JP2000511045A - Measurement of gene expression profiles in toxicity determination - Google Patents

Abstract

  (57) [Summary] Measuring the gene expression profile of a selected tissue provides a method for assessing the toxicity of a compound in a test organism. Gene expression profiles are measured by robust parallel signature sequencing of cDNA libraries constructed from mRNA extracted from selected tissues. Gene expression profiles provide extensive information about the effects of administering compounds to test organisms in both acute and long-term and chronic toxicity tests.

Description

DETAILED DESCRIPTION OF THE INVENTION               Measurement of gene expression profiles in toxicity determination                               Field of the invention   The present invention generally provides methods for assessing and / or determining the toxicity of a chemical compound. Detect and monitor phenotypic changes in in vitro and in vivo systems The present invention relates to a method, and more particularly, to a method for determining the toxicity of a drug candidate. Detect changes in gene expression patterns in in vitro and in vivo systems And how to monitor.                                   background   The ability to quickly and conveniently assess the toxicity of a new compound is very important. Thousands New compounds are synthesized each year, and many develop new commercial products and processes Often have little knowledge of its short and long term health effects , Introduced into the environment. In developing new drugs, the safety and efficacy of candidate compounds The cost of assessing is becoming astronomical: the pharmaceutical industry markets new pharmaceutical compounds Spend about $ 300,000,000 on average to bring to the budget, e.g., 13: 226-228 (19 95). Most of these costs are spent on candidate compounds at a later stage in the development process. It was due to failure. In other words, the evaluation of candidate drugs is, for example, relatively inexpensive As a drug candidate by combination assay or in vitro screening assay Progresses from the identification of the compound of interest, for pharmacokinetic studies, toxicity studies, model systems The cost of testing and analysis related to preclinical studies, etc. Increase tremendously. As a result, once promising candidate compounds, Determined to have side effects or cross-reactivity that would make development commercially impractical It can cost tens of millions of dollars to be defined. Big challenge for pharmaceutical development Expose compounds that fail later in the drug test as soon as possible. From consideration.   A drug development program is clearly established for this subject; Increasing costs are an early step in identifying the wrong lead as soon as possible Has created the need to develop more stringent and less expensive screens You. Toxicity assessments indicate that such improvements may be useful for drug development and generally for new compounds. To assess the effects of environmental, health, and safety is there.   Typically, the toxicity of a compound is controlled in one or more test animals under controlled conditions. Monitor effects by administering compounds and on a wide range of parameters To be determined. Parameters include blood chemistry, weight gain or Things like decrease, various behavior patterns, muscle tone, body temperature, respiratory rate, mortality etc. These collectively provide a measure of the health of the test animal. Normal range The degree of deviation of such parameters of the type provides a measure of the toxicity of the compound. This Tests such as acute, prolonged, or chronic toxicity of the compound Can be designed to evaluate Generally, an acute test is one of the test chemicals. Administration in a single dose. The observation period for test animals is short, such as a few hours. But usually at least 24 hours, and in some cases longer than a week Period. In general, long-term testing involves multiple administrations of the test chemical. Include. Test chemicals may be irregular if incorporated into the diet at least once daily Given at specific times, such as during pregnancy, or in some cases regularly but every other week Can be done. Also, for long-term tests, experiments are usually 90 It takes place for less than a day or for a year in dogs. Acute and long-term type tests Conversely, a chronic toxicity test is one in which test chemicals This is a test administered during a portion. In the case of a mouse or rat, this is 2 It is a period of ~ 3 years. For dogs, it is 5-7 years.   Large groups of test animals are used for such assays, especially in chronic toxicity assays. Establishing and maintaining larger animals loses significant costs. Further Omission of such toxicity tests due to species-specific effects may also Does not guarantee that there will be no toxic effects when used in However, such te Strike provides a standardized set of information to determine the safety of new compounds And a database for preliminary evaluation of related compounds. Toxicity determination Break An important area for improvement is new and predictable results in expensive and boring animal assays. It is the identification of a remarkable thing.   In other medical fields, in biotechnology, especially in DNA sequencing Recent advances in the identification of differentially expressed genes in healthy and diseased organisms. I was very interested in applying it to research and research. For example, Adams et al., Scien ce, 252: 1651-1656 (1991); Matsubara et al., Gene, 135: 265-274 (1993); Rosenber. g et al., International Patent Application PCT / US95 / 01863. The target of such applications is the disease process Increase knowledge and identify genes that play important roles in disease processes And diagnostic and therapeutic approaches utilizing expressed genes or their products To provide services. Such an approach is attractive, but manifests itself Thorough or even sampled sequencing of selected genes The roach is accompanied by a lot more effort required: 30,000-35,000 different Is presumed to be expressed in representative lactating tissues at any given stage . For example, edited by Ausubel et al., Current Protocols, 5.8.1-5.8.4 (John Wiley & Sons , New York, 1992). To sequence smaller samples of a certain number of gene products Is a major project that requires an industrial-scale source. Therefore, expressed The application of robust sequencing routines for selected genes is still a Not exceed.   Candidate drug that provides broader and more accurate information about the health of the test animal The availability of new assays to assess the toxicity of compounds such as Is desired. Such additional assays are preferably based on current test procedures. Less expensive, faster, and more convenient, and at the same time, new compounds Provide sufficient information to make a quick decision on the safety of the                               Summary of the Invention   It is an object of the present invention to provide a gene in an in vitro or in vivo test system. A new approach to toxicity assessment based on examination of expression patterns or profiles It is to provide.   Another object of the invention is to provide a basic decision regarding the toxicological properties of chemicals, especially drug candidates. Is to provide a database to   It is a further object of the present invention to provide a gene expression protein in a selected tissue of a test animal. The purpose is to provide a way to analyze turns.   A still further object of the invention is to express differentially in response to exposure to a test compound. The purpose of the present invention is to provide a system for identifying genes to be identified.   The purpose of the present invention is to couple gene expression with short and long term toxicity in test animals. To provide a quick and reliable way to correlate.   Another object of the invention is to identify genes whose expression is predictive of detrimental toxicity. is there.   The present invention provides for expression in one or more selected tissues of an organism exposed to a test compound. Massively parallel sign sequencing of selected genes achievable these and other objectives by providing a method of I do. An important feature of the present invention is to align the sequence of portions of a number of different polynucleotides. Gene expression profile for selected tissue by determining in line Application of novel DNA sorting and sequencing methods that allow the formation of DNA. this Such profiles can be single or multiple to identify expression patterns that predict toxicity. Or can be compared to profiles from control organism tissues at multiple time points .   The sorting methodology of the present invention provides a minimal cross-reacting cell for oligonucleotides. Use oligonucleotide tags that are members of the kit. Such a set of The sequence of oligonucleotides is the same set of at least two nucleotides Is different from the sequence of other members. So each member of such a set -Double with the complement of any other member with less than two mismatches Cannot form a chain (or triplex). Complement of oligonucleotide tag of the present invention Is referred to herein as a “tag complement”, which is a naturally occurring nucleotide or non-natural nucleotide. Nucleotide analogs may be included. Preferably, the tag complement is a solid support Adhered to. When used with the corresponding tag complement, such oligos Nucleotide tags are used to sort polynucleotides, such as cDNA. Means are provided to enhance the specificity of hybridization.   Each sorted polynucleotide has a different polynucleotide It has oligonucleotide tags attached with different tags. less than As explained more fully, this condition is substantially less than the group of polynucleotides. By using a large tag repertoire, and tagged polynus Enough of tagged polynucleotides from a complete ensemble of nucleotides This is achieved by employing a small sample. After such sampling , The support and the group of polynucleotides are each When mixed under conditions that allow specific hybridization with the complement , The same polynucleotide sorts on a particular bead or region. The sorted groups of polynucleotides are then described in more detail below. `` Single-base '' or `` base-by-base '' sequencing It can be sequenced on a solid support by routine methodology.   In one aspect, the method of the invention comprises the following steps: (a) providing a test organism Administering the compound; (b) mRNA content from each of one or more tissues of the test organism. Extracting a group of offspring; (c) oligonucleotides to which each cDNA molecule of a separate group is attached CD from each group of mRNA molecules extracted from one or more tissues to have a Forming separate groups of NA molecules, wherein the oligonucleotide tags have the same minimum Selecting from a cross-hybridization set; (d) substantially eliminating All different cDNA molecules have different oligonucleotide tags attached Separately sampling each group of cDNA molecules; Each by specifically hybridizing it to its complement. Sorting of a group of cDNA molecules, wherein each complement comprises one or more Homogeneity of substantially identical complements in spatially distinct regions on the solid support (F) expressed for each of one or more tissues, attached as a group Sorted cD of each separate group to form frequency distribution of genes Determining the nucleotide sequence of a portion of each of the NA molecules; and (g) one or more The frequency distribution of the expressed genes in each of the above tissues was linked to the toxicity of the compound. Steps to be related.   An important aspect of the invention is the identification of genes whose expression is predictive of compound toxicity. . Once such genes are identified, they are used to reverse transcribe Yeast It can be used in conventional assays such as elementary polymerase chain reaction (RT-PCR).                             BRIEF DESCRIPTION OF THE FIGURES   FIG. 1 shows the generation of a minimally cross-hybridized set of oligonucleotides. 5 is a flowchart display of the algorithm of FIG.   FIG. 2 illustrates an apparatus for performing polynucleotide sequencing according to the present invention. .                                   Definition   As used herein with respect to oligonucleotide tags, a "complement" or Is the "tag complement", which means that the oligonucleotide tag specifically hybridizes , Oligonucleotides that form a perfectly matched duplex or triplex. Special In embodiments where the heterogeneous hybridization results in a triplex, the oligonucleotide The dotag can be selected to be either double-stranded or single-stranded. Thus the triplex Are formed, the term "complement" refers to two single-stranded oligonucleotide tags Includes either a strand complement or a single-stranded complement of a double-stranded Saori nucleotide tag That means.   As used herein, the term “oligonucleotide” includes Watson Quartz Rick-type base pairing, base stacking, Hoogsteen or reverse foo Modulation of monomer-monomer interactions such as Gusteen base pairing Deoxyribonucleic acid, which can specifically bind to a target polynucleotide by turns Nucleoside, ribonucleoside, its anomeric form, peptide nucleic acid (PNA), etc. And natural or modified monomers or linear oligomers of bonds . Usually, the monomers are linked by a phosphodiester bond or an analog thereof. To a small number of monomer units, for example, from 3 to 4 to tens of monomer units. To form a range of oligonucleotides. Oligonucleotide is "ATGCCTG" Whenever a nucleotide sequence is represented by a 5 '→ In the order of 3 ', and unless otherwise noted, "A" indicates deoxyadenosine , "C" indicates deoxycytidine, "G" indicates deoxyguanosine, and It is understood that "T" indicates thymidine. Analogue of phosphodiester bond Is phosphorothioate, phosphorodithioate, phosphoranilide, phosphate Foramidate and the like. Usually, four oligonucleotides of the invention However, these also include non-natural nucleotide analogs. May be included. Use oligonucleotides with natural or unnatural nucleotides When enzymatic processing is required, e.g. It is obvious to those skilled in the art that oligonucleotides consisting of nucleotides are required. Is.   An "exact match" for a duplex is any nucleotide in each strand Undergo Watson-Crick base pairing with nucleotides in other strands The poly- or oligonucleotide chains that make up the duplex have a double-stranded structure with each other Means to form. This term may also be used, deoxyinosine, A pair of nucleoside analogs, such as a nucleoside having a 2-aminopurine base Formation. With respect to triplex, this term means that the triplex is a perfectly matched double The strand and any nucleotides must have a perfectly matched duplex base pair and Hoog Or a third strand that undergoes an inverse Hoogsteen bond . Conversely, a "mismatch" in the duplex between the tag and the oligonucleotide is A nucleotide pair or triplet in a duplex or triplex Click and / or Hoogsteen and / or inverse Hoogsteen Means that it does not undergo binding.   As used herein, "nucleoside" includes, for example, Kornberg and As described in Baker, DNA Replication, 2nd Edition (Freeman, San Francisco, 1992) Include natural nucleosides, including 2'-deoxy and 2'-hydroxyl forms It is. "Analogs" for nucleosides include specific hybridization Except for what can be done, for example, Scheit, Nucleotide Analog (John Wiley, New York, 1980); Uhlman and Peyman, Chemical Reviews, 90: 543-584 (1990). Have a modified base moiety and / or a modified sugar moiety described in Synthetic nucleosides. Such analogs have enhanced binding properties Synthetic oligonucleotides designed to reduce complexity, increase specificity, etc. Leotide.   As used herein with respect to polynucleotides, "sequencing" or "sequencing" Determining the nucleotide sequence "includes the portion of the polynucleotide as well as the entire sequence. Determining column information. That is, the term includes the target polynucleotide Sequence comparison, fingerprinting, and target polynucleotides in otide Expression identification and sequencing of otides and nucleosides, usually each nucleoside The similarity level of information about is included. This term also refers to the target polynucleotide Identification, ordering of 1, 2, or 3 of the 4 types of nucleotides in the And location determination. For example, in some embodiments, the determination is By identifying the ordering and position of a single type of leotide, For example, cytosine in the target polynucleotide "CATCGC ..." The array is a binary code for "C- (not C)-(not C) -C- (not C) -C ..." For example, it is represented as "100101 ...".   As used herein, the term "complexity" refers to a group of polynucleotides. Means the number of different species of molecules present in the group.   As used herein, the term “gene expression profile,” which is equally used, ”And“ gene expression pattern ”are sampled from a group of tag-cDNA conjugates Means the frequency distribution of a part of the sequence of the cDNA molecule. Generally, part of the array is Some of them are long enough to uniquely identify the resulting cDNA. Preferably determined The total number of sequences to be performed is at least 1000; more preferably, the gene expression profile The total number of sequences determined by the tool is at least 10,000.   As used herein, a "test organism" is defined as an exposure to a test compound. Means any in vitro or in vivo system that provides a measurable response . Typically, the test organism is a mammalian cell culture, in particular, hepatocytes, kidney cells, A mammalian cell culture of a particular tissue, such as colony forming cells, or Test organisms are rats, mice, hamsters, guinea pigs, dogs, cats, rabbits , Pigs, monkeys and the like.                             Detailed description of the invention   The present invention provides for gene expression in selected tissues of a test organism exposed to a compound. Analyze changes in the current profile to determine how compounds are toxic provide. The present invention also relates to rapidly expressed and correlated with long-term or chronic toxicity. Provides a method for identifying a toxic marker consisting of an individual gene or group of genes Alternatively, it indicates that the compound has an undesirable reactivity. Genetic The current profile is extracted from the tissue of the test organism exposed to the compound being tested Generated by sequencing a portion of the cDNA molecule assembly from the mRNA obtained. As used herein, the term “tissue” refers to an in vitro tissue, such as a cell culture. With regard to the strain system, other than simply referring to a sample from the culture, Is used in a biological sense. The gene expression profile from the test organism is Tested against gene expression profiles from control organisms Determine differentially expressed genes in test organisms for exposure to the compound being tested . In both cases, the sequence information in the gene expression profile Obtained by parallel signature sequencing, which is performed in steps (c) to (f) of the method described above. Will be                                 Toxicity assessment   To design and perform toxicity tests in vitro and in vivo systems Procedures are well known, and Loomis et al., Loomis's Essentials of Toxicology, No. 4 Edition (Academic Press, New York, 1996); Echobichon, The Basics of Toxicity T esting (CRC Press, Boca Raton, 1992); Frazier, In Vitro Toxicity Testin g (Marcel Dekker, New York, 1992). Listed in the quist.   For toxicity testing, two groups of test organisms are usually used: one group is controlled And a single dose for the other group (for acute toxicity testing). Or several dose regimens (for long-term or chronic toxicity testing) Is administered. In most cases, tissue extraction as required by the method of the present invention Requires sacrifice of test animals, so control group and compound For both treatment groups, it was desirable to observe the kinetics of gene expression during the duration of the experiment. If sufficient to allow removal of animals to sample tissue Must be large.   In assembling a toxicity study, the compound being tested, the route of administration, and the Extensive guidance is provided in the literature for selecting appropriate test organisms for It is. Water or physiological saline (0.9% NaCl in water) is Selectable solutes for test compounds as they allow administration by different routes It is. If this is not possible due to solubility limits, propylene glycol is commonly used. Relies on the use of vegetable oils, such as corn oil or even organic solvents, commonly used There is a need. Whenever possible, use of suspensions of emulsifiers is recommended, except for oral administration. Should be avoided. Regardless of the route of administration, it is necessary to administer a given dose. The capacity is limited by the size of the animal used. Within and between groups of animals It is desirable to keep each dose volume uniform. Where rats or mice are used Where the volume administered by the oral route should not exceed 0.005 ml / gram animal. Absent. An aqueous or physiological saline solution is used for parenteral injection. Although the volume that can be tolerated is limited, such solutions are usually considered harmless. Can be Intravenous LD in distilled water in mice₅₀Is about 0.044 ml / g, and LD with isotonic saline₅₀Is 0.068 ml / gram mouse.   If a compound is to be administered by inhalation, a Special techniques are needed. Dose estimation is very complicated. This method is It usually involves the aerosolization or nebulization of a liquid containing the compound. Drugs tested If the agent is a liquid with a considerable vapor pressure, air is produced under controlled temperature conditions. It can be administered by passing through a solution. Under these conditions, the dose is From the volume of air inhaled, the temperature of the solution, and the vapor pressure of the drug involved. Is determined. Gas is metered from the reservoir. When particles of a solution are to be administered If the particle size is less than about 2 μm, the particles will not reach the terminal alveolar sac of the lung. Various devices and chambers, when administered by inhalation, may cause irritants or other It can be used to conduct studies to detect the effects of toxic endpoints. Move the drug The preferred method of administering to a product is by intubation or by incorporating the drug into the diet. Thus, it is by the oral route.   Preferably, metabolism, absorption, excretion, tissue storage, etc. in designing toxicity assessments Used by two or more species that treat test compounds as similar as possible to humans in terms of Should be done. Preferably, multiple doses or dosing schedules at various concentrations are Should be used to establish a dose-response relationship for toxic effects. So Preferably, the route of administration to the test animal is the same as the route of administration of the compound to humans Or as similar as possible. One route of administration to test animals The effects obtained are not empirically applicable to the effects of alternative routes of administration to humans. Absent. For example, food additives for humans are produced by mixing substances in the diet of test animals. Should be tested.   Acute toxicity testing consists of administering the compound to the test organism in one shot. this The purpose of such tests is to determine the symptomatology of the consequences of Determining the degree of lethality of an object. The first step involves compounding in a single species Is a series of range finding doses. This depends on the choice of route of administration, the choice Requires preparation of compound in a form suitable for administration by route and selection of appropriate species I do. Preferably, the initial acute toxicity study is based on its low cost, its availability, and Due to the availability of abundant toxicological reference data on Or with one of the mice. Long-term toxicity test spans 3-4 months The process of administering the compound to the test organism repeatedly, usually on a daily basis. Become. We come across two real factors that constrain the design of such tests: The selected route is appropriate for repeated administration without inducing adverse effects. The available routes of administration are limited. And second, blood Fluids, urine, and possibly other samples may induce significant harm in test animals No, and should be collected repeatedly. Preferably, the method of the present invention The offspring expression profile is based on traditional toxicological parameters as listed in the table below. Obtained with the measurement of the data:Hematology blood chemistry urine analysis Red blood cell count Sodium pH Total leukocyte count Potassium Specific gravity Differentiated leukocyte count Chlorine Total protein Hematocrit calcium sediment Hemoglobin carbon dioxide glucose                     Serum glutamine-pyruvate ketone                     Transaminase                     Serum glutamine-oxaloacetate bilirubin                     Transaminase                     Serum protein                     Electrophoresis                     Blood sugar                     Blood urea nitrogen                     Total serum protein                     Serum albumin                     Total serum bilirubin                 Oligonucleotide tags and tag complements   Oligonucleotide tags are used for minimal cross-hybridization of oligonucleotides. Members. The sequence of such a set of oligonucleotides is At least two nucleotides differ from the sequence of any other member of the same seven Become. Thus, each member of such a set has less than two mismatches Cannot form a duplex (or triplex) with the complement of any other member of. Book The complement of an oligonucleotide tag, referred to herein as a "tag complement", is naturally occurring. It may include nucleotides or unnatural nucleotide analogs. Preferably the tag The complement binds to the solid support. Such oligonucleotide tags are Molecules, especially polynucleotides, when used with a complementary tag complement. Specificity of hybridization for tracking, tracking, or labeling To provide a means for enhancing   The minimum cross-hybridization set for oligonucleotide tags and tag complements is The desired set size and minimal cross-hybridization To the extent required (or, in other words, increased specificity) Re Depending on the degree), can be synthesized either in combination or individually. . For example, a minimally cross-hybridized set has at least 4 nucleotides in each other. May consist of a set of individually synthesized 10-mer sequences that differ , With a maximum size of 332 (consisting of three nucleotides and in Appendix Ic If counted using a computer program as disclosed). There The minimally cross-hybridized set of oligonucleotide tags is also Are combined from subunits selected from the smallest intersection hybridisation set Can be assembled. For example, different from the other by at least 3 nucleotides The set of minimally cross-hybridizing 12-mers is a minimal set of three nucleotides that differ from each other. Assemble three subunits selected from a set of cross-hybridizing 4-mers Can be synthesized. Such an embodiment is described in 9^ThreeI.e. 729, 12 Providing the largest set of mer. Number 9 is the computer program in Appendix Ia. The number of oligonucleotides listed in grams, which are as in 10 mer Assumes that only three of the four different types of nucleotides are used . The computer programs in Appendix Ia-c are provided with predetermined inputs (eg, length, composition, (The difference in the number of nucleotides between members). The cut is described as “maximum”. In addition, the smallest intersection hybrid set is Such a calculated set may be formed from a subset.   The oligonucleotide tag can be single-stranded and single-stranded due to duplex formation. For specific hybridization to tag complement or by triplex formation Designed for specific hybridization to a double-stranded tag complement. Oligonucleotide tags can also be double-stranded and single-stranded by triplex formation. It can be designed for specific hybridization to strand tag complement.   When synthesized in combination, the oligonucleotide tags are preferably multiple Subunits, each subunit being 3-9 nucleotides long. Each subunit has the same minimal cross-hybridization Selected from the set. In such embodiments, the available oligonucleotides Tag number depends on the number of subunits per tag and subunit length I do. In general, the number is much less than the number of all possible sequences and the length of the tag That is 4 for an n nucleotide long tag.ⁿIt is.   The complement of the oligonucleotide tag attached to the solid support is For sorting polynucleotides from each mixture of nucleotides used. Oligonucleotide tag complements have identical sequences in specific regions. Unique in the alignment of synthetic sites on microbeads or on a single support as generated It is synthesized on the surface of a solid support, such as a target site. That is, in the case of beads Or in the case of each region, in the case of alignment, the surface of each support is a phase having a specific arrangement. It is derivatized by only one type of complement. Such beads or areas Group comprises a repertoire of complements with different sequences. Oligonucleotide tags And tag complement as used herein and the term "repertoire" Is the minimum crossover hybridization of the oligonucleotides that make up the tag in certain embodiments. A set of soy sets or a corresponding set of tag complements is meant.   Each sorted polynucleotide has a different polynucleotide Having an oligonucleotide tag attached, like having a different tag. Less than As explained more fully, this condition is substantially more than the group of polynucleotides. By using a large tag repertoire for Fully tagged polynucleotides from a complete ensemble of nucleotides Achieved by taking a small sample. After such sampling, The support and the group of polynucleotides are the respective complements and oligonucleotides When mixed under conditions that allow specific hybridization of the tags, One polynucleotide is sorted on a particular bead or region.   The nucleotide sequence of the oligonucleotide of the minimum crossover hybridization set is: As the source code is exemplified by the programs listed in Appendix Ia and Ib It is enumerated for convenience by a simple computer program. Appendix Ia Program Lamb minhx has a 4-mer subunit composed of three nucleotides Calculate all minimum intersection hybridization sets. The program tagN in Appendix Ib is List the longer oligonucleotides in the minimally cross-hybridizing set. Kind Similar algorithms and computer programs may be implemented in any embodiment of the invention. Easy to list oligonucleotides with minimal cross-hybridization set To Written. Table I below shows the minimum crossover hybridization for the indicated lengths. An indication of the size of the set of rigonucleotides and the number of nucleotide differences Provide a needle. The numbers were generated using the computer program described above.   For some embodiments of the invention, a very large repertoire of tags is required. If not required, the oligonucleotide tag of the minimally cross-hybridizing set , Can be synthesized separately. Contains hundreds to thousands or tens of thousands of oligonucleotides Sets can be synthesized directly by various parallel synthesis approaches, for example, No. 4,689,405; Frank et al., Nucleic Acids Res. earch, 11: 4365-4377 (1983); Matson et al., Anal. Biochem., 224: 110-116 (1995); Fodor et al., International Patent Application PCT / US93 / 04145; Pease et al., Proc. Natl. Acad. Sci., 91: 5022-5026 (1994); Southern et al. Biotechnology, 35: 217-227 (1994); Bren nan, International Application PCT / US94 / 05896; Lashkari et al., Proc. Natl. Acad. Scl., 92: 7912 -7915 (1995).   Preferably, the oligonucleotide tag of the invention has between 3 and 6 nucleotides And the same minimum crossing hybrid Is selected from the set. For oligonucleotides in this range, Members of a set such as Program.   The algorithm of FIG. 1 first sub-units the least-cross hybrid set. Define the characteristics of the parameters: length, number of base differences between members, and composition. For example, they consist of two, three and four bases . Table M_n, N = 1 consists of all possible sequences of a given length and composition (100) Is generated. First subunit S₁Is selected and i for the end of the table = N + 1 consecutive subunits S_iAnd (120). Consecutive sub The unit requires a mismatch that is a member of the smallest intersection hybridization set Whenever a new table M_{n + 1}In (125), step 120 is The inclusion of preselected subunits before passing is also omitted. For example, compare In the first set of_TwoIs S₁; In a second set of comparisons, M_ThreeIs S₁And And S_Two; In a third set of comparisons, M_FourIs S₁, S_Two, And S_ThreeIncluding It is. Similarly, Table M_jThe comparison at_jAnd M_jAll successive services in Buunit. Subunits are eliminated in a continuous pass by step 130 So each successive table M_{n + 1}Note that is smaller than the previous one. table M_nAfter every subunit of is compared to (140), the old table becomes the new table M_{n + 1}To Replaced, and the next round of comparison begins. Selected subunit S_iWhen Table M without consecutive subunits for comparison_nIs achieved, Word, M_n= M_{n + 1}If, the process stops (160).   Preferably, the minimally cross-hybridized set is any other sub-set in the set. Contains subunits that make nearly equal contributions to duplex stability as a unit . Thus, a perfectly matched duplex between every subunit and its complement chain Are nearly equal in stability. Guidance for selecting such a set is based on optimal PCR Published on choosing primers and calculating duplex stability Provided by the following techniques: for example, Rychlik et al., Nucleic Acids Research, 1 7: 8543-8551 (1989) and 18: 6409-6412 (1990); Breslauer et al., Proc. Natl. Acad . Sci., 83: 3746-3750 (1986); Wetmur, Crit. Rev. Biochem. Mol. Biol., 26: 2 27-259 (1991). For shorter tags, e.g., about 30 nucleotides or less, The algorithm described in Rychlik and Wetmur is preferred, and longer For example, for more than about 30-35 nucleotides or more, see Suggs et al., Pp. 683-693, Brow. n, ICN-UCLA Symp. Dev. Biol. Opened in 23 volumes (Academic Press, New York, 1981) The algorithm shown can be used conveniently. Obviously, the best within the scope of the present invention. Many approaches were used to design a set of small cross-hybridizing subunits. Switches are available to those skilled in the art. For example, base sticks with different terminal nucleotides Sub-units are assembled to minimize the effects of switching energy , Subunits having the same terminal nucleotide may be provided. Thus, Subunits are linked, the base stacks of all linked terminal nucleotides are The sum of the king energies is the same, thereby reducing the variability of the tag melting point Or eliminate.   Terminal word "words", shown in italics below, also indicate exact matches. As always formed between it and a similar terminal "word" on any other tag complement Can be added to each end of the tag. Such augmented tags have the following form: Do: Here, primed W 'indicates the complement. Tags whose ends match exactly All mismatched words are internal mismatches because they always form a heavy chain. Thereby reducing the safety of the tag-complement duplex with the other mismatched terminus. Decrease qualitative. Duplex with internal mismatch, same mismatch at end It is well known that they are significantly less stable than duplexes having   A preferred embodiment of a minimally cross-hybridizing set has four subunits. A set made up of three of the natural nucleotides. Less than As discussed in one minute, one of the nucleotides in the oligonucleotide tag The absence of the type is due to the use of the 5 'to 3' exonuclease activity of DNA polymerase. Thus, it allows the target polynucleotide to be loaded on a solid support. The following includes four nucleotides selected from the group consisting of A, G, and T Here is a representative minimal intersection hybrid set for each subunit: In this set, each member has three misses with the complement of every other member. Form a duplex with matched bases.   Additional representative minimal cross-hybridization sets are listed below in Table III. Obviously, the additional set is by substituting different groups of nucleotides Or by using a subset of a known minimal crossover hybridization set. And can be generated.                                 Table III         Representative minimal cross hybridizing set of 4 mer subunits   The oligonucleotide tags of the present invention and their complements may be synthesized by phosphoramidite chemistry. Using standard chemical reactions such as reactions, an automated DNA synthesizer, e.g., Applied B iosystems, Inc. (Foster City, California) Model 392 or 394 DNA / RNA Synt  It is conveniently synthesized on a hesizer and is disclosed, for example, in the following references: Beaucage And Iyer, Tetrahedron, 48: 2223-2311 (1992); Molko et al., US Patent No. 4,980,46. No. 0; Koster et al., US Pat. No. 4,725,677; Caruthers et al., US Pat. No. 4,415,732; No. 4,458,066; and No. 4,973,679. The resulting oligonucleotide is Except for being able to specifically hybridize, for example, phosphorothio Alternative chemical reactions that produce non-natural backbone groups such as ate, phosphoramidite, etc. , Can be used. In some embodiments, the tag is processed by the enzyme Or may contain naturally occurring nucleotides that allow manipulation However, the corresponding tag complement promotes the formation of a more stable duplex during sorting Non-natural nucleotide analogs, such as peptide nucleic acids, or analogs I can see.   If microparticles are used as the support, oligonucleotide tags and tag phases The complement repertoire is based on subunit-related synthesis by "split and mix" technology. Thus, for example, Shortle et al., International Patent Application PCT / US93 / 03418 or Lyt tle et al., Biotechniques, 19: 274-280 (1995). Simply put, synthetic A basic unit is a subunit of an oligonucleotide tag. Preferably, Suphoramidite chemistry is used, and the 3 'phosphoramidite ribonucleotide Are prepared for each subunit in the minimally hybridized set. For example, for the first set listed above, eight 4-mer 3'-phospho There is lamidite. The synthesis may be performed as disclosed in Shortle et al. To generate different oligonucleotide libraries using sid monomers Proceeded in a direct analogy using the techniques used in, for example, Telenius et al. , Genomics, 13: 718-725 (1992); Welsh et al., Nucleic Acids Research, 19: 5275. -5279 (1991); Grothues et al., Nucleic Acids Research, 21: 1321-1322 (1993); Har. tley, European Patent Application 90304496.4; Lam et al., Nature, 354: 82-84 (1991); Zuckerman Et al., Int. J. Pept. Protein Research, 40: 498-507 (1992). one In general, these techniques simply involve growing oligonucleotides during the coupling step. Requires application of a mixture of activated monomers to the tide. Preferably, oligo Nucleotide tags and tag complements are different types of terms used in the construction of tags. On a DNA synthesizer with many synthesis chambers greater than or equal to Is done. That is, preferably, there is a synthesis chamber corresponding to each type of word. You. In this embodiment, each synthetic channel is as if the word consisted of 5 nucleotides. The term is appended to each nucleotide so that there are five monomer couplings in the Be added. After the words have been completely synthesized, the synthetic support is removed from the chamber. , Mixed, and redistributed into the chamber for the next cycle of word addition. Subsequent embodiments may include, for example, phosphoramidite chemistry, where the monomer addition is high. Utilize the coupling yield.   The double-stranded form of the tag allows the complementary strands to be synthesized separately, followed by duplex formation By mixing under the following conditions. Alternatively, double-stranded tags One linked to a known oligonucleotide sequence that serves as a primer binding site It can be formed by first synthesizing a chain repertoire. Then the second strand Combine single-stranded repertoire with primers and extend with polymerase Are synthesized by This latter approach is described in Oliphant et al., Gene, 44. : 177-183 (1986). Such double-stranded tags can then be used in accordance with the present invention. Standard for sorting and manipulation of target polynucleotides Together with the target polynucleotide.   Enhanced, such as PNA or oligonucleotide N3 '→ P5' phosphoramidite When tag complements made from nucleotides with different binding characteristics are used , Sorting is 3 '→ 5' exo of DNA polymerase to make the tag single-stranded As an alternative to the “stripping” reaction using nuclease activity, D-loop between tags, including otide and its PNA or phosphoramidite complement Can be performed.   Oligonucleotide tags of the invention can be 12-60 nucleotides or base pairs in length. Range. Preferably, the oligonucleotide tag comprises 18 to 40 nucleotides. Or base pair length range. More preferably, oligonucleotide tags Ranges in length from 25 to 40 nucleotides or base pairs. Good and good From a more favorable number standpoint, these ranges can be expressed as:                                   Table IV             Number of subunits in tag in preferred embodiment SubunitMonomer in Nucleotides in oligonucleotide tags                   (12-60) (18-40) (25-40)     3 4-20 subunit 6-13 subunit 8-13 subunit     4 3-15 subunit 4-10 subunit 6-10 subunit     5 2-12 subunit 3-8 subunit 5-8 subunit     6 2-10 subunit 3-6 subunit 4-6 subunit More preferably, the oligonucleotide tag is single stranded and specific high Hybridization occurs by Watson-Crick pairing with the tag complement. You.   Preferably, the repertoire of single-stranded oligonucleotide tags of the invention is low Comprises at least 100 members; more preferably, the repertoire of such tags And at least 1000 members; and most preferably Pa The tree contains at least 10,000 members.                               Triple chain tag   In embodiments where specific hybridization results from triplex formation, Coding sequence follows the same principles as for double-strand forming tags; There are further restrictions on the choice of the cut sequence. In general, Hoogsteen type The third strand due to binding is homopyrimidine-homoprint in the double-stranded target Most stable with racks. Usually, the base triplet is T-A^*T or C-G^*C Motif (where "-" indicates Watson-Crick pairing and "^*Is Ho ogsteen type binding); however, other motifs are also possible is there. For example, Hoogsteen base pairing, depending on conditions and strand composition, can be a third Between the first strand (Hoogsteen strand) and the double-stranded purine-rich strand to which the third strand binds. Allows row and anti-parallel orientation. Triple-strand stability desired in certain embodiments Proper alignment, orientation, and conditions to maximize or otherwise adjust , Nucleoside type (eg, ribose or deoxyribose nucleoside Is used), base modification (for example, methylated cytosine, etc.) There are extensive guidelines in the literature to do this. See, for example, Roberts et al., Proc. Natl. Acad. S ci., 88: 9397-9401 (1991); Roberts et al., Science, 258: 1463-1466 (1992); Robert. s et al., Proc. Natl. Acad. Sci., 93: 4320-4325 (1996); Distefano et al., Proc. Natl . Acad. Sci., 90: 1179-1183 (1993); Mergny et al., Biochemistry, 30: 9791-9798 (1 991); Cheng et al. Am. Chem. Soc., 114: 4465-4474 (1992); Beal and Dervan Nucleic Acids Research, 20: 2773-2776 (1992); Beal and Dervan, J. Mol. Am. C hem. Soc., 114: 4976-4982 (1992); Giovannangeli et al., Proc. Natl. Acad. Sci. , 89: 8631-8635 (1992); Moser and Dervan, Science, 238: 645-650 (1987); McS. han et al. Biol. Chem., 267: 5712-5721 (1992); Yoon et al., Proc. Natl. Acad. Sc i., 89: 3840-3844 (1992); Blume et al., Nucleic Acids Research, 20: 1777-1784 (19 92); Thuong and Helene, Angew. Chem. Int. Ed. Engl. 32: 666-690 (1993); E scude et al., Proc. Natl. Acad. Sci., 93: 4365-4369 (1996)). Single stranded or Add a double-stranded tag to its single-stranded or double-stranded complement Conditions are well known (eg, Ji et al., Anal. Chem. 65: 1323-1328 (1). 993); Cantor et al., US Patent No. 5,482,836. The use of triple-stranded tags complements their complement 'Stripy' with the polymerase to expose the tag for annealing to the product It has the advantage that no "ging" reaction is required.   Preferably, the oligonucleotide of the invention using triplex hybridization Tide tags are double-stranded DNA, and the corresponding tag complement is single-stranded. Than Preferably, the range of pH stability of the triplex formed between the tag and its complement is broadened. 5-methylcytosine is used instead of cytosine in the tag complement You. Preferred conditions for forming triplex are well disclosed in the above references. You. Briefly, hybridization is less than 5.5 (or 5-methyl Strong salt solution (eg, 1.0 M NaCl, 1.0 M (Eg, potassium acetate). Hybridization temperature depends on tag length and Depending on composition; however, for longer 18-20 martags, hybridization at room temperature Distillation is appropriate. Washing is carried out at room temperature with a non-strong salt solution, e.g. mM sodium acetate, 100 mM MgCl_TwoPH 5.8. Tags are similar at pH 9.0 Can be eluted from these tag complements by incubation in salt solutions. You.   The minimum cross-hybridization set of oligonucleotide tags that form a triplex is Generated by the computer program of Attachment Ic or a similar program obtain. A representative set of double-stranded 8-mer languages is listed below in capital letters and the corresponding Complements are listed in lower case. Each of these words is represented by three base pairs Different from each of the other words in the set.Table V Exemplary minimal cross-hybridization set of double-stranded 8-mer tags                                   Table VI     Repertoire size of various double-stranded tags that form triplex with tag complement   Preferably, the repertoire of double stranded oligonucleotide tags of the invention is low. Contains at least 10 members; more preferably, the repertoire of such tags is small. Includes at least 100 members. Preferably, the word is a combination of two For a strand oligonucleotide tag, the length is between 4 and 8 nucleotides And the oligonucleotide tag is between 12 and 60 base pairs in length. Than Preferably, such tags are between 18 and 40 base pairs in length.                                 Solid support   Solid supports for use in the present invention include microparticles, beads, and membranes, slides , Plates, micromachined chips, etc. of It can have a form. Similarly, the solid supports of the present invention can be glass, plastic, silicon Con, alkanethioate derivatized gold, cellulose, low and high crosslinked polices A wide variety of compositions can be included, including styrene, silica gel, polyamide, and the like. Preferably, a uniform code of complementary sequences, each of the same tag (and not the other). Which of the separate particle groups is used to have Or a uniform coating of the complementary sequence on the same tag (and not others) Or a small number of supports, each having a spatially distinct area, each containing a group Is used. In later embodiments, the area of the region may vary according to the particular application. Usually, the area is several μm^Two, For example, 3-5 to several hundred μm^Two, For example, 100-500 The area is up to. Preferably, such a region is an event in an adjacent region ( For example, the signal generated by fluorescence Thus, they are spatially separated so that they can be analyzed. In some applications, for example, For simultaneous sequence analysis or to bring the tagged molecules separately proximally And have regions with a uniform coating of more than one tag complement. May be desired.   See, for example, Lund et al., Nucleic Acids Research, 16: 10861-10880 (1988); n et al., Anal. Biochem., 189: 40-50 (1990); Wolf et al., Nucleic Acids Research, 15 : 2911-2926 (1987); or Ghosh et al., Nucleic Acids Research, 15: 5353-5372 (19 As disclosed in 87), tag complements can be used with solid supports, The complement is synthesized there, or synthesized separately, and attached for use Can be done. Preferably, the tag complement is synthesized on the same solid support and used together. Used, which may include various forms and may include various connecting portions. Tag phase Such a support in the area where a uniform group of complement is synthesized may be finely divided or aligned, Alternatively, it may include a matrix. A wide variety of particulate supports can be used in the present invention. , This is controlled pore glass (CPG), highly crosslinked polystyrene, acrylic copoly Mer, cellulose, nylon, dextran, latex, polyacrolein And microparticles produced from such sources and are disclosed in the following exemplary references: Meth. En zymol., Section A, pp. 11-147, vol. 44 (Academic Press, New York, 1976); U.S. Pat. No. 4,413,070; and No. 4,046,720; and Pon, Chapter 19, Agrawal, Methods in Molecular Biology, Volume 20, (Humana Press, Totowa, NJ , 1993). Microparticle supports include commercially available nucleoside derivatized CPG and polystyrene Beads (eg, available from Applied Biosystems, Foster City, CA); Embedded magnetic beads; polystyrene grafted with polyethylene glycol (eg, For example, TentaGel^TM, Rapp Polymere, Tubingen Germany). material Support characteristics such as, porosity, size, shape, etc., and the connections used The choice of the minute type depends on the conditions under which the tag is used. For example, a series of enzymes In applications involving continual processing, steric hindrance of the enzyme is minimized and Supports and linkers that facilitate access are preferred. Of the most appropriate particulate support Other important factors to consider in selection are size uniformity, synthetic support As well as known degrees of surface area, and optical properties. As a more complete explanation below, when handling large numbers of beads on a surface, Beads provide a mechanical advantage.   An exemplary linking moiety for attaching and / or synthesizing a tag on a microparticle surface is Pon et al., Biotechniques, 6: 768-775 (1988); Webb, U.S. Patent No. 4,659,774; Ba. rany et al., International Patent Application PCT / US91 / 06103; Brown et al. Chem. Soc. Commun., 1989 : 891-893; Damha et al., Nucleic Acids Research, 18: 3813-3821 (1990); Beattie et al. , Clinical Chemistry, 39: 719-722 (1993); Maskos and Southern, Nucleic Ac. ids Research, 20: 1679-1684 (1992).   As described above, the tag complement can also be uniformly coated with the tag complement. Can be synthesized on a single (or a small number) of solid supports to form an array of regions . That is, the same tag complement is synthesized in each region in such a sequence. This A technique for synthesizing such sequences is described in McGall et al., International Patent Application PCT / US93 / 03767. Pease et al., Proc. Natl. Acad. Sci., 91: 5022-5026 (1994); Southern and Mas. kos, International Patent Application PCT / GB89 / 01114; Maskos and Southern (supra); Southern Et al., Genomics, 13: 1008-1017 (1992); and Maskos and Southern, Nucleic Ac. ids Research, 21: 4663-4669 (1993).   Preferably, the invention relates to a microparticle uniformly coated with the complement of the same tag sequence. Done with particles or beads. Oligonucleotides on microparticle support and its surface Methods for covalently or non-covalently bonding a compound are exemplified by the following references: As is well known: Beaucage and Iyer (supra); Ed. Gait, Oligonucleotide Synthesis: A Practical Approach (IRL Press, Oxford, 1984); Literature. In general, the size and shape of the microparticles is not important; For example, one to two to several hundred, for example, 200 to 1000 μm diameter fine particles having a size range preferable. Because these are oligonucleotides with minimal reagent and sample usage Whether to facilitate the construction and manipulation of large repertoires of pseudotags It is.   In some preferred applications, commercially available controlled pore glass (CPG) or poly Styrene supports are used as solid supports in the present invention. Such a support The carrier has a base-labile linker and the first nucleotide to be attached. Will be available. For example, Applied Biosystems (Foster City, CA). Preferably Fine particles having a pore size of 500-1000 Angstroms are used.   In other preferred applications, non-porous microparticles are used for their optical properties. This is useful when tracking large numbers of particles on a flat support such as a microscope slide. Can be used to advantage. Particularly preferred non-porous microparticles are Bangs laboratories (Car Glycidal methacrylate (GMA) beads available from mel, IN). this Such microparticles are useful in a variety of sizes and combine tags or tag complements. Derivatized with various linking groups. Preferably, tagged microparticles 5 μm diameter GMA beads were used for a large number of parallel manipulations It is.       Tag to polynucleotide for sorting on solid support Attach process   Important aspects of the invention include, for example, polynucleotides from cDNA libraries, Grouping, and polynucleus to which each microparticle or region is attached to only one species Polynucleotides to discrete regions on a solid support to have substantial nucleotides This is the adhesion of tide. The purpose is to attach virtually all different polynucleotides. Achieved by ensuring that you have different tags worn. Next, Conditions include taking a sample of the complete set of tag-polynucleotide conjugates for analysis. Achieved by taking. (Manipulated or analyzed twice at two different locations Only obtain the same polynucleotide that is It is acceptable to have a tag. ) Such a sampling is After being attached to a polynucleotide, it becomes apparent--e.g., From a larger mixture to a smaller volume. By taking the amount--which can be done, or Can be done essentially as a secondary effect of the technology used to Or sampling is both obvious and as an inherent part of the processing process Either can be done.   Preferably, a cDNA library in which substantially all different cDNAs have different tags In building a tree, the complexity or number of different tags Uses a tag repertoire that greatly exceeds the total number of mRNAs extracted from tissue samples. It is. Preferably, the complexity of the tag repertoire is low for the population of polynucleotides. At least ten times; and more preferably, the complexity of the tag repertoire is At least 100 times the nucleotide population. The following is a complete review of an exemplary 9-word tag. About cDNA library constructs using primer mixtures containing parts Is disclosed. Such a mixture of tag-containing primers would⁹Ma Or 1.34 × 10⁸With the complexity of Winslow et al., Nucleic Acids Research, 19: 325 As shown in 1-3253 (1991), mRNA for library construction is 10-100 It can be extracted from a small number of mammalian cells. Single mammalian cells can be prepared using standard techniques. About 3.4 × 10^FourAbout 5 × 10 of different types of mRNA molecules^FiveAbout 100 cells, including copies MRNA from, ie (theoretically) about 5 × 10⁷mRNA molecules can be isolated. This number Comparing with the complexity of the primer mixture, without any additional steps, In addition, mRNA is converted to cDNA with full efficiency (less than 1% is more accurate). The cDNA library construction protocol can be used to Shows that a population containing only 37% of the total number results. That is, With no clean sampling steps, this protocol is compatible with the tag repertoire. Essentially produces material containing 37% or less. Double stranded under these conditions Is about 5%, which is within the preferred range. From 10 cells MRNA has a fraction of only 3.7% of the tag repertoire sampled Reduced, and all processing steps are assumed to be performed at 100% efficiency You. In fact, the efficiency of the processing steps to construct a cDNA library is very And the "approximate guide" is that there are 10 good libraries⁶From mammalian cells About 10 from the extracted mRNA⁸CDNA clone should be included.   Larger use of mRNA in the above protocols, or generally polynucleotides Higher usage of tides (where the number of such molecules is more complex than the tag repertoire) Tag-polynucleotide conjugate mixture, Potentially includes pairing and mRNA or polynucleotide types. like this In cases where obvious sampling is the starting mixture of the tag-polynucleotide conjugate This can be done by removing the sample volume after serial dilution of Required dilution Depends on the amount of starting material and the efficiency of the processing step, which are easily Presumed.   mRNA is 10⁶Cells (about 0.5 μg of poly (A)⁺Extracted from (corresponding to RNA) , And if the primers are present in about a 10-100 fold concentration excess--a representative protocol Col, for example, Sambrook et al., Molecular Cloning, 2nd edition, page 8.61 [10 mg / mL. μL 1.8 kb mRNA is about 1.68 × 10^-11Equal to 10 mol / mL at 1 mg / mL -Primer is about 1.68 x 10^-9Equivalent to a mole). The total number of tag-polynucleotide conjugates in the library is simply the start of mRNA Less than or equal to about 5x10, including tag-polynucleotide conjugate^{1 1} Vector--this is also the step in cDNA construction--first strand synthesis, second strand Strand synthesis, ligation to a vector--is assumed to occur with full efficiency. (This is a very conservative estimate). The actual number is significantly lower.   A sample of n tag-polynucleotide conjugates is randomly taken from the reaction mixture. If possible-use the same tag, as can be done by taking a sample volume The probability of collecting a conjugate having a Poisson distribution, P (r) = e^{- λ}(λ)^rDescribed by / r Where r is the number of conjugates with the same tag and λ = np, where Here, p is a probability of a predetermined tag to be selected. n = 10⁶And p = 1 / (1.34 × 10⁸ ), Then λ = 0.00746 and P (2) = 2.76 × 10^-FiveIt is. Therefore, 1,000,000 A sample of molecules will yield the expected number of duplicate wells within the preferred range. this Such a sample is easily obtained as follows: 5 × 10¹¹mRNA as insert 5 × 10 with tag-cDNA conjugate¹¹Completely converted to vector and 5 × 10¹¹ Assume that the vector is in a reaction solution having a volume of 100 μl. Four 10-fold reams Serial dilutions include 10 μl from the original solution containing 90 μl of the appropriate buffer (eg, TE). Can be carried out by transferring to a container. This process is 5 × 10^Five Repeat for three additional dilutions to obtain a 100 μl solution containing vector molecules. Can be returned. A 2 μl aliquot from this solution was used as an insert for tag-cDNA 10 including conjugates⁶Get the vector. This sample is then used to competent host cells. Amplified by straight forward transformation, followed by culturing. It is.   Of course, as mentioned above, the steps in the above process do not proceed at full efficiency. No. In particular, vectors are used to amplify samples of tag-polynucleotide ligatures. When used, the process of transforming a host is very inefficient. Usually Up to 1% of the vector is adopted and replicated by the host. Therefore, For such a method of width, even less dilution is 10⁶Obtain a sample of the conjugate Needed for   Oligonucleotide tag repertoire includes primers containing tag sequences, etc. Uses a number of methods, including direct enzymatic ligation, e.g., amplification by PCR. In a manner, it can be bound to a population of polynucleotides. The first ligating process is simply The tag is such that one tag is generally attached to many different polynucleotides. -Produces a very large population of polynucleotide conjugates. But as mentioned above By taking a sufficiently small sample of the conjugate, the "double" (ie, The probability of obtaining the same tag on two different polynucleotides) can be ignored You. In general, the larger the sample, the greater the probability of obtaining a double. But Thus, trade-off is a large sample of tag-polynucleotide conjugates. Selecting--this can be, for example, a target polynucleotide in a shotgun sequencing operation. Ensure proper coverage of nucleotides or proper representation of fast-changing mRNA pools -And select a small sample that ensures that the least number of doubles are present Exists between things. In most embodiments, the presence of the double is noisy. Only provide additional sources or, in the case of sequencing, multiple fireflies. Light sig Scanning and signal It only adds a small complexity in processing.   As used herein, a tag is attached to a molecule, especially a polynucleotide. The term “substantially all” in terms of -The statistical properties of the sampling procedure used to obtain the population of molecular conjugates Means to reflect. Substantially no change in the actual percentage of tag-molecule conjugate The meaning of all depends on how the tags are used. Preferably, the nucleic acid sequence For the decision, substantially all are at least 80% polynucleotides It means that the tide has a unique tag attached. More preferably, less At least 90% of polynucleotides have unique tags attached Means Even more preferably, at least 95 percent of the polynucleotides Means that the tag has a unique tag attached. And most preferably , At least 99% of polynucleotides have unique tags attached That means.   Preferably, the population of polynucleotides consists of messenger RNA (mRNA) If so, the oligonucleotide tag is preferably a set comprising the complement of the tag sequence. Can be attached by reverse transcribing the mRNA with the primers. Such a ply An exemplary set of mer may have the following sequence (SEQ ID NO: 1): Here, "W, W, W, C]₉"" Indicates the orientation of nine subunits of each four nucleotides. Represents the sequence of the oligonucleotide tag, and “W, W, W, C] is the subunit listed above. Represents a knit arrangement (ie, “W” represents T or A). The underlined array is If one is used, polynucleotides from attachment to a solid support via biotin Identify any restriction endonuclease sites that can be used to release the tide I do. For the above primers, the complement attached to the microparticles has the following form Do:   After reverse transcription, mRNA is removed (eg, by RNase H digestion) and cDNA The second strand is synthesized using, for example, a primer having the following form (SEQ ID NO: 2). RU: Wherein N is any one of A, T, G, or C; R is a purine-containing nucleic acid. Otide and Y is a pyrimidine-containing nucleotide. This particular program The primer is used to bind the resulting double-stranded DNA (Sal I site to a vector (eg, Bam B) which facilitates cloning into (with HI and Xho I sites) Creates a st Y1 restriction site. After Bst Y1 and Sal I digestion, exemplary conjugates are: With form: The polynucleotide-tag conjugate is then converted using standard molecular biology techniques. Can be treated. For example, the conjugate above--which is actually a mixture-- Commercially available cloning vectors (for example, Stratagene Cloning System (La Jolla, CA)); transfected into a host (eg, a commercially available host bacterium). It is then cultured to increase the number of conjugates. Then clonin Vectors are constructed using standard techniques (eg, Sambrook et al., Molecular Cloning, 2nd edition). (Cold Spring Harbor Laboratory, New York, 1989)). . Alternatively, a suitable adapter such that the conjugate population can be increased by PCR And primers can be used.   Preferably, when a ligase-based method of sequencing is used, Bst Y1 and And the SalI digested fragment had the following single copy restriction site (SEQ ID NO: 3) Cloned into a Bam HI- / Xho I-digested vector with: This allows the start of the sequencing process, discussed more fully below, Add an I site.   Tags are ligated to the cDNA of an existing library using standard cloning methods. Can be combined. cDNAs are excised from those existing vectors, isolated and then Is linked to a vector containing the repertoire of tags. Preferably, the tag-containing vector The two restriction enzymes so that the excised cDNA can be ligated in a predetermined orientation. It is linearized by cutting. The concentration of the linearized tag-containing vector is CDNA inserts so that ligation provides the original sampling of tags Is substantially above the concentration.   A common method for exposing single-stranded tags after amplification is to use T4 DNA polymerase. Digestion of target polynucleotide-containing conjugate with 5 '→ 3' exonuclease activity Such as enzymes or enzymes. Single deoxynucleoside triphosphate When used in the presence, the complement of a single deoxynucleoside triphosphate is Until the template strand is reached, such polymerases Cleave nucleotides from the 3 'recessed end present on the non-template strand. When such nucleotides are achieved, the elongation activity of the polymerase increases 5 'because the sex adds nucleotides at a higher rate than it removes nucleotides. → 3 'digestion ends efficiently. As a result, for loading on solid supports In addition, single-stranded tags composed of three nucleotides are easily prepared.   This technique also preferentially methylates the internal Fok I site of the target polynucleotide. At the end of an unmethylated polynucleotide. Leave one Fok I site. First, the terminal Fok I site is associated with deoxycytidine triphosphate. Both are made single-stranded using a polymerase. Then the double strand of the fragment The moiety is methylated and then the single-stranded ends are replaced with all four nucleoside triphosphates In the presence of DNA polymerase, thereby regenerating the Fok I site. joy Anyway, this procedure can be generalized to endonucleases other than Fok I.   Oligonucleotide tags (eg, as described above, make these single-stranded) After being prepared for specific hybridization) Otides favor the formation of a perfectly matched duplex between the tag and its complement. Under the same conditions as microparticles containing the complementary sequence of the tag. In the literature these conditions There are extensive guidelines for producing Example references that provide such guidance Wetmur, Critical Reviews in Biochemistry and Molecular Biology, 26 : 227-259 (1991); Sambrook et al., Molecular Cloning: A Laboratory Manual, 2nd edition. (Cold Spring Harbor Laboratory, New York, 1989). Preferred Alternatively, hybridization conditions should be such that only perfectly matched sequences are stable duplexes. Is sufficiently stringent to form Under these conditions, it Polynucleotides specifically hybridized via these tags are attached to the microparticles. Can be ligated to the attached complementary sequence. Finally, the microparticles are washed and linked Polynucleotides with missing and / or mismatched tags are removed You.   When CPG microparticles that are conveniently used as a synthetic support are used, the microparticle surface The density of the above tag complement is typically the complement required for some sequencing operations. Greater than. That is, continuous processing of attached polynucleotides with various enzymes. Closely spaced polynucleotides in sequencing approaches that require Tend to block the access of relatively bulky enzymes to polynucleotides obtain. In such cases, the polynucleotide preferably has a tag complement, Significant excess over polynucleotides (eg, 10: 1 to 100: 1, or more) Is mixed with the microparticles so that This is due to the polynucleotide on the surface of the microparticles. Ensure that the density of the tide is not high enough to inhibit access of the enzyme. Preferably The average spacing between polynucleotides on the surface of the microparticle is about 30 to 100 nm. Standard CPG support and Ballotini beads (solid glass support type) Guidance for selecting the ratio can be found in Maskos and Southern, Nucleic Acids Research, 20: 167. See 9-1684 (1992). Preferably, for sequencing applications, a range of 20-50 μm is used. Standard CPG beads with a diameter of^FiveLoaded with a polynucleotide, and GMA beads having a diameter in the range of 5-10 μm can be used for tens of thousands of polynucleotides, for example, 4 × 1 0^Four~ 6 × 10^FourLoaded in.   In a preferred embodiment, the tag complement is synthesized in combination on a microparticle; Therefore, at the end of the synthesis, the sample loads the tagged polynucleotide To obtain a complex mixture of microparticles to be collected. What is the size of the particulate sample? This includes the size of the repertoire of tag complements, The nature of the device used to observe the particles that have fallen-for example, their capabilities, the same tag Resistance to multiple copies of the microparticles with complement (i.e., "bead dub And so on. The table below shows the particle sample size, particle diameter, and various The approximate physical size of a packed array of particulates of different diameters Provide guidance on what to do. The probability that the sample of microparticles contains the given tag complement or is present in multiple copies The rates are described by the Poisson distribution, as shown in the table below. Highly specific sorting and panning   The kinetics of selection is based on the hybridization of the oligonucleotide tag to the tag complement. Of the tag in the hybridization reaction. Depends on complexity. Therefore, trade off is a tradeoff between sorting speed and tag complexity. Increase in the sorting speed is associated with the hybridization reaction Can be achieved at the expense of reducing the complexity of the tag. As explained below, The effect of trade off can be recovered by "panning".   The specificity of hybridization can be obtained by collecting sufficiently small samples. So that a high percentage of tags in the sample are unique and in the sample The nearest neighbors of substantially all of the tags differ by at least two words. This latter Conditions are about 0.1% or less of the size of the repertoire used Attained by taking a sample containing multiple tag-polynucleotide conjugates. Can be For example, if the tag is constructed with 8 words selected from Table II, then 8⁸(S That is, about 1.67 × 10⁷Repertoire) tags, and tag complements are generated. Up For a library of tag-cDNA conjugates as described above, a 0.1 percent sample would be approximately 16,7 It means that there are 00 different tags. Repertoire equivalent of fine particles (sun In this example, 1.67 × 10⁷If this is loaded directly on the A low-density subset of the charged microparticles is loaded. Of loaded particulates Density--for example, for more efficient sequencing--sampled tag-cDNA binding Used to separate loaded particulates from unloaded particulates It can be increased by performing a "panning" step that is performed. Therefore, above In the example above, even if the "0.1 percent" sample contains only 16,700 cDNA, The sampling and panning steps can be performed as many loaded granules as desired This can be repeated until children are accumulated.   The panning step can be performed by providing a sample of the tag-cDNA conjugate. , Each of which captures at the opposite end, or distal, of the oligonucleotide tag Including parts. Preferably, the capture moiety is a tie that can be released from the tag-cDNA conjugate. Therefore, the tag-cDNA conjugate was sequenced using the single nucleotide sequencing method. Can be determined. Such moieties may include biotin, digoxigenin, or similar In , Triplex binding regions, and the like. Preferably, such capture moieties are bioactive Contains chin component. Biotin is attached to the tag-cDNA conjugate by a number of standard techniques. Can be worn. A suitable adapter containing a PCR primer binding site is used for tag-cDNA binding. When attached to the body, biotin is biotinylated during amplification after sampling. Can be attached by using a primer. Alternatively, tag-cDNA binding If the body is an insert in a cloning vector, biotin Excising the tag-cDNA conjugate by enzymatic digestion and isolation; and Protrude distal to tag with DNA polymerase in the presence of biotinylated uridine triphosphate After filling the chain, it can be attached.   After the tag-cDNA conjugate is captured, it is cleaved by reduction (eg, Herman Et al., Anal. Biochem., 156: 48-55 (1986)) or photochemically cleaved (eg, Olejnik et al., Nucleic Acids Research, 24: 361-366 (1996)) or PCR By chemical bonds that are enzymatically cleaved by introducing restriction sites into the immer Can be released from the biotin moiety in a number of ways. The latter embodiment is based on the tag described above. -Can be exemplified by considering a library of polynucleotide conjugates : The following adapters are used to terminate these fragments to allow amplification by PCR. Can be concatenated to: Here, “ACTAGT” is the Spe I recognition site (this is the attachment site prepared for single nucleotide sequencing). X and Z are annealing of respective primers And nucleotides selected to have approximately the same dissociation temperature. Adap Ligation and amplification by PCR using biotinylated primers followed by ligation The body tag is single-stranded by the exonuclease activity of T4 DNA polymerase. And the conjugate, along with the attached tag complement, a sample of the microparticles (eg, Repertoire equivalent). (To minimize tag mis-adhesion ) After annealing under stringent conditions, the conjugate is preferably a tag phase. The microparticles linked to and loaded with complement can be avidinated magnetic beads, or Separation from unloaded particulates by capture with a similar capture technique.   Returning to the example, this process works for about 10,500 (= 16,700 × 0.63) with various tags. ), Resulting in the accumulation of loaded microparticles, which was cut by Spe I Can be released from A new sample of microparticles and tag-cDNA conjugate 4 to 5 × 10^FivecDNA uses the released microparticles Can be accumulated by pooling. The pooled microparticles are then It can be sequenced simultaneously by base sequencing techniques.   Determine how many times the sampling and panning steps are repeated-or More generally, determining the number of cDNAs to be analyzed depends on the purpose. Purpose Monitor, for example, changes in the abundance of relatively common sequences that make up more than 5% of the population. If smaller, a relatively small sample, i.e., a small fraction of the entire It may allow for a statistically significant estimate of the relative richness. On the other hand, for example, If you want to monitor the abundance of rare sequences that make up less than 0.1% of the population , Large samples are needed. In general, sample size and relative abundance based on sample There is a direct relationship between the reliability of wealth estimates. Reliable statistical view Extensive literature describes how to determine the appropriate sample size to create a stack. What There are guidelines, for example, Koller et al., Nucleic Acids Research, 23: 185-191 (1994); Good, Biometrika, 40: 16-264 (1953); Bunge et al. Am. Stat. Assoc., 88: 364- 373 (1993). Preferably, 3.0-3.5 × 10^Four10 of different arrays of^Five~Ten⁸German Changes in gene expression based on analysis of a series of cDNA libraries containing established clones At least 10 to monitor^FourSequence samples are used to analyze each library. To be accumulated. More preferably, at least 10^FiveSamples of the array are Accumulated for analysis of L .; and most preferably at least 5 × 10^FiveArray Samples are accumulated for analysis of each library. Or sampling The number of sequences performed is preferably 0.1% with a 95% confidence limit of about 0.1% of the population size. Sufficient to estimate the relative abundance of sequences present at a frequency in the range of is there.                           Single base DNA sequencing   The present invention is described, for example, in Hultman et al., Nucleic Acids Research, 17: 4937-4946 (1989). ) Can be used in conventional methods of DNA sequencing. But parallel , Or for simultaneous sequencing of multiple polynucleotides, a similar size DN Electrophoretic separation of A fragments is also a separate analysis as in peptide sequencing DNA sequencing methodology that does not require analysis of truncated nucleothiol by procedures Is preferred. Preferably, the methodology is based on successive cycles of processing and detection. Allows for the stepwise identification of nucleotides in a sequence, usually one at a time. I do. Such a methodology is referred to herein as a "single base" sequencing method. A single base approach is disclosed in the following references: Cheeseman, US Pat. 02,509; Tsien et al., International Application WO 91/06678; Rosenthal et al., International Application WO 93/21340; Canard et al., Gene, 148: 1-6 (1994); and Metzker et al., Nucleic Acids Research, 22: 4259-4267 (1994).   Suitable for use in the present invention and requires electrophoretic separation of DNA fragments A "single base" method of DNA sequencing is described in International Application PCT / US95 / 03678. You. Briefly, the method comprises the following steps: (a) forming the linked complex Connecting a probe to the end of a polynucleotide having a protruding strand to form The probe has an overhanging strand that is complementary to the polynucleotide strand, and The step wherein the lever probe has a nuclease recognition site; (b) the linked complex Removing the unligated probe from (c) the identity of the ligated probe Identifying one or more nucleotides in the overhanging strand of the polynucleotide by gender (D) cleaving the complex linked by nuclease; and (e) polynucleic acid. Steps (a) through (d) until the nucleotide sequence of the reotide or a portion thereof is determined. Repeating the process.   A single signal-generating moiety, such as a signal fluorescent dye, can be used for parallel sequencing operations. Various spatially addressable (addressab) le) sequencing several different target polynucleotides attached to a solid support It is used when This is due to the large number of target polynucleotides in various microparticles. Achieved by providing four sets of probes that are applied sequentially to the ochides Can be achieved. A representative set of such probes is shown below: Here, each of the listed probes is therefore a 3 ' The identity of the tide is fixed, and the other positions of the overhanging strand are every three nucleotides of the nucleotide. Be filled by analogs with reduced permutation or complexity Una, 4^Three= Represents a mixture of 64 oligonucleotides. The listed probes also , "T^*Single-stranded polypeptides with a signal-generating moiety attached to terminal thymidine Shown with Re-T tail. "D" in unlabeled probe is linked 3 'to prevent ligation of unblocked or unlabeled probes - Hydroxyl indicates absence. Preferably, such a 3'-terminal nucleotide is Dideoxynucleotide. In this embodiment, the probes in set 1 Initially applied to multiple target polynucleotides and labeled 3 'end of probe A target polynucleotide with a thymidine complementary to the terminal adenosine will be ligated. And treated with ligase. Unlabeled probes should be Applied at the same time to minimize Along with probes ending in "A" The location of the target polynucleotide forming the ligated complex is Identified by the signal generated by the resulting label. After cleaning and cutting , Set 2 probes are applied. In this case, with a probe ending in "C" Target polynucleotides forming linked complexes are identified by position . Similarly, the probes of sets 3 and 4 were applied and a positive signal Is identified. This process of applying four sets of probes sequentially Continue until the desired number of nucleotides has been identified in the target polynucleotide . Obviously, one skilled in the art will recognize, for example, protruding strands of various lengths, Various parts to block the ligation of the probe, to label the probe A similar set of probes that can have many changes, such as having various means of Can build.               Apparatus for sequencing populations of polynucleotides   It is an object of the present invention to provide for specific hybridization of a tag and its complement. The same molecules, especially polynucleotides, on the surface of the microparticles. Is Rukoto. Once such sorting is performed, the molecules that are performed there Or the presence of the operation, whether the particulates are detected separately or in "batch", Depends on the nature of the tagged molecule, such as whether repeat measurements are desired And can be detected in many ways. Typically, the sorted molecules are For example, in drug development, exposure to ligands for binding or Subject to chemical or enzymatic processes in polynucleotide sequencing . In both of these uses, such events or processes in large numbers of microparticles It is often desirable to observe the signals corresponding to the processes simultaneously. Saw Te Microparticles with loaded molecules (referred to herein as "loaded" microparticles). For example, as demonstrated by Lam et al. (Supra). Can be adapted for concurrent operations.   Preferably, the light producing signal, e.g., chemiluminescence, fluorescence, etc., is In any case used to detect process, the loaded microparticles International Patent Applications PCT / US91 / 09217, PCT / NL90 / 00081, and PCT / US95 / 01886 For experiments with scanning systems, a flat substrate, e.g. Spread out on a slide. The scanning system scans the substrate reproducibly Scanning and defining the position of each particle in a given area by a coordinate system. Should be possible. For polynucleotide sequencing applications, It is important that the location identification be repeatable in successive scanning steps.   Such scanning systems include commercially available components (eg, one or more photoelectric A photomultiplier tube, or alternatively, a CCD array, and, for example, to excite and collect fluorescent signals Detection system including suitable optics for collection and sorting X-y conversion table controlled by digital computer used with ). In some embodiments, a confocal lens based system is desired Can be done. A typical scanning system suitable for use in four-color sequencing is This is illustrated in FIG. Substrate 300 (eg, microscope slide with immobilized microparticles ) Is placed on the x-y conversion table 302, which is a properly programmed digital Connected to and controlled by a computer 304, which For sale personal computers, such as 486-based machines or Apple Computer ( PowerPC model 7100 or 8100 available from Cupertino, CA) obtain. Computer software for table conversion and data collection functions Commercial laboratory software such as Lab Windows, available from National Instruments. Software.   The substrate 300 and the table 302 are connected to a microscope 306 having one or more objective lenses 308. Operationally coupled, this lens collects and directs light on the particles fixed to the substrate 300. Can be delivered. The excitation light 310 (preferably a laser) from the light source 312 A beam splitter 314 (eg, a dichroic mirror), which Redirect the beam through the objective lens 308 and then focus the beam onto the substrate 300 You. A lens 308 collects the fluorescent light 316 emitted from the microparticles, and It directs it through signal 314 to signal distribution lens system 318, and then One or more suitable for converting a property (eg, strength, lifetime, etc.) to an electrical signal Direct fluorescence to the upper photoelectric element. The signal distribution lens system 318 is standard in the art. Substantially various components, such as bandpass filters, fiber optics s), rotating mirrors, fixed position mirrors and lenses, diffraction gratings and the like. FIG. As shown, the signal distribution lens system 318 converts the fluorescence 316 into four separate photomultipliers. To tubes 330, 332, 334, and 336, and then the output is Photon counters 350, 352, 354, and 356. Photon counter Output is collected by computer 304, where it is stored and analyzed And displayed on video 360. Alternatively, the signal distribution lens system 318 It can be a diffraction grating that directs the light signal 318 to a CCD array.   Stability of positional localization in scanning And reproducibility are almost smelling resolution to separate nearby particles To decide. Preferably, the scanning system is located nearby (eg, Fine particles (which are separated by a particle diameter or less) should be able to be resolved. did Thus, for most applications, for example, the use of CPG microparticles The system should be capable of resolving objects of at least 10-100 μm. It is. Even higher resolutions may be desired in some embodiments, but increase Resolution increases the time required to completely scan the substrate; Thus, in some embodiments, a compromise must be made between speed and resolution Can be The increase in scanning time is due to the fact that particles A system that only scans locations that are known to be located) Thus, it can be achieved. Preferably, particle size and scanning system Resolution is cm^TwoRandomly distributed on a plane at a density between about 10,000-100,000 particles per It is selected to allow resolution of the placed fluorescently labeled microparticles.   In sequencing applications, loaded microparticles are immobilized on the surface of a substrate in various ways. Can be Immobilization results in significant loss of microparticles over successive cycles of reagent exposure and washing Should be strong enough to be able to receive without. The substrate is glass In some cases, the surface can be prepared using commercially available reagents (eg, Pierce Chemical) Can be derivatized with an alkylamino linker, and then re-acquired using conventional chemistry. It can be crosslinked to vidin to form an avidinated surface. Biotin part is for many people Can be introduced into the loaded microparticles. For example, tag to polynucleotide The fraction of the cloning vector used to attach (eg, 10-15%) , Immediately adjacent to the polynucleotide insert at the end of the polynucleotide opposite the tag Engineered to include a unique restriction site flanking (providing cohesive ends during digestion) You. This site is used for loading polynucleotides and tags for loading onto microparticles. Cut out by After loading, about 10-15% of loaded polynucleo Tide has a unique restriction site distal from the microparticle surface. Combined restriction endonu After digestion with creatase, a suitable double-stranded adapter containing a biotin moiety is Connected to the end. The resulting microparticles are then ligated by biotin-avidin ligation. Spread on the avidinized glass surface to be fixed.   Alternatively and preferably, when sequencing by ligation is used In the first ligation step, a mixture of probes is applied to the loaded microparticles Yes: Some fractions of the probe are type IIs restricted, as required for sequencing Contains recognition sites and some fractions of the probe do not have such recognition sites But instead contains a biotin moiety at its non-linked end. Preferably, the mixture is Contains about 10-15 percent of the otylated probes.   In yet another alternative, when DNA loaded microparticles are applied to a glass substrate DNA is non-specific on glass surface after several hours (eg, 24 hours) incubation Adsorption to reagents and washing solutions without significant loss of particulates Produces a bond strong enough to allow exposure of Preferably, such a gala The substrate is a flow cell, which is a chamber etched into a glass slide. Flannel. Preferably, such channels allow liquid to pass through them. Closed so that it can be sucked out with a pump, and a monolayer of particulates within the defined viewing area Has a depth close enough to the diameter of the microparticles to be captured bycDNA Identification of novel polynucleotides in libraries   Novel polynucleotides in the cDNA library, as described above, It can be identified by constructing a library of attached cDNA molecules. live A large fraction of the rally, or even the entire library, is then partially distributed in parallel. Columns can be determined. Isolation of mRNA, and possibly, Soares et al., Proc. Natl. Acad. After population normalization as taught in references such as Sci., 91: 9228-9232 (1994). The following primers were used for first strand synthesis with reverse transcriptase using conventional protocols. Can be hybridized to a poly A tail for formation (SEQ ID NO: 1): Where [W, W, W, C]₉Represents a tag as described above, and “ACCAGCTGATC” is a double-stranded An arbitrary sequence that forms a restriction site in the form, and a "primer site" To amplify the target polynucleotide by using the primer binding site later. It is an ordinary sequence for all members of the library used.   After reverse transcription and second strand synthesis by conventional techniques, the double-stranded fragment Inserted into a cloning vector and amplified as described. Then amplified The library obtained is sampled and the sample is amplified. Amplified trial The cloning vector from the source was isolated and tagged cDNA fragment Are cut out and purified. One tag with polymerase as above After chaining, the fragments are methylated and, according to the invention, are solubilized on microparticles. Will be Preferably, as described above, the cloning vector comprises a tag So that the attached cDNA can be cut out with an endonuclease such as Fok I. Once constructed and sorted and coupled to microparticles, the preferred single base method Allows for immediate sequencing.   Then a sufficient number of nucleotides are found in the genome of the organism from which the library is derived. Until identified in each cDNA for a unique expression of According to the invention, the entire library or one or more large fractions of the library At the same time. For example, if the library is derived from mammalian mRNA, The 14-15 nucleotide length chosen for the sequence is between 2000 and It is predicted to have a unique designation of 3000 megabases. Of course, much less Identification of nucleotides not found in bacterial or other lower animal derived libraries Sufficient for a unique display. Preferably, at least 20-30 nucleotides are unique Labeling and allows for the construction of appropriate primers as described below. To be identified. The tabulated sequences are then used to identify unique cDNAs. It can be compared to known sequences.   The unique cDNA is then determined by conventional techniques, e.g., by priming and sequencing. PCR amplicon generated with primers directed to the selected cDNA portion By constructing a probe from Then the probe is To identify cDNAs in a library using our screening protocol Can be used for   The above methods for identifying new cDNAs are also useful in isolated assays. Fingerprint the mRNA population, either in terms of Can be used to The partial sequence information is obtained as described in the method above. Large samples attached to separate, e.g., 10,000 to 100,000 or more cDNAs Are obtained at the same time.                                 Example 1                           Building a tag library   A representative tag library comprises nucleotides A, G, defined by the following formula: Construct as follows to form a chemically synthesized 9-word tag of T and T: here,"[^Four(A, G, T)₉] "Indicates a tag mixture, where each tag is A, G, and T And "p" indicates 5 'phosphate. This mixture is Ligation to lower right and left primer binding regions (SEQ ID NO: 4 and SEQ ID NO: 5) : The right and left primer binding regions are ligated to the above tag mixture and then ligated. The single-stranded portion of the isolated structure is filled with DNA polymerase and then And the left primer and amplify to obtain a tag library (SEQ ID NO: No. 6). The underlined portion of the left primer binding region indicates the Rsr II recognition site. Right ply The leftmost underlined regions of the mer-binding regions are Bsp120I, ApaI, and Eco01. The recognition site for 09I and the cleavage site for HgaI are shown. Right primer The rightmost underlined region of the binding region indicates the recognition site for HgaI. If necessary, the right or left primer can be used for purification after amplification and / or cleavage. (For example, available from Clontech Laboratories, Palo Alto, CA) Functional (using conventional reagents) with biotin attached. The plasmid is cut with Ppu MI and Pme I (Rsr so that the insert is oriented).  II to obtain compatible and blunt ends) and then use DAM methylase Chill. The tag-containing construct is cut with Rsr II and then ligated into the open plasmid. The conjugate was then cut with Mbo I and Bam HI to ligate and plasmid Enables the closure of The plasmid is then amplified and isolated, and Use as specified.                                 Example 3         In rat liver tissue exposed to various xenobiotic drugs Changes in gene expression profiles   In this example, genes induced as a result of exposure to xenobiotic compounds were detected. To test the ability of the method of the present invention to produce Expression profiles are known to induce the expression of cytochrome P-450 isoenzymes. The following doses of some compounds are tested. Results obtained from the method of the invention Was obtained from the reverse transcriptase PCR assay and the immunochemical assay for cytochrome P-450 isozymes. Compare with the results obtained. Protocols and materials for the latter assay are rris et al., Biochemical Pharmacology, 52: 781-792 (1996).   Using male Sprague-Dawley rats 6-8 weeks of age and 200-300 g body weight, and Food and water are provided ad libitum to the animals. The test compound was phenobarbital (PB) , Metyrapone (MET), dexamethasone (DEX), clofibrate (CLO), Corn oil (CO), and β-naphthoflapon (BNF); emical Co. (St. Louis, MO). Anti-specific P-450 enzyme Bodies are available from the following sources: Rabbit anti-rat CYP3A1 is available from Human Biolog ics, Inc. (Phoenix, AZ); Goat anti-rat CYP4A1 is from Daiichi Pure Chemicals Co. (Tokyo, Japan); Monoclonal mouse anti-rat CYP1A1, Monoclona Mouse anti-rat CYP2C11, goat anti-rat CYP2E1, and monoclonal mouse Anti-rat CYP2B1 is available from Oxford Biochemical Research, Inc. (Oxford, MI) . Secondary antibodies (goat anti-rabbit IgG, rabbit anti-goat IgG, and goat anti-mouse IgG) Available from Jackson ImmunoResearch Laboratories (West Grove, PA) .   Wang et al., Arch. Biochem. Biophys. 290: 355-361 (1991) Following the same administration method as above, animals were given PB (100 mg / kg), BNF (100 mg / kg), MET (100mg / kg), DEX (100mg / kg) or CLO (250mg / kg) for 4 days Administer by continuous intraperitoneal injection. H_TwoControl animals treated with O and CO Used as a tool. Animals were sacrificed 2 hours after the last injection (day 4) and liver Remove the offal. Livers are immediately frozen and stored at -70 ° C.   Total RNA was modified by the method described in Xie et al., Biotechniques, 11: 326-327 (1991). Prepared from frozen liver tissue using About 100-200 mg of liver tissue, Xie Homogenize in the RNA extraction buffer described above to isolate total RNA. RN obtained A was reconstituted in diethyl pyrocarbonate-treated water and subjected to spectrophotometry at 260 nm. Therefore, quantify and adjust to a concentration of 100 μg / ml. Total RNA, apparent degradation In water treated with diethyl pyrocarbonate at -70 ° C for up to one year without any Exist. RT-PCR and sequencing are performed on samples from these preparations.   For sequencing, about 0.5 μg of poly (A)⁺Use an RNA sample corresponding to the RNA `` Attaching Tags to Polynucleotides for Sorting onto  Sol tag-cDNA binding according to the protocol described in the section entitled “id Phase Supports”. Build a body library: The tag repertoire contains the six 4 nucleotides from Table II. Constructed from Otide. Therefore, the complexity of the repertoire is 8⁶Or about 2 .6 × 10^FiveIt is. For each tag-cDNA conjugate library constructed, about 10,0 Ten samples of the 00 clone are taken for amplification and sorting. Amplified Each of the samples contains approximately 10⁶Immobilized 10 μm diameter GMA beads Apply separately to single layer. That is, the tag complementation in the GMA beads group in each monolayer The “sample” of an object is a repertoire of about four times the total size, Each Ringed Tag-cDNA Conjugate Finds Its Tag Complement on a Monolayer Make sure there is a high probability. The oligonucleotide tag of the amplified sample is After being single-stranded as described above, the tag-cDNA conjugate of the sample is Specific hive only between the tag and the tag complement forming a perfectly matched duplex Applied separately to the monolayer under conditions that allow for redidation. Amplified trial Select the concentration of the reagents and the hybridization time About 5 × 10 on beads^Four~ 2 × 10^FiveAllows loading of tag-cDNA conjugate. Consolidation Later, the 9-12 nucleotide portion of the attached cDNA was replaced by Brenner, International Patent Application PCT / US9. Determined in parallel by the single base sequencing technique described in 5/03678. Heredity The frequency distribution for the offspring expression profile was determined from the sequence obtained from each of the 10 samples. Build from information.   For cytochrome P-450 gene and constitutively expressed cyclophilin gene RT-PCR of the corresponding selected mRNA is performed as described in Morris et al., Supra. Simple Speaking of 1x reverse transcriptase buffer (Gibco BRL), 10 nM dithiothreitol, 0.5 nM dNTP, 2.5 μM oligo d (T)₁₅Primer, 40 units RNasin (Promega, Madis on, WI), 200 units RNase H-reverse transcriptase (Gibco BRL), and 400 ng of total RNA A 20 μl reaction mixture containing (diethyl pyrocarbonate-treated water) is prepared. You. The reaction was incubated at 37 ° C. for 1 hour and then at 95 ° C. for 5 minutes Inactivate. The resulting cDNA is stored at -20 C until use. cDNA PCR For amplification, 10x polymerase reaction buffer, 2mM MgCl_Two, 1 unit Taq DNA Polymerase (Perkin-Elmer, Norwalk, CT), 20 ng cDNA, and Morris et al. 10 μl containing 200 nM concentrations of 5 ′ and 3 ′ specific PCR primers of the sequence described above An L reaction mixture is prepared. PCR was performed on a Perkin-Elmer 9600 thermal cycler. Thaw at 94 ° C for 30 seconds, 56 ° C for 1 minute, and 72 ° C for 1 minute, respectively, annealing And 23 cycles using elongation and extension conditions. 5% less cDNA product Separate by PAGE using a denaturing gel. Band stained with ethidium bromide Detect by color.   Western blots of liver proteins were run on standard Perform using tocol. Briefly, proteins were reduced to 10% SDS-PA under reducing conditions. Separation on GE gels and Harris et al., Proc. Natl. Acad. Sci., 88: 1407-1410 (1 Using a modification of the method described in 991), immunoblotting for detection of P-450 isoenzymes To The protein was loaded at 50 μg / lane and incubated at 2 ° C. for about 4 hours. Disconnect under current (250V). Proteins are converted to 120 mM glycine and 20% (v / v) meta Nitrocellulose membrane (Bio-Rad, He- rcules, CA). Block the nitrocellulose membrane with 2.5% BSA and Primary monoclonal and polyclonal antibodies and secondary alkaline phosphatase Immunoblot for P-450 isoenzymes using tatase-conjugated anti-IgG. Immunoblots are developed with the Bio-Rad alkaline phosphatase substrate kit.   The three types of P-450 isoenzyme induction measurements showed substantial agreement.                                 Appendix Ia           To produce minimally cross-hybridizing sets Exemplary computer program                   (Single-stranded tag / single-stranded tag complement)                                 Appendix Ib           To produce minimally cross-hybridizing sets Exemplary computer program                   (Single-stranded tag / single-stranded tag complement)                                 Appendix Ic           To produce minimally cross-hybridizing sets Exemplary computer program                   (Double-stranded tag / single-stranded tag complement)

Claims (1)

[Claims]   1. A method for determining the toxicity of a compound, comprising:   Administering the compound to a test organism;   Extracting a group of mRNA molecules from each of one or more tissues of the test organism;   A separate group of each cDNA molecule has an oligonucleotide tag attached to it. Form a separate group of cDNA molecules from each group of mRNA molecules from one or more tissues Step, wherein the oligonucleotide tags have the same minimum crossover hybridization set. A process selected from:   Different oligonucleotides attached to substantially all different cDNA molecules in separate groups Separately sampling each group of cDNA molecules to have leotide;   Specifically hybridize the oligonucleotide tag with its respective complement Sorting the cDNA molecules of each separate group. Each complement is substantially identical to a spatially distinct region on one or more solid supports. Attached as a homogeneous group of essentially identical complements;   Forming a frequency distribution of the expressed genes for each of the one or more tissues For each of the separate groups of each of the sorted cDNA molecules, Determining the nucleotide sequence; and   Transforming the frequency distribution of the expressed genes in each of the one or more tissues Correlating the toxicity of the compound, A method comprising:   2. The oligonucleotide tag and the complement of the oligonucleotide tag 2. The method of claim 1, wherein the object is single-stranded.   3. 3. The method of claim 2, wherein the oligonucleotide tag comprises a plurality of oligonucleotide tags. Subunits, each subunit being an oligonucleotide 3-9 nucleotides in length Leotide, and each subunit has the same minimal crossover hybridization set. Method selected from the list.   4. 4. The method of claim 3, wherein said one or more solid supports are microparticles. And the step of sorting the cDNA molecules on the microparticles is loaded. A method of producing a subgroup of loaded microparticles and a subgroup of unloaded microparticles.   5. A process for separating the loaded particles from the unloaded particles. 5. The method of claim 4, further comprising the step of:   6. 6. The method of claim 5, wherein the number of loaded microparticles is at least The above steps of sampling, sorting, and separation until also accumulated at 10,000 The method further comprising the step of repeating the steps.   7. 7. The method of claim 6, wherein the number of loaded microparticles is at least 100,000. The described method.   8. 9. The method of claim 7, wherein the number of loaded microparticles is at least 500,000. The described method.   9. 6. The method of claim 5, wherein the 95% confidence limit of about 0.1% of the group is zero. Estimate the relative abundance of cDNA molecules present in the group with a frequency in the range of 1% to 5% Sampling until the number of loaded microparticles is sufficient to accumulate Further comprising the step of repeating said steps of sorting, sorting and separating. Law.   10. 5. The method of claim 4, wherein said test organism is a mammalian tissue culture. .   11. 11. The method of claim 10, wherein said mammalian tissue culture comprises hepatocytes.   12. The test organism is a rat, mouse, hamster, guinea pig, rabbit , 5. An animal selected from the group consisting of cats, dogs, pigs, and monkeys. The method described in.   13. The one or more tissues may be liver, kidney, brain, cardiovascular, thyroid, spleen, Consists of kidney, large intestine, small intestine, pancreas, bladder, stomach, ovary, testis, and mesenteric lymph glands 13. The method of claim 12, wherein the method is selected from a group.   14． Differentially expressed in selected tissues of test animals after treatment with compound A method for identifying a gene, comprising:   Administering the compound to a test animal;   Extracting a group of mRNA molecules from the selected tissue of the test animal;   So that each cDNA molecule has an oligonucleotide tag attached to it. Forming a group of cDNA molecules from the group, wherein the oligonucleotide tag comprises Selected from the same minimum intersection hybridization set;   Virtually all different cDNA molecules have different oligonucleotide tags attached to them. Sampling said group of cDNA molecules to have   Specifically hybridize the oligonucleotide tag with its respective complement Thereby sorting the cDNA molecules, wherein the respective Complements are substantially identical at spatially distinct regions on one or more solid supports; Attached as a uniform group of objects;   To form a frequency distribution of the expressed genes, the sorted cDNA Determining the nucleotide sequence of a portion of each of the offspring; and   The frequency distribution of the expressed genes in the selected tissue of the test animal is By comparing the frequency distribution of expressed genes in selected tissues of roll animals Thus, identifying a gene expressed in response to administering the compound, A method comprising:   15. The oligonucleotide tag and the complement of the oligonucleotide tag 15. The method of claim 14, wherein is single-stranded.   16. 16. The method of claim 15, wherein the number of oligonucleotide tags is multiple. And each subunit is an oligonucleotide having a length of 3 to 9 nucleotides. And each subunit has the same minimal crossover hybridization sequence. Method selected from the set.   17． 17. The method of claim 16, wherein said one or more solid supports are microparticles. And the step of sorting the cDNA molecules on the microparticles comprises: To produce a subgroup of loaded microparticles and a subgroup of unloaded microparticles. Law.   18. Separating the loaded microparticles from the unloaded microparticles 18. The method of claim 17, further comprising the step of:   19. 19. The method of claim 18, wherein the number of loaded microparticles is low. Before sampling, sorting, and separation until accumulated at least 10,000 A method further comprising the step of repeating the above steps.   20. 2. The method of claim 1, wherein the number of loaded microparticles is at least 100,000. 10. The method according to 9.   21. 3. The method of claim 2, wherein the number of loaded microparticles is at least 500,000. The method according to 0.   22. 19. The method of claim 18, wherein the group has a 95% confidence limit of as much as 0.1%. The relative abundance of cDNA molecules present in the group with a frequency in the range of 0.1% to 5% Until the number of loaded particulates is sufficient to accumulate, Further comprising repeating the steps of pulling, sorting, and separating How.   23. The test organism is a rat, mouse, hamster, guinea pig, rabbit 18. The composition of claim 17, wherein the composition is selected from the group consisting of, cats, dogs, pigs, and monkeys. the method of.   24. The selected tissues may be liver, kidney, brain, cardiovascular, thyroid, spleen, Consists of kidney, large intestine, small intestine, pancreas, bladder, stomach, ovary, testis, and mesenteric lymph glands 24. The method of claim 23, wherein the method is selected from a group.   25. Robust parallel determination of compound toxicity in test organisms Using the technique of signature sequencing of   Administering the compound to a test organism;   Extracting a group of mRNA molecules from each of one or more tissues of the test organism. And forming a group of cDNA molecules for each from the one or more tissues;   Different oligonucleotides attached to substantially all different cDNA molecules in separate groups Separately sampling each group of cDNA molecules to have leotide;   Forming a frequency distribution of the expressed genes for each of the one or more tissues In order to use robust parallel signature sequencing, each separate group of cDNA molecules Determining a partial nucleotide sequence of each; and   Transforming the frequency distribution of the expressed genes in each of the one or more tissues Correlating the toxicity of the compound, Use, including. 26. 26. The use according to claim 25, wherein said test organism is a mammalian tissue culture. 27. 27. The use according to claim 26, wherein said mammalian tissue culture comprises hepatocytes.   28. The test organism is a rat, mouse, hamster, guinea pig, rabbit An animal selected from the group consisting of, cats, dogs, pigs, and monkeys. 2 Use according to 5.   29. The one or more tissues may be liver, kidney, brain, cardiovascular, thyroid, spleen, Consists of kidney, large intestine, small intestine, pancreas, bladder, stomach, ovary, testis, and mesenteric lymph glands 29. Use according to claim 28 selected from the group.   30. Differentially expressed in the test organism after treatment with the compound, and Robust parallel signature sequencing to identify genes associated with toxicity A use of the technique, wherein the use comprises:   Administering the compound to the test organism;   Extracting a group of mRNA molecules from the selected tissue of the test organism to form a group of cDNA molecules Forming step;   Tight parallel signature sequences to create a frequency distribution of expressed genes Using the determination to determine the nucleotide sequence of a portion of each of the cDNA molecules. About;   Controlling the frequency distribution of the expressed genes of the selected tissue of the test organism. By comparing the frequency distribution of expressed genes in selected tissues of the rolled organism Identifying a gene that is expressed in response to administering the compound; And   The gene expressed in response to administering the compound is expressed in the test organism. Determining whether it is correlated with the toxicity of the compound, Use, including.