the protein properties while maintaining the three-dimensional folding properties of the protein. Multiple sequences can be manipulated using a PDA that can design the structure of a protein well (sequences, contiguous sequence sites, etc.). PDAs are described, for example, by Malakauskas and Mayo (1998) "Design, Structure and Stability of Hyperthermophilic Protein Variant" Nature Struc. Biol. 5: 470; Dahiyat and Mayo (1997) "De Novo Protein Design: Fully Automated Sequence Selection" Science, 278, 82-87; DeGrado (1997) "Proteins from Scratch" Science, 278: 80-81; Dahiyot, Sarisky and Mayo (1997) "De Novo Protein Design: Towards Fully Automated Sequence Selection" J. Mol. Biol. 273: 789-796; By Dahiyot and Mayo (1997) "Probing the Role of Packing Specificity in Protein Design" Proc. Natl. Acad. Sci. USA 94: 10172-10177; By Hellinga (1997) "Rational Protein Design-Combining Theory and Experiment" Proc. Natl. Acad. Sci. USA 94: 10015-10017; By Su and Mayo (1997) "Coupling Backbone Flexibility and Amino Acid Sequence Selection in Protein Design" Prot. Sci. 6: 1701-1707; By Dahiyat, Gordon and Mayo (1997) "Automated Design of the Surface Positions of Protein Helices" Prot. Sci. 6: 1333-1337; By Dahiyat and Mayo (1996) "Protein Design Automation" Prot. Sci. It is described in numerous publications, including 5: 895-903. Further description of the PDA can be obtained, for example, at http://www.xencor.com/.

In the description of the present invention, PDAs and other design methods can be used to modify intra-silico sequences that can be synthesized / recombinated in the shuffling methods presented herein. Similarly, PDAs and other design methods can be used to engineer nucleic acid sequences obtained from the following selection methods. Therefore, the design method can be used repeatedly during an iterative shuffling process.

Shuffling Methods in Biooligonucleotide-dependent Silicos

As discussed herein, many of the methods of the present invention include following the induction of diversity in the sequence column in silico, followed by performing oligonucleotide gene recombination / synthesis methods. However, recombination methods based on bioligonucleotides are also suitable. For example, instead of synthesizing oligonucleotides, genes corresponding to any of the diversity induced in silico may be prepared without the use of oligonucleotide intermediates.

This is particularly useful when the gene is short enough to induce synthesis.

In addition, peptide sequences can be synthesized directly from various string populations rather than through oligonucleotides. For example, solid phase polypeptide synthesis can be performed. For example, the solid phase peptide sequence can be constructed by standard solid phase peptide synthesis as a member of sequences selected to correspond to the sequence sequence synthesized in silico.

In this regard, solid phase synthesis of biological polymers comprising peptides is described, for example, in J. Am. Chem. Soc. As described in 85: 2149-2154 (1963), this may be done even by the initial “Merrifield” solid phase peptide synthesis method. Solid phase synthesis techniques are useful, for example, for synthesizing some peptide sequences on a number of “pins”. Geyser et al. (1987) J. Immun. Meth. See 102: 259-274. Other solid phase techniques include, for example, synthesizing various peptide sequences on different cellulose disks supported on a column. See Tetrahedron 44: 6031-6040, by Frank and Doring (1988). Other solid phase techniques are described in US Pat. No. 4,728,502 and WO 90/00626 published in Hamill. Methods of forming large arrays of peptides are also useful. For example, Pirung et al., U. S. Patent No. 5,143, 854 and Fodor and many, PCT Publication No. WO 92/10092 disclose, for example, the arrangement of peptides and other polymer sequences using light-directed synthesis techniques. It is disclosed about how to form. Stewart and Young, Solid Phase Peptide Synthesis, 2d.ed., Pierce Chemical Co. (1984); Col. Artherton et al. (1989) Solid Phase Peptide Synthesis, IRL Press, Greene et al. (1991) Protective Groups In Organic Chemistry, 2nd ed., John Wiley & Sons, New York, NY and Bodanzszyky (1993) Principles of Peptide Synthesis 2nd ed., Springer Verlag, Inc. See N.Y. Other useful information about proteins can be found in R. Scopes, Protein Purification, Springer-Verlag, N.Y. (1982); Deutscher, Methods in Enzymology Vol. 182: Guide to Protein Purification, Academic Press, Inc. N.Y. (1990); Sandana (1997) Bioseparation of Proteins, Academic Press, INC .; Bollag et al. (1996) Protein Methods, 2nd ed., Wiley-Liss, NY; Walker (1996) The Protein Protocols Handbook Humana Press, NJ, Harris and Angal (1990) Protein Purification Applications: A Practical Approach IRL Press at Oxford, Oxford, England; Harris and Angal Protein Purification Methods: A Practical Approach IRL Press at Oxford, Oxford, England; Scopes (1993) Protein Purification: Principles and Practice 3rd Edition Springer Verlag, NY; Jason and Ryden (1998) Protein Purification: Principles, High Resolution Methods and Applications, Second Edition Wiley-VCH, NY; And Walker (1998) Protein Protocols on CD-ROM Humana Press, NJ; And the references cited in these documents.

In addition to proteins and nucleic acids, it will be appreciated that the string generated in silico may correspond to other biopolymers. For example, strings may correspond to peptide nucleic acids (PNAs) that can be synthesized according to useful techniques and screened for activity in any suitable assay. See, for example, Nielsen and Michael Egholm (eds) (1999) Peptide Nucleic Acids: Protocols and Applications ISBN 1-898486-16-6 Horizon Scientific Press, Wymondham, Norfolk, U.K. See for an introduction to PNA synthesis and activity screening.

Assay--Physical Choice

As with DNA shuffling, or traditional chain amplification or any functional genetics technique, GAGGS-induced evolution can use any physical assay known in the art to detect polynucleotides encoding desired phenotypes. have.

Synthetic genes are easy to apply traditional cloning and expression studies; Therefore, the properties of these genes and the proteins they encode can be readily observed after they are expressed in host cells. Synthetic genes can also be used to synthesize polypeptide products by in vitro transcription and translation. Thus, polynucleotides and polypeptides are observed for their many predetermined ligands, small molecules and ions, or polymers and heteropolymeric substrates comprising other proteins and polypeptides, and their ability to bind microbial cell walls, viral particles, surfaces and membranes. Can be.

For example, a number of physical methods can be used to detect polynucleotides that encode phenotypes directly by either polypeptide or associated with the catalysis of a chemical reaction by an encoded polypeptide. For the purpose of explanation and by the nature of the particular chemical reaction intended for the purpose, these methods are characterized by differences between the substrate and the product or changes in the reaction medium associated with the chemical reaction (e.g. And a variety of techniques well known in the art for explaining the change in fluorescence for visible or far infrared (heat). These methods may also be selected from any combination of the following: mass spectroscopy; Nuclear magnetic resonance method; Fractions and spectroscopy illustrating the formation of isotopically labeled substances, isotope distribution or labeled products; Spectroscopic and chemical methods for detecting changes in the ionic or elemental composition of the reaction product (changes in pH, changes in inorganic and organic ions, etc.). Other methods of physical assays suitable for use in GAGGS include reaction products, including products with antibodies with reporter properties or products based on in vivo affinity recognition coupled with expression and activity of reporter genes. It may be based on using specific biometric sensors. Enzyme coupled assays for in vivo reaction product detection and cell survival-kill-growth selection may also be used where appropriate. Regardless of the specific nature of the physical assay, all of them are used to select the desired property, or combination of preferred properties, encoded by the GAGGS synthetic polynucleotide. Therefore, polynucleotides that are shown to have desirable properties are selected from the library.

The methods of the present invention optionally comprise a selection and / or screening step of selecting nucleic acids having the desired characteristics. The appropriate assay used for the selection will depend on the method of application. Many assays are known, such as proteins, receptors, ligands and the like. These methods include binding to modified components, cell or organism viability, generation of reporter compositions, and the like.

In high throughput assays, thousands of different shuffled variants can be screened per day. For example, each well of a microtiter plate may be used to perform an independent assay, or if an enrichment or incubation time effect is observed, every 5-10 wells may be a single variant (e.g., At concentrations). Therefore, a single standard microtiter plate can be assayed for about 100 (eg 96) reactions. If 1536 well plates are used, a single plate can easily be assayed from about 100 or about 1500 different reactions. Several different plates can be assayed per day; Assays with up to about 6,000-20,000 different assays (eg, including different nucleic acids, encoded proteins, enrichments) can be performed using the integrated system of the present invention. More recently, microfluidic approaches to reagent manipulation have been developed, for example, by Caliper Technologies (Moutain View, Calif.), Which can provide very high throughput microflow assays.

In one aspect of the invention, cellular viral plaques, spores, etc. comprising GAGGS shuffled nucleic acids are separated on solid medium to produce respective colonies (or plaques). Using an automated colony collector (eg Q-bot, Genetix, UK), a colony or plaque was identified and collected and up to 10,000 different mutants containing 96 well microtiter dishes containing two 3 mm glass balls per well. Is inoculated. The Q-bot does not collect the entire colony but inserts a pin in the center of the colony to extract a small amount of cells (or mycelia) and spores (or plaque virus) samples. The time the pin stays in the colony, the number of times it is taken to inoculate the culture medium, the time the pin stays in the medium to perform the inoculation, and each parameter can be adjusted and optimized.

The same method of automated colony collection, such as Q-bot, reduces human error and increases the rate of successful culture production (about 10,000 / 4 hours). The culture is optionally stirred under temperature and humidity controlled by an incubator. The glass balls in the fine titration plates promote the aeration of cells and the dispersion of cell fragments (eg mycelia) uniformly, similar to the blades of fermenters. Clones from the desired culture can be separated by limiting dilution. As mentioned above, plaques or cells comprising the library can also be screened directly for protein production by detecting hybridization, protein activity, protein binding to antibodies, and the like. To increase the chance of identifying pools of sufficient size, a preliminary screening method can be used that increases the number of mutants by 10 times. The purpose of this preliminary screening is to quickly identify mutants with as much production titration as the source chain or more than the source chain and to transport these mutants to cell culture for use in subsequent assays.

One approach to screening multiple libraries is to use a massively parallel solid-phase procedure for screening cells expressing shuffled nucleic acids encoding enzymes with enhanced activity. Large parallel solid phase screening devices using adsorption, fluorescence or FRET are useful. See, eg, Bylina et al. (1999) US Pat. No. 5,914,245; http://www.kairos-scientific.com/; Youvan et al. (1999) "Fluorescence Imaging Micro-Spectrophotometer (FIMS)" Biotechnology et alia <www.et-al.com> 1: 1-16; Yang et al. (1998) "High Resolution Imaging Microscope (HIRIM)" Biotechnology et alia <www.et-al.com> 4: 1-20; And Youvan et al. (1999) "Calibration of Fluorescence Resonance Energy Transfer in Microscopy Using Genetically Engineered GFP Derivatives on Nickel Chelating Beads" www.kairos-scientific.com. After screening by the above techniques, the desired sequence is typically isolated and optionally sequenced, and the sequences as set forth herein are used to design new sequences in silico or in other shuffling methods.

Similarly, a number of robotic systems for chemicals in solution are also being developed that are useful for well known assay systems. The system uses a robotic arm and an automated workstation, such as an automated synthesis device developed by Takeda Chemical Industries, LTD. (Osaka, Japan), and a number of robotic systems (Zymate II) that mimic the passive synthesis operations performed by scientists. , Zymark Corporation, Hopkinton, Mass .; Orca, Beckman Coulter, Inc. (Fullerton, Calif.). Any of these devices is suitable for use in the present invention, such as, for example, high throughput screening of molecules encoded by codon modified nucleic acids. The nature and manner of modification to these devices operable as mentioned herein will be apparent to those skilled in the art.

High throughput screening systems are commercially available [eg, Zymark Corp., Hopkinton, MA; Air Technical Industries, Mentor, OH; Beckman Instruments, Inc. Fullerton, CA; See Precision Systems, Inc., Natick, MA et al.]. The system is typically automated throughout the process, including all sample and reagent pipetting, liquid dispersion, incubation time adjustment and final readout of the detector microplates. The definable system provides high throughput and fast start-up and a high degree of flexibility and customization.

The manufacturers of these systems provide detailed procedures for various high throughputs. Thus, for example, Zymark Corp. provides a technical report describing screening systems that detect the regulation of gene transcription, ligand binding, and the like.

Many commercially available peripherals and software are, for example, PCs (machines based on Intel × 86 or Pentium chip compatible DOS ^™ , OS2 ^™ WINDOWS ^™ , WINDOWS NT ^™ or WINDOWS95 ^™ ), MACINTOSH ^™ , or UNIX. It is useful for digitizing, storing and analyzing digitized video or digitized optical or other calibration images using a computer based (eg, SUN ^™ workstation).

Integrated systems for analysis typically include a digital computer with GO software for GAGGS, optionally with high throughput liquid control software, image analysis software, data decoding software, and a solution operatively connected to the digital computer from a source. Robotic liquid control armature to deliver to destination, high throughput liquid transfer by an input device (eg computer keyboard) or robotic liquid control arm to introduce data into digital computer to control GAGGS operation and optionally labeled An image scanner that digitizes the label signal from the assay component. The image scanner adjusts the image analysis software to measure the intensity of the probe markers. Typically the probe label intensity is interpreted by data interpretation software to indicate whether the labeled probe hybridizes with DNA on a solid support.

Currently, computer hardware support in the industry is sufficient for use with GAGGS [any medium or low cost Unix system (eg Sun Microsystems) or a more expensive Macintosh or PC will be sufficient]. Current software technology is sufficient to design upgradeable open-architecture object-oriented genetic algorithm packages that are specific to GAGGS users with biological basic knowledge.

Digital device for GO

As mentioned herein, various methods and genetic algorithms (GOs) can be used to perform the desired functions. Moreover, a digital or analog system, such as a digital or analog computer system, can control other functions such as displaying and / or controlling output files.

For example, standard desktop applications such as word processing software (eg, Microsoft Word ^TM or Corel WordPerfect ^TM) and database software (eg, Microsoft Excel ^TM, spreadsheet software such as Corel Quattro Pro ^TM, or Microsoft Access. ^A database program such as ^TM or Paradox ^™ may be applied to the present invention by inputting one or more character strings into software loaded into the memory of the digital system to perform GO as mentioned herein. For example, the system may be a user interface that manipulates strings (for example, a GUI of a standard operating system such as a Windows, Macintosh, or LINUX system), a GO programmed into the application, or a GO performed manually by the user. It may include the software having the appropriate string information used with. As mentioned, specific sorting programs such as PILEUP and BLAST to align nucleic acids or proteins (or corresponding strings) as a preliminary step to perform further GO on the aligned sequences are also incorporated into the system of the present invention. Can be. Software for performing PCA may also be included in the digital system.

GO operating systems are typically incorporated into software systems comprising, for example, digital software and engineered sequences equipped with GO software for alignment and manipulation of sequences according to the GO mentioned herein, or GO software for performing PCA. Contains a data set. The computer is, for example, a PC (Intel × 86 or Pentium chip compatible DOS ^TM , OS2 ^TM , WINDOWS ^TM , WINDOWS NT ^TM , WINDOWS 95 ^TM , WINDOWS 98 ^TM , LINUX, Apple Compatibility, MACINTOSHTM Compatibility, Power PC Compatibility, or UNIX compatible (eg, SUN ^™ workstation) devices or other commercially available common computer systems known to those skilled in the art. Sequence alignment or manipulation software can be constructed by one skilled in the art using standard programming languages such as Visualbasic, Fortran, Basic, Java, etc. in accordance with the methods herein.

Any controller or computer optionally includes such things as, for example, a bipolar tube ("CRT") display, a horizontal panel display (eg, an active matrix liquid crystal display, a liquid crystal display), and the like. Computer circuitry is often placed in a box containing a number of integrated circuit chips such as microprocessors, memory, interface circuits, and the like. The box may also optionally include a hard disk drive, a floppy disk drive, a high capacity removable drive such as an information retractable CD-ROM, and other common peripheral elements. Input devices such as keyboards or mice optionally allow the user to provide input or to select a sequence to be compared or manipulated in the relevant computer system.

Typically a computer system includes suitable software that accepts a user's instructions, where the software is in the form of instructions programmed by the user for example in a set parameter field in the GUI or for a number of different specific operations. Included as. The software then translates these instructions into a language suitable for operating the system to perform any desired operation. For example, in addition to performing GO manipulation of strings, the digital system allows the oligonucleotide synthesizer to synthesize oligonucleotides for gene reconstruction or even to align oligonucleotides of commercially available sources (eg, in the proper order form). By printing or linking the order form on the Internet).

The digital system may also include an output element that controls the synthesis of the nucleic acid (eg, based on the sequence or sequence alignment herein), ie the integrated system of the present invention is optionally an oligonucleotide synthesizer or Nucleotide synthesis controllers. The system can include other operations that occur along the flow from alignment or other operations performed using strings corresponding to sequences herein, for example, as mentioned above as assays.

In one embodiment, the GO of the present invention is embodied as a logical indication and / or data modification medium or transferable program component when a computing device suitably configured for one or more strings is loaded to perform GO. 13 is a standard digital device that should be understood to be a logical device capable of reading instructions generated from medium 717, network port 719, user input keyboard 709, user input 711, or other input means. (700). The apparatus 700 can then use these instructions to induce one or more strings to be GO modified, for example by constructing one or more data sets (including a plurality of GO modification sequences corresponding to nucleic acids or proteins, for example). Can be. One form of logic device that may embody the present invention is to display a CPU 707, an optional user input device keyboard 709 and a GUI specific device 711 and a disk drive 715 and a monitor 705 (GO variant strings). And a computer system 700 including peripheral components such as by the user to provide a simplified selection of a subset of strings.

Modification medium 717 is optionally used to program the entire system and may include, for example, optical or magnetic media or other electronic memory storage elements in the form of disks. The communication port 719 can be used to program the system and can represent any form of communication connection state.

The invention may also be embodied in circuits of specifically integrated circuit (ASIC) applications or programmed logic devices. In such cases, the invention may be embodied in a computer readable descriptor language that may be created as an ASIC or PLD.

The invention may also be embodied in circuits or logic processors in many other digital devices such as PDAs, laptop computer systems, displays, image editing devices, and the like.

In one preferred embodiment, the digital system computer comprises learning of components in which the results of the physical oligonucleotide assembly scheme (composition, product abundance, different processes) are monitored and correlated in association with physical assays. do. Successful and unsuccessful combinations have been validated in the database to include parameters for subsequent GAGGS processes that include the same set of source strings / nucleic acids / proteins (or even irrelevant sequences if the information provides method improvement information). Provide coordination / priority for the user base or digital system base selection of the set. The correlation is used to modify subsequent GAGGS processes to optimize the process. The cycle of physical synthesis, selection, and correlation is randomly repeated to optimize the system. For example, you can learn neural networks to optimize your results.

Materialization on the Web Site

The method of the present invention may be performed in a zoned or distributed computing environment. In a distributed environment, the method may be performed on a single computer including multiple processors or multiple computers. The computer may for example be connected via a general bus, more preferably the computer may be a node on a network. The network may be a generalized or deductive logic or wide area network, and in some embodiments the computer may be a component of an intranet or the Internet.

In one Internet embodiment, the client system is typically coupled to a server computer running a web browser to run a web server. The web browser is typically a program such as IBM Web Explorer, Internet Explorer, NetScape or Mosaic. The web server is typically an IBM HTTP Daemon or other WWW Daemon (e.g., a LINUX-based version of the program) but need not be. The client computer is interactively coupled with the server computer via a line or silent system. In contrast, the server computer is interactively coupled to a website (server hosting the website) that is accessed by software that performs the method of the present invention.

A user of a client connected to an intranet or the Internet can cause the client to request a resource that is part of a website that hosts an application that performs the method of the present invention. The server program then processes to return the specified resources (assuming they can now be used). Standard naming conventions have been adopted and are called Uniform Resource Locators ("URLs"). The convention includes some form of zone name, which currently includes subclasses such as Hypertext Transport Protocol ("http"), File Transport Protocol ("ftp"), gopher, and Wide Area Information Service ("WAIS"). Include. When a resource is downloaded, it may include a URL that is an additional resource. Therefore, users of the client can easily determine the existence of the new resource they specifically requested.

The software that performs the method of the present invention can be run locally on a server that hosts a website in a true client server architecture. Therefore, the client computer post requests the host server to regionally operate the requested process and download the results back to the client. Alternatively, the method of the present invention may be performed in a "multi-tier" manner in which the components of the method are regionally performed by the client. This may be performed by software (eg, Java application) downloaded from the server on request by the client or by software that is "permanently" installed on the client.

One embodiment of an application that performs the method of the present invention is divided into frames. In this paradigm, it is useful to see an application as an abstract frame or set of perspectives rather than as a set of features or functionality. For example, a typical application typically includes a set of menu items, each of which requests a particular frame--that is, clarifies any function of the application. In this respect, an application is considered as a collection of applets rather than as a single piece of cryptography or as a bundle of functions. In this way within the browser, the user selects a web page connection that requests a particular frame of application (ie, a sub-application). Thus, for example, one or more frames provide the ability to enter and / or encode biological molecules into one or more strings, provided that another frame provides a tool for generating and / or increasing the diversity of an encrypted string.

In certain preferred embodiments, the method of the present invention is performed as one or more frames, for example providing the following functions.

Encoding at least two biological molecules into a string providing a collection of at least two different initial strings, each subunit comprising a selected subunit; Selecting at least two substrings from the string; Concatenating the substrings to form one or more generated strings that are approximately the same length as one or more initial strings; Adding the generated string (place) to a string collection; And the ability to implement any feature of GAGGS or any GO or GA presented herein.

The ability to encode two or more biological molecules can provide one or more widows from which a user can insert an indication of the biological molecule. In addition, the encryption function also provides access to private and / or public databases, optionally accessible through local networks and / or intranets, so that one or more sequences contained in the databases can be entered into the method of the present invention. do. Thus, for example, in one embodiment, if the end user introduces the sequenced nucleic acid sequence into a coding function, the user may optionally request a search from GenBank and / or encode one or more sequences returned to this search method. Or may be introduced into a mutation generating function.

Methods of implementing intranet and / or intranet refinement of computational and / or data access methods are well known to those skilled in the art and are described in great detail in the following literature [eg, Cluer et al. (1992) A General Framework for the Optimization of Object-Oriented Queries, Proc SIGMOID International Conference on Management of Data, San Diego, California, Jun. 2-5, 1992, Sigmoid International Conference on Management of Data, San Diego, California, Jun.2- 5, 1992, SIGMOID Record, vol. 21, Issue 2, Jun., 1992; Stonebraker, M., Editor; ACM Press, pp. 383-392; ISO-ANSI, Working Draft, "Information Technology-Database Language SQL", Jim Melton, Editor, International Organization for Standardization and American National Standards Institute, Jul. 1992; Microsoft Corporation, "ODBC 2.0 Programmer's Reference and SDK Guide.The Microsoft Open Database for Microsoft Windows.TM. And Windows NT.TM., Microsoft Open Database Connectivity.TMSoftware Development Kit", 1992, 1993, 1994 Microsoft Press, pp. 3-30 and 41-56; See ISO Working Draft, "Database Language SQL-Part 2: Foundation (SQL / Foundation)", CD9075-2: 199.chi.SQL, Sep.11, 1997, etc.] Filed by Selifonov and Stemmer and found at "METHODS OF POPULATING DATA STRUCTURES FOR USE IN EVOLUTIONARY SIMULATIONS", Attorney Docket Number3271.002WO0.

The following examples are intended to illustrate the present invention in detail and are not intended to limit the present invention. Those skilled in the art will be familiar with various variables that can vary substantially with similar results.

Example 1 Decision Tree for an Example of a GAGGS Process

Attached is a set of flow schematics generally illustrating an example of an inductive evolution (DE) process with GAGGS (FIGS. 1-4). 1 provides an aspect determination method from the idea of desirable properties to the selection of genetic algorithms. 2 provides an inductive evolution decision tree from the selection of genetic algorithms to a library of refined source strings. FIG. 3 illustrates the aspect processing steps from the refined source library to the natural derived library of strings. 4 shows a process of progressing a natural string into a string of desirable properties.

In general, these charts are schematic diagrams of the arrangement of members and schematic diagrams of the process decision tree structure. It will be apparent that various modifications to the particular DEGAGGS arrangement can be developed and implemented as described herein. Although the absence of most genetic artificial neuronal network learning as well as specific property control modules and links have been omitted for clarity of construction, those skilled in the art will be familiar with it. These charts are arranged in series, with each head connected. Additional materials and implementations of each GO module, work order and various GO arrangements in the tree (eg GAGGS) are available as various software packages. Suitable references for existing sample software include, for example, http://www.aic.nrl.navy.mil/galist/ and http://www.cs.purdue.edu/coast/archive/clife/FAQ/www There is /Q20_2.htm. It will be apparent that the various decision steps shown in FIGS. 1-4 are most easily performed by a computer using one or more software programs that facilitate the selection / decision process.

Example 2 Modeling Cost Estimation

The use of very definite / low levels of site degenerate degenerate synthetic oligos (<0.05 to 5% per site) can yield significant cost savings in constructing libraries that are substantially mutagenically convoluted. However, PCR assembly gene synthesis uses the synthesis of two related oligos per mock crossover mode, typically as all crossover embodiments between source entry sequences.

However, as can be clearly seen from the combinatorial properties of the following examples and nucleic acid evolution algorithms, even when constructing a very large (10 ⁹ to 10 ¹⁰ ) gene library for physical selection, a gene family of about 1.6 kb in size can be evolved. 40-mer oligos of less than 10 ³ are used.

Some general examples below provide examples of the cost of gene synthesis components for GAGGS, where the cost estimate is, for example, $ 0.7 per base (amount from 40 to 50 nmol suitable for gene reassembly procedure) on an arbitrary basis. It is based. The use of higher doses in oligo synthesis substantially reduces unit cost (eg, a tenfold reduction) and also reduces the overall cost of oligo synthesis. Oligo synthesis is a common process inherently parallel that is easy to automate and therefore easy to increase yield. At present, the non-chip parallel apparatus for oligo synthesis has an effective capacity capable of simultaneously (single loading) synthesizing 196 (2 x 96) of each 60-mer oligos within 5 hours. At this point, the hardware cost is less than $ 100K and the reagent cost is less than $ 0.07 per base. Thus, in view of these costs, the cost estimates made in the examples below can be reduced by more than eight times.

Example 3: GAGGS of a single source low mutagenic library

This example is directed to GAGGS of a single source low mutagenic library derived from the mean gene (approximately 1.6 kb) if sequence information is provided for one 1.6 kb gene (500 aa + "convenience" initiation / end oligo code). will be. The goal of this example is to construct a library of genetic variants with all possible single amino acid changes, one aa change for each gene copy in the library.

Relevant parameters include the number of oligos, such as the cost of constructing one source 1.6 kb gene, for example from a 40-mer oligo having a complete 20 + 20 base overlap, by non-error easy assembly PCR, the number of all possible aa substitution mutations. , The number of unique non-degenerate 40-mer oligos used to "construct" all possible single aa mutations, the minimum number of single codon degenerate oligos at all unique fixed positions used to feed in all possible single aa mutations except the stop codons, and There is a minimum number of 1 codon fully degenerate oligos of all unique fixed positions used to inject all possible single aa mutations.

For a 1.6 kb gene, 1∞1,600: 40∞2 = 80 oligos; $ 0.7∞40∞80 = $ 2,240. N = 500∞19 = 9,500. 9,500x2 = 19,000; $ 532,000 @ $ 0.7 / base $ 56 / gene, 1 / pool 500x2x3 = 3,000; $ 84,000 @ 0.7 / base, $ 8.85 / gene, 20 phenotype / pool, defined multiplicity (eg, using only three variable codons, two of which are degenerate: NNT, VAA, TGG) 500 × 2 = 1,000; $ 28,000 @ 0.7 / base $ 2.94 / gene, 20 phenotypes / pools, asymmetric multiplicity (this results in the presence of a significant number of truncated genes in the synthesized library).

The same physical oligo list used for the primary GAGGS is used for the secondary GAGGS to synthesize a library comprising about 95% of all possible combinations of any two single aa changes. Addition oligos are used to make 100% to include a combination of mutations present in the +/- 20bp proximal position. If one or more mutations obtained from the previous GAGGS were found to be beneficial, the synthesis of up to 42 new oligos was used to make all combinations of new mutations present within +/- 20 bp of the beneficial mutations. The cost used for subsequent GAGGS is marginal and only increases linearly linearly, while the diversity sampled in a recursive manner increases exponentially.

Example 4 GAGGS (GAGGS Corresponding to Primary Family DNA Shuffling) of Recombinant (Nonmutant) Libraries inherited from a Group of Gene Families

Sequence information is provided for six normal mean (1.6 kb) genes, each with six source regions and six homology regions (six "tofu" and "tail" to chimeric each homology region). If there is.

Related variables include: The number of oligos and thus the cost used to construct the six source 1.6 kb genes (from 40-mer oligos with full 20 + 20 overlap by non-error easy assembly PCR), 1 crossover per paired homology region The number of unique pair crossovers that appear between all matching homology regions, the number of all possible chimeras using the crossover, the theoretical library size, and the cost and number of specific oligos used to construct all possible chimeras. As noted above, 6∞1,600: 40∞2 = 480 oligo, $ 0.7∞40∞480 = $ 13,440, ie $ 2,240 per gene construct. N = 180 when calculated according to the formula N = k∞m∞ (m-1) (where m = 6, the number of sources and k = 6, the number of pair homology regions suitable for the crossover condition). When calculated according to the following formula, X = ∼5.315∞10 ⁹

Where X is the theoretical library size and n is the number of crossovers present in each library list (an integer from 1 to k), 2∞180 + 480 = 840; $ 0.7∞40∞840 = $ 23,520; $ 0.000048 per gene construct. If only 10 ⁶ is selected, the cost of the oligo is $ 0.24 per gene construct. For screening 10 ⁴ , the oligo cost is $ 2.35 per gene construct.

The cost of running GAGGS multiple times is not additive because most of the excess oligo obtained from previous GAGGS can be reused for the synthesis of subsequent generation libraries. Although virtually only a small amount of all constructed genes are selected (eg 10 ⁴ , the cost of oligos used per gene construct is $ 2.35), the cost of oligos is comparable to the cost of analyses used per gene assay criteria. In addition, the cost of oligo synthesis widely used industrially is decreasing.

Example 5 Step by Step GAGGS

This example provides a GAGGS family model step-by-step protocol.

A group of genes / proteins (DNA or AA sequences) are selected. Pick a crossover point. All possible pair alignments are made to identify pair homology regions that meet the crossover manipulation conditions (length, percent identity, stringency). Crossover points are selected, one for each pair homology subcolumn, with each subcolumn in the middle or randomly, or according to a probabilistic model based on an annealing made from a histogram of crossover probability ratings for each pair of sources. Oligos are selected and synthesized for assembly PCR. Assemble the gene / library from the synthesized oligos. These libraries are screened / selected as described above.

Example 6 Subtilisin Family Model

Align the amino acid sequence (optimize the codons used for reverse reads for the preferred expression system and minimize the number of oligos used for synthesis). Point blot pair alignments of all possible pairs of seven sources were prepared (FIGS. 5, 6, 7). 5 is the% similarity in alignment of the seven sources. The amino acid sequences except the leader sequence were aligned. 6 is a dot plot alignment of sequences to identify similarity regions. 7 is a point plot showing paired crossover points in an aligned sequence.

Pairs 6 and 7 indicate 95% identity for each window of ≧ 7 aa, while all other pairs show 80% identity for each window of ≧ 7 aa. Note that the alignment of the alignment (and the crossover between subsequent appearing sources) can be adjusted individually for each pair so that low homology crossovers may appear despite the highly homologous source. Thus, this model also does not reflect structural or active site bias.

As an example, assuming that there are about 400 amino acids each, 1200 bp (including the leader sequence), or 7 sources of about 275 amino acids and 825 bp for mature proteins, GAGGS for the subtilisin family model is calculated by gene synthesis. The total sequence to be produced is 7 x 825 x 2 + about 500 = 12 kb. In the case of 40-mers with full overlap assembly (20 + 20 bp overlap) about 300 oligos are used.

Based on the alignment results, the paired crossover oligos for constructing chimeras with one crossover per each homology subcolumn have about 180 homology subcolumns consisting of 170 in the crypto area and 10 sub-rows in the leader area. In the case of having two 60-mers for each crossover point and two sets of heads for each source phase, about 360 additional oligos involved in constructing the crossover can be used. The total number of oligos is about 660 (300 40-mers and 360 60-mers). Operating at $ 0.70 per base, the total cost of the oligo is about $ 23,520. The reagent cost is approximately $ 0.07 per base, with a total cost of approximately $ 2,252.

Example 7 Naphthalene Deoxygenase

Naphthalene deoxygenase is a non-hemene reducing deoxygenase. There are three or more types of naphthalene deoxygenase that are closely related but unique in their catalytic properties. FIG. 12 is a schematic of the% similarity plot for three types of naphthalene deoxygenase, providing amino acid sequences of ISP large subunits (subunits involved in substrate specificity).

In the case of about 1,400 amino acids in size, 3x1,400 total base pairs = 260 of 40-mer oligos for 20 + 20 overlapping gene synthesis are used. Sequence alignment plots confirmed 14 + 19 + 23 = 112 60-mer high stringency oligos used for recombination. The oligo cost was $ 0.70 per base and the cost used for the synthesis was about $ 12,000 with a synthesizer time of about 9 hours to prepare the oligo. The estimated library size is about 9.4 × 10 ⁹ libraries.

Example 8 Type 1 GAGGS Computation

As mentioned above, one aspect of the present invention is to provide one source GAGGS. This method provides a polynucleotide of desirable properties. This method is carried out through the following steps. (a) providing a string of origin sequence properties encoding a polynucleotide or polypeptide, (b) an overlapping sequence fragment of the string of full origin properties and a full polynucleotide strand complementary to the string of this origin property (the source sequence is Providing a set of strings of a predetermined length encoding single-stranded oligonucleotide sequences comprising oligonucleotides suitable for assembly PCR), (c) one mutation per strain string containing all possible single point mutations (e.g., Constructing a set of derivatives of the original sequence comprising a variant having all possible single point mutations, (d) a sequence of lengths of overlapping strings encoding the both strands of the source oligonucleotide sequence and the sequence comprising the mutations Set of nested strings of predetermined length to encode regions (same assembly PCR Providing an oligo comprising a suitable single point mutation), (e) synthesizing a single-stranded oligonucleotide set according to step (c) (e.g., incorporating a single point mutation during gene assembly, etc.) Or constructing or reconstructing a variant thereof), (f) assembling a library of mutated genes required for assembly PCR from single-stranded oligonucleotides (provide one collection, partial collection, or one per container).

In one genetic method per container (or other physically isolated library components such as other methods included in an array), wild type oligos are excluded from mutations, and (g) select recombinant polynucleotides that have evolved to the desired properties or Screening. Optionally, in further step (h), an unwinding sequence of the mutant polynucleotides evolved with the desired properties in order to determine useful mutations (i.e., where the library member has the desired sequence and which sequence is the sequence) (A step of determining cognition), which is carried out by position sequencing rather than actual sequencing if the assembly PCR is 1 form per container, ie whether the physical position of the component is appropriate to provide the function of the sequence. do). A further optional step (i) comprises the step of assembling a library of recombinant variants from single-stranded oligos by assembly PCR, combining all possible or some useful mutations in some or all possible combinations. This is done using the same oligo set, where if some of the mutations are positionally close (in either oligo), an additional single-stranded oligo containing a combination of mutations is prepared. Optional step (j) may also include selecting or screening recombinant polynucleotides further evolved to the desired properties.

As an example, the single source GAGGS operation per kb of sequence is as follows.

Genome length: 1000 bp.

Primary mutation rate: 1 amino acid / gene

Number of oligonucleotides used in wild-type gene construction: 52 (40-mer, 20 + 20 overlapping synthetic scheme).

Number of oligonucleotides giving all non-terminal possible single point mutations (total possible number 333): non-degenerate oligos: 13320. Partially degenerate 40-mer, 1 pg per oligo position: 1920. Fully degenerate oligo-, 40-mer, 1 fg per oligo: 666.

In addition, error prone PCR assembly is performed, but no sequencing is performed prior to subsequent steps.

The number of additional oligos used to construct all possible recombinants with useful mutations is about 10% of the previous oligos. However, about 95% of all possible recombinants with useful mutations can be made in the primary set prepared above.

Example 9 Design of Crossover Oligonucleotides to Synthesize Chimeric Polynucleotides

First, we identify and select a substring from the source string to apply the crossover operator to the chimeric junction. This step includes the steps of a) identifying all or part pair homology regions between all source strings, b) all or part pair homology identified to index one or more crossover points present in each selected pair homology region. Selecting an area, c) selecting one or more paired nonhomologous areas to index one or more crossover points present within each selected phase nonhomologous area (“c” is an optional step that can be excluded, It is also a step in which active system elitism can be applied), providing a location of the source string suitable for further selection of the crossover point and providing a set of source indexed regions / regions (sub-rows).

Second, the crossover point existing in each sub-row among the sub-row sets selected in step 1 is further selected. This step may be based on: a) randomly selecting one or more crossover points from each selected subcolumn, and / or b) one or more annealing simulations to measure the probability of crossover point selection within each selected subcolumn. Using the model, selecting one or more crossover points existing within each selected subcolumn, and / or c) constructing a set of paired crossover points by selecting one crossover point approximately in the middle of each selected subcolumn, Wherein each point is indexed at a corresponding character position in each source string desired for formation of the chimeric junction.

Third, optionally adjust codon usage. The process may vary depending on the method used to measure homology (heat encoding DNA or AA). For example, if a DNA sequence was used: a) Codon adjustments for all source strings for the selected expression system. B) Adjust codons between sources to normalize codon usage for all given aa present at all corresponding positions. This process can significantly reduce the total number of specific oligos that require gene library synthesis, which would be particularly advantageous when AA homology is higher than DNA homology or in highly homologous genes (eg, 80% + identity) family. Can be.

This step should essentially be carried out with similar attention to the expression of the elliptically mutant effector. That is, the benefits of introducing this bias versus the reduction in the number and final cost of oligomers that may produce undesirable results should be considered. Most preferably, it is preferable to use codons that encode AA present at a given position among the majority of sources.

If an AA sequence is used, a) reverse translation of the sequence into degenerate DNA, b) confine the degenerate nucleotides by reference to the codon usage in the original DNA (of the majority or corresponding source) from position to position (or Codon adjustments are made to the selected expression system in which physical analysis can be performed.

In addition, this step can be used to introduce any restriction site into the coding region of the gene to subsequently identify / QA / unlock / manipulate the library list if necessary. All crossover points (indexed with pairs of sources) identified in step 2 above are indexed correspondingly to the adjusted DNA sequences.

Fourth, the oligo sequence is selected for the gene assembly schematic. This step includes several decision steps.

Generally, homogeneous 40-60-mer oligos are used (if longer oligos are used, additional oligos are used to reduce the number of oligos used to construct the source and to provide a sample of closely located crossovers / mutants. use.

Choose whether short / long chain oligos are allowed (ie yes / no crystals). A determination of “yes” reduces the total number of oligos for high homologous genes of various lengths with gaps (deletions / insertions), especially gaps of 1 to 2 aa.

The overlap length (usually 15 to 20 bases, which can be symmetrical or asymmetrical) is chosen.

Select whether degenerate oligos are allowed (Yes / No). This is another powerful feature of cost savings and a powerful means of gaining additional sequence diversity. Partial degenerate schemes and minimal degenerate schemes are particularly advantageous for constructing mutagenic libraries.

If you use software tools for this manipulation, you can use several parameter parameters to select the maximum library complexity and minimum cost. Complex assembly schemes using oligos of varying lengths significantly complicate the assembly of the library in the indexing process and subsequently in the location-coded parallel or partial pooling format. If this is done without high software, a simple and uniform scheme (eg total oligo 40 base length with 20 base overlap) can be used.

Fifth, design a "convenience sequence" before and after the source string. It would be ideal if the last set of all the library listings was the same set. Such sequences include any restriction sites, primer sequences for assembly product identification, RBS, leader peptides and other special or desirable features. Originally, the convenience sequence can be obtained at a later stage, at which stage a set of "imitation" of appropriate length, such as substrings of easily distinguished forbidden letters can be used.

Sixth, an indexed matrix of oligo strings is constructed to construct all sources according to the chosen scheme. Indices of all oligos include the source identifier (source ID), an indication of the code or complementary chain, and the location number. The crossover point is determined from the indexed cipher strings of all sources using the DumiBu Convenience Substring. Construct a complementary chain of all strings. All cipher strings are selected according to the assembly PCR scheme selected in step 4 (eg, an increase of 40 bp). All complements are split according to the same scheme (eg 40 bp with 20 bp shift).

Seventh, construct an indexed matrix of oligos for all phase crossover manipulations. First, all oligos with paired crossover markers are determined. Next, determine the total set of all oligos that have the same location and the same pair of source crossover markers (four per crossover point). Next, take all sets of four oligonucleotides labeled with the same crossover marker, and set another derivative of four chimeric oligonucleotides containing letters encoding two code chains and two complement chains (e.g. 40 = 20 + 20 With a 20 bp shift in the scheme). It is possible for two code strings to retain the rear end of the second source after the subrow of the forward end sequence of one source and the next crossover point. In addition, the complement sequence is designed in the same manner to obtain an indexed complete list of strings encoding oligos suitable for assembly of the gene library by PCR.

This list optionally detects all excess oligos, counts them, removes them from the list, and improves them, while assigning count values to the "rich = amount" range in the index of each oligo row. This is a very useful step that can reduce the total number of oligos for library synthesis, particularly when the source sequence is highly homologous.

Example 10 Program Algorithms for Designing Synthetic Oligonucleotides

The following is a program overview useful for the design of oligonucleotides used in the synthesis / recombination protocol.

If an alignment of the protein and codon bias tables is provided:

For each position in the aligned protein,

The codon bias table is used to search for a minimal set of degenerate codons that encode amino acids at each position.

For each sequence aligned,

A three letter codon (DNA) encoding the amino acid at each position in this sequence is added to the DNA form of the sequence.

The gap is represented by a specific codon.

For each sequence of DNA constructed as described above,

At this stage, ignore the gap.

Each window = approximate raised size

Check for terminal degeneracy.

The window length is increased and decreased to maintain the length limit while minimizing terminal degeneracy.

Add oligos and all sequences provided with window limits.

Add oligos and all sequences provided with reverse window limits.

for each position in dnaseps,

For each sequence,

Add oligos provided a limit comprising a minimum amount of sequence from the current sequence on the 5 'side of the gap and a minimum amount of sequence from the current sequence on the 3' side of the gap.

Add oligo for reverse limit repeatedly.

Add oligos: if a list of sequences (DNA) and limits is provided

For each position within the limit,

All unique bases in this position are obtained from the DNA sequences in the list.

The base (or degenerate base symbol) at this position is prepared.

If the total degenerate position is greater than the number you define,

Divide the sequence listing in two, add the oligo provided sequence listing 1 and add the oligo provided sequence listing 2 (repetitive).

In addition, the oligo (base set at each position) is raised and added to all oligo list displays in the list.

Example 11 Crossover Point Selection

8-11 are schematic diagrams of various processes and process criteria for selecting oligonucleotides for recombination between source nucleic acids. Panel A of FIG. 8 shows an increase in the crossover probability of creating a common point plot alignment and similarity regions of the two sources. Panel B shows that the crossover can be selected based on a simple logical / physical filter, ie using the physical or actual annealing temperature of the oligonucleotide, such as a linear annealing temperature. Panel C shows a variety of more complex filters through varying annealing temperatures, ie, by appropriately adjusting the annealing temperature physically or practically to achieve a specific crossover.

9 schematically illustrates the process of introducing an indexed crossover point into each aligned original sequence. Briefly, the sequences were aligned and the position index (marker range) of each crossover point was schematically illustrated with a vertical identifier marker. As shown in Fig. 8, the crossover points for the root m and the n are indicated by the identifier, the position number of the root m (the head), and the position number of the root n (the tail). This process was repeated for all sources in the data set by applying an oligonucleotide gridding operator (grid of position indices representing the beginning and end of all oligonucleotides in the PCR assembly operation). FIG. 10 schematically shows a complete list of oligonucleotide sequences for assembling all sources. This data set provides a semi-list of oligos with crossover markers by identifying and simplifying all pairs of oligonucleotide sequences with a pairwise matching crossover index. FIG. 11 provides a schematic for obtaining a sequence listing of chimeric oligonucleotides for each selected crossover point. Briefly, two pairs of oligo sequences with matching pair crossover indices are selected (under arrow 1). The sequence of chimeric oligonucleotides is made around a pair crossover point (head, head, s or a chain). In the "40 = 20 + 20" assembly scheme (when the 40-mer has 20 residues from each source), more than 60 bp for each chimerization (two for each crossover point regardless of the relative position within each oligo) Use only one oligo long. Such experimental findings can be explained through changes in cleavage and binding manipulations in the indicated S or A chains (eg, as in FIG. 8), which rules are summarized in the guide. In chimeric oligos, the A or S chains and the sequence head and tail subfragments were indicated using a subset of the predetermined rules of the guidance table. The selection or rule obtained from the guide can be automated based on a comparison of the relative positions of the crossover points in the oligo (non-logical operation) (under arrow 3). This process can be repeated for every set of oligos having the same crossover index to obtain a list of chimeric oligo sequences for each selected crossover point.

The methods and materials described so far may be modified without departing from the spirit or scope of the invention as claimed. In addition, the present invention can be used for a variety of uses, such as:

Use for an integrated system for testing shuffled nucleic acids and / or generating shuffled nucleic acids, including repeating processes.

An assay, kit, or system using any one of the aforementioned selection strategies, materials, ingredients, methods, or substrates. The kit may optionally further comprise one or more containers containing instructions, packaging materials, assays, devices or system components, etc. for carrying out the method or assay.

In another aspect, the present invention provides a kit incorporating the methods and apparatus described herein. The kits of the invention optionally include (1) shuffled components as described herein, (2) instructions for carrying out the methods as described herein and / or for the manipulation of the selection procedures described herein, (3) a container containing one or more analytical components, (4) nucleic acids or enzymes, other nucleic acids, mutant plants, animals, cells, and the like, (5) packaging materials, and (6) determination steps described herein with respect to GAGGS At least one of the software to perform any of the steps.

In another aspect, the invention relates to the use of any component or kit described herein, to the implementation of any method or assay described herein and / or to the use of any device or kit implementing any assay or method described herein. to provide.

The foregoing embodiments are illustrative and not restrictive. Those skilled in the art will be familiar with various noncritical parameters that can be modified to obtain practically similar results. All patents, applications, and publications cited herein are hereby incorporated by reference in their entirety.

Claims (88)

Providing a plurality of character strings corresponding to the plurality of nucleic acids, wherein the strings comprise one or more heterologous regions when aligned for maximum identity; Sorting the strings; Defining a quasi-sequence set of strings comprising at least two quasi-sequences of the plurality of source strings; Providing a set of oligonucleotides corresponding to the set of string subsequences; Annealing the oligonucleotide set; And Extending one or more members of the oligonucleotide set with a polymerase or ligation of two or more members of the oligonucleotide set with a ligase to produce one or more recombinant nucleic acids. 2. The method of claim 1, wherein the string includes one or more similarity regions when aligned for maximum identity. The method of claim 1, wherein at least one of said source strings is an evolutionary or artificial intermediate. The method of claim 1, wherein the one or more source strings correspond to the designed nucleic acid. The method of claim 4, wherein the designed nucleic acid is designed to minimize energy for the encoded polypeptide. The method of claim 1, further comprising applying one or more gene operators to one or more source strings or one or more string subsequences, wherein the genetic operator is selected from: : Mutation of one or more source strings or one or more string subsequences, amplification of one or more source strings or one or more string subsequences, fragmentation of one or more source strings or one or more string subsequences, one or more source strings or one or more strings subsequences, or additional Crossover between any of the strings, ligation of one or more source strings or one or more string subsequences, elitism operations, calculation of sequence homology or sequence similarity of aligned columns, repeated use of one or more genetic operators for string evolution , Random operator of one or more source strings or one or more string subsequences, deletion mutations of one or more source strings or one or more strings subsequences, one or more source strings or one or more strings Select from among mouth mutations, subtraction of one or more source strings or one or more string subsequences with inactive sequences, selection of one or more source strings or one or more string subsequences with active sequences, and killing one or more source strings or one or more string subsequences How to be. The method of claim 1, further comprising selecting a diplomat sequence with moderate sequence similarity between two or more of the plurality of strings. The method of claim 1, wherein said oligonucleotide set comprises a plurality of overlapping oligonucleotide sequences. 2. The method of claim 1, wherein the set of string subsequences is defined by selecting a string length and subdividing at least two of the plurality of source strings into segments of the selected length. The method of claim 1, wherein the step of sorting the strings is performed in a digital computer or web based system. The method of claim 1, wherein the single-stranded oligonucleotide set corresponding to the string subsequence set is synthesized to provide the oligonucleotide set. The method of claim 1, Pooling all or a portion of the oligonucleotide set; Hybridizing the resulting pooled oligonucleotides; And Extending the resulting hybridized plurality of oligonucleotides, wherein the resultant stretched one or more double-stranded nucleic acids further comprise a sequence comprising from two or more source strings. 12. The method of claim 11, further comprising denaturing the double stranded nucleic acid to produce a heterologous mixture of single stranded nucleic acids. The method of claim 11, (Iii) denaturing the double stranded nucleic acid to produce a heterogeneous mixture of single stranded nucleic acids; (Ii) rehybridizing the heterogeneous mixture of single-stranded nucleic acids; And (Iii) extending the rehybridized double stranded nucleic acid with the polymerase as a result; Characterized in that it further comprises. 14. The method of claim 13, further comprising repeating steps (iii), (ii) and (iii) two or more times. The method of claim 1, further comprising selecting one or more recombinant nucleic acids for the property of interest. The oligonucleotide set of claim 1, wherein the oligonucleotide set is prepared by synthesizing an oligonucleotide comprising one or more modified source string subsequences, wherein the subsequence is substituted one or more characters of the source string subsequence with another character one or more times Source string subsequence modified by means of; A source string subsequence modified by deleting or inserting one or more characters of the source string subsequence one or more times; Source string subsequences modified by including degenerate sequence letters in one or more selected positions randomly or nonrandomly selected; A source string subsequence modified by including a string from a different string from a second source string subsequence at one or more locations; Source string subsequence biased based on frequency of occurrence in selected libraries of nucleic acids; And A method characterized by combining at least one of the source string subsequences comprising one or more sequence motifs artificially included in the subsequences. 18. The method of claim 17, wherein said sequence motif is RNA 2 affecting N-linked glycosylation sequence, O-linked glycosylation sequence, protease sensitive sequence, collagenase sensitive sequence, Rho dependent transcription termination sequence, transcription efficiency And a secondary structural sequence, an RNA secondary structural sequence that affects translation efficiency, a transcription enhancing sequence, a transcriptional promoter sequence, or a transcriptional silencing sequence. The method of claim 1, wherein said oligonucleotide set comprises one or more modified or degenerate positions compared to corresponding subsequences of one or more source strings. The method of claim 1, further comprising selecting one or more recombinant nucleic acids based on hybridization with the selected nucleic acid or selected nucleic acid set. The method of claim 1, wherein the one or more source strings comprise two or more source strings, the oligonucleotide set comprises one or more oligonucleotide members comprising a chimeric nucleic acid sequence, and the one or more oligonucleotide members comprises two or more oligonucleotide members Wherein said at least two oligonucleotide member subsequences correspond to at least two subsequences obtained from at least two source strings and are separated by crossover points. 22. The method of claim 21, wherein the crossover points identify a plurality of source character subcolumns from a plurality of source strings of two or more, and then align the subcolumns to indicate pair identity between the subcolumns, and a point in the aligned sequence. Characterized by being obtained as a crossover point. 22. The method of claim 21 wherein the crossover point is selected randomly. 22. The method of claim 21 wherein the crossover point is selected non-randomly. 22. The method of claim 21, wherein the crossover point is selected non-randomly by selecting a crossover point that is nearly centered in one or more identified pair identity zones. The method of claim 21, wherein one or more crossover points for one or more oligonucleotide members are selected from the outer region of the identified pair homology region. The method of claim 1, further comprising adding one or more oligonucleotide members of the oligonucleotide set to a concentration higher than one or more additional oligonucleotide members of the oligonucleotide set. The method of claim 1, further comprising incubating one or more members of the oligonucleotide set with recombinant nucleic acid and polymerase. The method of claim 1, further comprising denaturing the recombinant nucleic acid and contacting the recombinant nucleic acid with one or more additional nucleic acids from the oligonucleotide set. The method of claim 1, further comprising denaturing the recombinant nucleic acid and contacting the recombinant nucleic acid with an additional one or more nucleic acids prepared by cleaving the source nucleic acid encoded by the one or more source strings. Way. The method of claim 1, further comprising denaturing the recombinant nucleic acid and contacting the recombinant nucleic acid with an additional one or more nucleic acids prepared by cleaving the source nucleic acid encoded by the one or more source strings, wherein the source Cleavage of the nucleic acid is by at least one of chemical cleavage, cleavage with DNase, and cleavage with restriction endonucleases. The method of claim 1, wherein the source string is EPO, insulin, peptide hormones, cytokines, epidermal growth factor, fibroblast growth factor, hepatocyte growth factor, insulin-like growth factor, interferon, interleukin, keratin synthetic cell growth factor, leukemia inhibitory factor Oncotin M, PD-ESCF, PDGF, Pliotropin, SCF, c-kit ligand, VEGEF, G-CSF, tumor gene, tumor suppressor, steroid hormone receptor, plant hormone, disease resistance gene, manufacturer resistance A method characterized by encoding one or more nucleic acids corresponding to one or more proteins or genes selected from among genes, bacterial genes, monooxygenases, proteases, nucleases and lipases. The method of claim 1, wherein said oligonucleotide set comprises one or more oligonucleotide members that are between about 20 to about 60 nucleotides in length. The method of claim 1, further comprising providing a selected recombinant nucleic acid by selecting for the feature or characteristic desired for the recombinant nucleic acid. 35. The method of claim 34, further comprising recombining the selected recombinant nucleic acid with homologous nucleic acid and one or more of the oligonucleotide members obtained from the oligonucleotide set. The method of claim 1, further comprising selecting for the feature or feature of interest and providing the selected recombinant nucleic acid, wherein the feature or feature of interest is an in vivo selection assay or a parallel solid phase. A method characterized by being selected from the assay. The method of claim 1, further comprising the step of selecting for the feature or feature of interest and providing the selected recombinant nucleic acid, wherein the feature or feature of interest is selected in an in vivo assay. Characteristic method. The method of claim 1, further comprising the step of deconvolution of said recombinant nucleic acid. The method of claim 1, further comprising sequencing or cloning the recombinant nucleic acid. The method of claim 1, wherein said recombinant nucleic acid is synthesized in vitro by assembly PCR (assembly PCR). The method of claim 1, wherein said recombinant nucleic acid is synthesized in vitro by error-prone assembly PCR. The method of claim 1, wherein said source string or oligonucleotide set is selected from a computer. a) providing a source string encoding a polynucleotide or polypeptide; b) providing a set of oligonucleotide strings of a predetermined length, encoding a plurality of single-stranded oligonucleotide sequences comprising the sequence fragments of the source string and the complements thereof; And c) producing a set of source sequence derivatives comprising a sequence variant sequence having one mutation per variant string and comprising a plurality of mutations. The method of claim 43, wherein the plurality of single-stranded oligonucleotide sequences overlap in the sequence. The method of claim 43, further comprising applying one or more gene effectors to the source string or one or more oligonucleotide strings, wherein the genetic effector is a mutation, source string or one or more oligos of the source string or one or more oligonucleotide strings. Amplification of a nucleotide string, fragmentation of a source string or one or more oligonucleotide strings, crossover between any string in the source string or one or more oligonucleotide strings or additional strings, ligation of the source string or one or more oligonucleotide strings, An operation of elitism, sequence homology or sequence similarity of an aligned string comprising a source string or one or more oligonucleotide strings, one for evolution of a string Use of gene effectors on a gene, application of random operator to a source string or one or more oligonucleotide strings, deletion mutations in the source string or one or more oligonucleotide strings, insertion mutations in the source string or one or more oligonucleotide strings, inactive And a subtraction of the source string or one or more oligonucleotide strings with sequence, the selection of the source string or one or more oligonucleotide strings with active sequence and the killing of the source string or one or more oligonucleotide strings. The method of claim 43, d) providing a set of overlapping strings of predetermined length that encode both chains of the source string sequence; And e) synthesizing a single-stranded oligonucleotide set according to steps (c) and (d). 47. The method of claim 46 wherein f) assembling the recombinant nucleic acid library from single-stranded oligonucleotides by assembly PCR. A library prepared by the method of claim 47. The method of claim 47, g) selecting or screening the library for one or more recombinant polynucleotides having the desired properties. The method of claim 48, h) deconvoluting one or more selected polynucleotide sequences. The method of claim 46, wherein the one or more selected polynucleotide sequences are solved by sequencing the selected polynucleotides or by digesting the one or more selected polynucleotides. 47. The method of claim 46, wherein said sequence is unannealed by unwinding one or more selected polynucleotides. 47. The method of claim 46, further comprising repeatedly shuffling or selecting the recombinant nucleic acid library. Aligning the source string corresponding to the divergent nucleic acid to identify whether it is a sequence identity region or a sequence divergence region; Defining a diplomat string that is an intermediate of the sequence between the source strings; Synthesizing at least some of the Diplomat sequences to produce Diplomat nucleic acid; And Recombining the selected nucleic acid mixture comprising the source nucleic acid or fragments thereof and the diplomat nucleic acid, thereby facilitating recombination between two or more divergent nucleic acids. 55. The diploid of claim 54, wherein a plurality of overlapping oligonucleotides corresponding to the sequences of the diplomat nucleic acid are synthesized, the overlapping oligonucleotides are hybridized, and the overlapping oligonucleotides are incubated with a polymerase. A method characterized by synthesizing a roman nucleic acid. 55. The method of claim 54, further comprising synthesizing an oligonucleotide pool corresponding to one or more source strings, wherein the oligonucleotide pool is present in a mixture of selected nucleic acids. A mixture of selected nucleic acids prepared by the method of claim 56. Inputting a plurality of amino acid sequence strings into a digital system; The amino acid string entered in the digital system is translated back into a plurality of nucleic acid strings, the reverse translated nucleic acid sequence being selected for at least one of the species codon bias in the selected expression host and the optimized sequence similarity between the plurality of nucleic acid strings. Becoming; And A method of synthesizing and recombining nucleic acids, characterized in that it comprises the step of synthesizing one or more sets of oligonucleotides corresponding to one or more reverse translated nucleic acid sequences. 59. The method of claim 58, further comprising hybridizing members of the one or more oligonucleotide sets to a set of fragmented nucleic acids encoding one or more amino acid polymers corresponding to each other or one or more amino acid sequence strings. 60. The method of claim 59, further comprising extending the hybridized one or more nucleic acids with a polymerase. 59. The method of claim 58, wherein the resulting one or more nucleic acids are fragmented and the resulting secondary fragmented nucleic acids are hybridized with one another or with one or more members of a set of oligonucleotides or corresponding to one or more amino acid sequence strings. And hybridizing with the first set of fragmented nucleic acids encoding the polymer. Parameterizing a set of nucleic acids or proteins to provide a multidimensional set of data points; Extrapolating one or more multidimensional datapoints by hypothesis from the multidimensional datapoint set; And converting the multidimensional data point according to the hypothesis into a new string corresponding to the nucleic acid or protein according to the hypothesis. 63. The method of claim 62 comprising synthesizing a nucleic acid or protein by the hypothesis. 63. The method of claim 62, further comprising analyzing the principal components of the multidimensional set of data points. 63. The method of claim 62, further comprising shuffling said hypothesis nucleic acid or subsequence thereof into additional nucleic acid. 63. The method of claim 62, wherein said nucleic acid or protein set is parameterized by correlating each residue thereof to a matrix of radix indicators. 67. The method of claim 66, wherein an origin is specified over the center of the tetrahedron to depict the matrix as a tetrahedron, each corner is represented as an radix, and each residue of the nucleic acid is located at a different corner to produce a matrix of radix indicators. How is it characterized. 63. The method of claim 62, comprising correlating each multidimensional data point with an output vector that identifies a relationship between a matrix of dependent variable Y and a matrix of predictor variable X. 69. The method of claim 68, wherein said correlation is performed by Partial Least Square Projection to Latent Structure (PLS) for latent structure analysis. 63. The method of claim 62, wherein each multidimensional data point comprises one or more different parameters, wherein the parameters are plotted with respect to each other in an n-dimensional hyperspace including one or more dimensions for each parameter. Preparing an initial library of about 10 ⁶ or more recombinant nucleic acids comprising forms of at least about 10 ⁵ different members that are not identical; Hybridizing the library with one or more populations of nucleic acids corresponding to one or more subsequences of different library members; Separating library members that hybridize to one or more nucleic acid populations to aggregate nucleic acid libraries for members that hybridize to one or more nucleic acid populations; And Providing a recombinant nucleic acid library in which the sequence of interest is concentrated, the method comprising selecting the integrated library members for one or more desired properties. The method of claim 71, wherein the initial library has about 10 ⁹ to 10 ¹² members. The method of claim 71, wherein said at least one nucleic acid population is immobilized on a solid substrate. 74. The method of claim 73, wherein said solid phase substrate comprises at least one of a column matrix material and a nucleic acid chip. The method of claim 71, wherein the initial library is prepared by synthesizing one or more homologous nucleic acids. An integrated library, characterized in that it is produced by the method of claim 71. 76. The method of claim 71, wherein the initial library is Providing a plurality of source strings comprising one or more similarity regions and one or more heterologous regions when aligned for maximum identity and corresponding to the plurality of nucleic acids; Sorting the strings; Defining a set of string subsequences comprising the subsequences of two or more source strings; Providing a set of oligonucleotides corresponding to the set of string subsequences; Annealing the oligonucleotide set; And Wherein said method is characterized by extending one or more members of a set of oligonucleotides with a polymerase to produce an initial library of nucleic acids. Computer-generated various clusters of strings created by modifying existing strings; And A method of synthesizing a library of biological polymers, the method comprising synthesizing various string communities comprising a biological polymer library. 79. The method of claim 78, wherein said modification comprises recombining an existing string. 79. The method of claim 78, wherein said biological polymer is selected from nucleic acids, polypeptides, and peptide nucleic acids. 79. The method of claim 78, further comprising selecting a member of said biological polymer library for one or more activities. 82. The method of claim 81, further comprising filtering the additional string set or additional library by subtracting additional libraries or additional string sets retaining members of a biological polymer library exhibiting activity below a desired threshold. How it is characterized. 82. The method of claim 81, further comprising filtering the additional set of strings or the additional library by biasing the additional library or set of additional strings that retain members of a biological polymer library exhibiting activity above a desired threshold. How is it characterized. A first data set comprising a first string, a second data set comprising a second string, software to align the first string and the second string, and genetic manipulation on the first string and the second string Software to perform, an output file comprising a third string data set containing a third string containing string subsequences from the first and second strings, and an oligo comprising a plurality of overlapping oligonucleotide sequences corresponding to the third string An integrated system comprising a computer with a nucleotide sequence output file. 85. The integrated system of claim 84, wherein said system further comprises an oligonucleotide synthesis machine that synthesizes a plurality of overlapping oligonucleotides. 85. The method of claim 84, further comprising a plurality of oligonucleotides encoded by a plurality of overlapping oligonucleotide sequences, wherein the oligonucleotides are encoded by a third string when incubated in one or more chain stretch cycles. Integrated system characterized by producing nucleic acids. 85. The integrated system of claim 84, wherein the system further comprises a program having a set of instructions for applying one or more gene operators to the first or second string or any other string. 85. The system of claim 84, wherein the system further comprises a program with a set of instructions for applying one or more gene actuators to the first or second string or any other string, One for mutation, amplification, fragmentation, crossover between one or more columns, ligation of columns, computation of sequence homology or sequence similarity, alignment, elitism, string evolution, randomness, deletion mutation, insertion mutation and death Integrated system, characterized in that selected from the repeated use of the above gene effector.