JP7376867B2 - How to identify sequence clusters containing antigen-specific antibodies - Google Patents

Description

The present invention relates to methods for identifying sequence clusters containing sequence data indicative of antigen-specific antibodies. The present invention also relates to a determination system for determining whether a sequence cluster contains an antigen-specific antibody, and a method for producing an antigen-specific antibody.

Antibodies are biomolecules that are used in many fields, including the fields of medicine and molecular biology. For example, in the pharmaceutical field, they are used as therapeutic agents and diagnostic agents, and in the molecular biology field, they are used as carriers for antigen purification and/or detection reagents.

Antigen-specific monoclonal antibodies can be obtained by various methods. For example, an antigen-specific monoclonal antibody is prepared by fusing B cells obtained from an animal to which the antigen has been administered with myeloma cells to produce hybridomas, and selecting a specific monoclonal antibody against the antigen from among the hybridomas. It can be obtained by a screening method that involves selecting antibody-producing hybridomas that exhibit reactivity.

In addition, by using biopanning methods such as phage display and ribosome display, we can extract antigen-specific monoclonal antibodies that exhibit a predetermined reactivity toward a specific antigen from an antibody library containing antigen-specific antibody candidates. Obtainable.

In recent years, various sequencers (also referred to as "next generation sequencers") that enable large-scale and rapid measurement of gene sequences have been developed and are commercially available. A method has been reported that combines a conventional biopanning method and genetic information analyzed by a next-generation sequencer to rapidly and efficiently screen for antigen-specific antibodies (Patent Document 1).

A method to obtain antigen-specific monoclonals for frequently detected cells from genetic information obtained by analyzing immune cells increased in specific lymph nodes using a next-generation sequencer without performing biopanning has also been reported. (Non-patent Document 1).

Japanese Patent Application Publication No. 2015-119637

Scientific Report 2015(5)13926:1-10

However, if we focus only on the genetic information of antibodies that are frequently detected in specific organs, as in the method described in Non-Patent Document 1, for example, it is difficult to The useful antigen-specific antibodies that they have will be overlooked.

The present inventors were able to develop antigen-specific antibodies by selecting antibody populations with large accumulation of mutations over time from among the genetic information of antibody groups obtained from animals administered with antigen at multiple time points. They discovered that genetic information can be selected and completed the present invention. The present inventors further discovered a method for efficiently obtaining genetic information of antigen-specific antibodies from genetic information of antibody populations obtained at multiple time points from animals administered with antigen, and completed the present invention. .

That is, the present invention relates to:
A method for identifying sequence clusters containing sequence data representing antigen-specific antibodies that are candidates for antibody production, the method comprising preparing a sequence data group containing sequence data of antibodies obtained from animals that may have received immune stimulation. a step of creating a molecular phylogenetic tree from the sequence data group; and a change in the degree of sequence identity between the sequence of the antibody shown in the sequence data in the sequence cluster in the molecular phylogenetic tree and the original sequence of the antibody or the antibody the sequence cluster includes a step of determining whether the sequence cluster includes sequence data indicating an antigen-specific antibody that is a candidate for antibody production, based on a change in the type of and a sequence that has the highest degree of sequence identity with the genome sequence of the animal species corresponding to the animal from which the antibody was obtained.

A determination system for determining whether or not a sequence cluster includes sequence data indicating an antigen-specific antibody that is a candidate for antibody production, comprising a control unit and a storage unit, the storage unit including a sequence cluster containing sequence data indicating an antigen-specific antibody that is a candidate for antibody production. A database containing sequence data groups containing sequence data of antibodies obtained from animals of different sexes is stored, and the control unit includes a molecular phylogenetic tree creation unit that creates a molecular phylogenetic tree from the sequence data groups; The sequence cluster is an antigen for which the sequence cluster is a candidate for antibody production based on a change in sequence identity between the sequence data of the sequence cluster in the molecular phylogenetic tree and the original sequence of the antibody or a change in the type of antibody. a sequence data determination unit that determines whether or not the original sequence contains sequence data indicating a specific antibody; A determination system that determines which sequence has the highest degree of sequence identity with the genome sequence.
A method for producing antigen-specific antibodies, the step of preparing a sequence data group containing sequence data of antibodies obtained from animals that may have received immune stimulation; creating a molecular phylogenetic tree from the sequence data group. Step: The sequence cluster is determined as a candidate for antibody production based on a change in sequence identity between the sequence data of the sequence cluster in the molecular phylogenetic tree and the original sequence of the antibody or a change in the type of antibody. Step of determining whether or not sequence data indicating an antigen-specific antibody is included; From sequence clusters determined to include sequence data indicating an antigen-specific antibody, sequence data indicating an antigen-specific antibody that is a candidate for antibody production is determined. A production method comprising the steps of selecting; and producing an antibody having an amino acid sequence shown in the selected sequence data.

A method for selecting sequence data indicating antigen-specific antibodies that are candidates for antibody production, the step of preparing a sequence data group containing sequence data of antibodies obtained from animals that may have received immune stimulation; , the step of selecting sequence data indicative of an antigen-specific antibody from a sequence data group, wherein the preparation step selects sequence data relating to a sequence having a predetermined degree of sequence identity with a sequence indicated by specific sequence data in the sequence data group; is less than a threshold regarding the number of sequence data, the selection step includes the step of excluding the specific sequence data from the sequence data group, and the selection step includes the step of excluding the specific sequence data from the sequence data group, and the selecting step and the original sequence of the antibody, and a threshold value, the original sequence is the sequence of the antibody shown in the sequence data, and the animal corresponding to the animal from which the antibody was obtained. The selection method is the sequence with the highest degree of sequence identity with the genome sequence of the species.
A selection system for selecting sequence data indicative of antigen-specific antibodies that are candidates for antibody production, comprising a control unit and a storage unit, the storage unit comprising sequence data obtained from an animal that may have received immune stimulation. The controller stores a database containing sequence data groups containing sequence data of antibodies that have been obtained, and the control unit stores sequence data regarding sequences having a predetermined degree of sequence identity with a sequence indicated by specific sequence data in the sequence data group. is less than a threshold regarding the number of sequence data, the sequence data preparation unit includes the step of excluding the specific sequence data from the sequence data group, and the sequence data shown in the sequence data in the sequence data group. a sequence data selection unit that includes a step of comparing the degree of sequence identity between the antibody sequence and the original sequence of the antibody and a threshold value, and the original sequence is the sequence of the antibody shown in the sequence data and the antibody sequence. is the sequence that has the highest sequence identity with the genome sequence of the animal species corresponding to the obtained animal.

A method for selecting sequence data indicating antigen-specific antibodies that are candidates for antibody production, the step of preparing a sequence data group containing sequence data of antibodies obtained from animals that may have received immune stimulation; selecting sequence data indicative of an antigen-specific antibody from a sequence data group; the preparation step selects sequence data relating to a sequence having a predetermined degree of sequence identity with a sequence indicated by specific sequence data in the sequence data group; is less than a threshold regarding the number of sequence data, the selection step includes the step of excluding the specific sequence data from the sequence data group, and the selection step includes the step of excluding the specific sequence data from the sequence data group, and the selecting step and the original sequence of the antibody, and a threshold value, the original sequence is the sequence of the antibody shown in the sequence data, and the animal corresponding to the animal from which the antibody was obtained. The selection method is the sequence with the highest degree of sequence identity with the genome sequence of the species.

A selection system for selecting sequence data indicative of antigen-specific antibodies that are candidates for antibody production, comprising a control unit and a storage unit, the storage unit comprising sequence data obtained from an animal that may have received immune stimulation. The controller stores a database containing sequence data groups containing sequence data of antibodies that have been obtained, and the control unit stores sequence data regarding sequences having a predetermined degree of sequence identity with a sequence indicated by specific sequence data in the sequence data group. is less than a threshold regarding the number of array data, an array data preparation unit including executing a process of excluding the specific array data from the array data group; and array data in the array data group. and a sequence data selection unit that performs a process of comparing the degree of sequence identity between the antibody sequence shown in the sequence of the antibody and the original sequence of the antibody with a threshold value, and the original sequence is shown in the sequence data. A determination system in which the sequence of an antibody has the highest sequence identity with the genome sequence of the animal species corresponding to the animal from which the antibody was obtained.

A method for producing antigen-specific antibodies, the step of preparing a sequence data group containing sequence data of antibodies obtained from an animal that may have received immune stimulation; sequence data indicating antigen-specific antibodies from the sequence data group. and a step of producing an antibody having an amino acid sequence shown in the selected sequence data, and the preparation step includes a step of selecting a sequence shown in specific sequence data in the sequence data group and a predetermined sequence. If the number of sequence data related to sequences having a degree of matching is less than a threshold regarding the number of sequence data, the specific sequence data is excluded from the sequence data group, and the selection step includes the step of excluding the specific sequence data from the sequence data group. The step of comparing the degree of sequence identity between the sequence of the antibody shown in the sequence data and the original sequence of the antibody with a threshold value, and when the degree of sequence identity is less than the threshold value, the sequence data is The original sequence is selected as sequence data indicating an antibody, and the original sequence has the highest sequence identity between the antibody sequence shown in the sequence data and the genome sequence of the animal species corresponding to the animal from which the antibody was obtained. , fabrication method.

According to one aspect of the present invention, genetic information of an antigen-specific antibody that is a candidate for antibody production can be quickly and efficiently obtained without using experimental techniques such as biopanning. Furthermore, according to another aspect of the present invention, antigen-specific antibodies can be rapidly and efficiently produced based on the acquired genetic information.

A series of graphs showing the integration of sequencing errors that can occur during the sequencing process. In FIG. 1a, the specific reference array data α (n=0) has the same array length as the array of the reference array data α (hereinafter referred to as "reference array α"), but n pieces ( This is a histogram regarding the reference sequence α, which includes sequence data of sequences having base substitutions (n=1 to 5) (hereinafter referred to as "query sequence data"). FIG. 1b is a line graph showing that the distribution of the class values of the histogram (solid line) for the reference array α follows the probability density function (dashed line) for the Poisson distribution. FIG. 1c is a bar graph after integrating the reference sequence data α with the query sequence data (n=1 to 5). FIG. 1d shows reference array data β different from the reference array data α in FIG. This is a histogram regarding the reference sequence β, which includes query sequence data (n=1 to 5) of sequences having base substitutions of ~5). Figure 1e is a histogram for reference sequence β after excluding reference sequence data γ (n=2) different from reference sequence data β and query sequence data having sequences presumed to be derived from reference sequence γ. be. FIG. 1f is a histogram for reference sequence γ, including reference sequence data γ and query sequence data excluded from the histogram for reference sequence β of FIG. 1d. FIG. 2a is a line graph showing classification based on the sequence length of antibody variable regions. FIG. 2b is a table in which the sequence data of the sequence length 354 bp in FIG. 2a is classified into classes based on V gene fragments and J gene fragments. Figure 3a shows a case where non-cluster-forming sequence data were not excluded for the sequence data group classified in Figure 2b (V gene fragment is vhh3-S1, J gene fragment is ighJ-4). (U40≧0). FIG. 3b is a molecular phylogenetic tree obtained by excluding sequence data in which the number of sequence data having up to 40 base substitutions is less than 10 (U40≧10). FIG. 3c is a molecular phylogenetic tree obtained by excluding sequence data in which the number of sequence data having up to 40 base substitutions is less than 50 (U40≧50). In FIGS. 3a to 3c, red circles (dark color) indicate sequence data of IgG2, green circles (light color) indicate sequence data of IgG3, and the size of each circle represents the frequency of appearance (integration). Figure 4a is a molecular phylogenetic tree for extracting sequence clusters from the molecular phylogenetic tree. FIG. 4b is a diagram showing nine sequence clusters (b1 to b9) extracted from the molecular phylogenetic tree of FIG. 4a. FIG. 5a is a molecular phylogenetic tree created for a classified sequence data group based on a sequence length of 372 bp (the V gene fragment is VHH3-S8 and the J gene fragment is ighJ-6). FIG. 5b is a molecular phylogenetic tree created for the classified sequence data group based on the sequence length of 354 bp (the V gene fragment is vhh3-S1 and the J gene fragment is ighJ-4). Figures 5c and 5d are scatter plots showing the relationship between the bit score of each sequence data in the sequence cluster indicated by a red circle (circled by a curved line) in Figures 5a and 5b, respectively, and the timing of blood collection. be. Figures 5e and 5f are Biacore sensorgrams showing the avidity of antibodies reconstructed from sequence data in the sequence clusters indicated by red circles (circled by curves) in the phylogenetic trees of Figures 5a and 5b, respectively. It is. FIG. 6a is a molecular phylogenetic tree created for a classified sequence data group based on a sequence length of 381 bp (the V gene fragment is VHH3-S8 and the J gene fragment is ighJ-6). Part 6b is a scatter diagram showing the relationship between the bit score of sequence data in the sequence cluster indicated by a red circle (circled by a curved line) in FIG. 6a and the timing of blood collection. Figure 6c shows a molecular phylogenetic tree created for the sequence data in the sequence clusters indicated by red circles (circled by curved lines) in the phylogenetic tree of Figure 6a, as well as the molecular phylogenetic tree shown by the dashed arrows in the molecular phylogenetic tree. This is a Biacore sensorgram showing the binding strength of antibodies reconstructed from sequence data (c1, c2, c3). FIG. 2 is a block diagram showing a schematic configuration of a determination system 1 according to a second embodiment. FIG. 2 is a block diagram showing a schematic configuration of a database 107 according to a second embodiment. 7 is a flowchart showing steps for preparing an array data group according to the second embodiment. 3 is a flowchart showing sequence library preparation steps according to Embodiment 2. 10 is a flowchart illustrating steps for integrating sequencing errors according to Embodiment 2. 7 is a flowchart showing a step of selecting sequence data for integration according to the second embodiment.

[Definition]
As used herein, "sequence data" refers to computer-readable information that includes at least the base sequence encoding part or all of an antibody, or information regarding the amino acid sequence thereof (also referred to as "sequence information" or simply "sequence"). refers to electronic data containing sequence information of antibodies obtained from animals that may have received immune stimulation. Antibody sequence data can be obtained, for example, from lymphoid cells present in a blood sample taken from an animal that may have undergone immune stimulation, using known sequencing methods or commercially available sequencing equipment. . The blood sample can be, for example, whole blood, plasma or serum taken from an animal, or a sample that has been processed to isolate antibodies. The known sequencing method may be, for example, an analysis method based on the Sanger method. A commercially available sequencing device may be, for example, a next generation sequencer such as Illumina's Miseq or HiSeq 4000.

Antibody sequence information includes, but is not limited to, information regarding the base sequence encoding one or more complementarity determining regions (CDRs) or the amino acid sequence thereof. The antibody sequence information includes, for example, information regarding the base sequence encoding a part of the antibody variable region (V region) including at least the CDR or the amino acid sequence thereof. In addition to the information regarding the base sequence or amino acid sequence, the sequence data includes, for example, information regarding the quality of sequencing (also referred to as "quality score"), information regarding the time when the antibody was collected, and information regarding the animal species from which the antibody was collected. May include.

As used herein, the term "sequence data group" refers to a collection of multiple sequence data having antibody sequence information. A sequence data group is, but is not limited to, a collection of a huge number of sequence data containing sequence information for each antibody obtained by sequencing using a next-generation sequencer (also referred to as a "sequence data group after sequencing"). It may be. The sequence data after sequencing may be, for example, a combination of multiple sets of sequence data obtained from the same animal at different times. For example, a sequence data group is processed by cleaning up the sequence data group after sequencing, aggregation into unique sequence data, integrating sequencing errors, classification, and removing sequence data that does not form sequence clusters. It may be a collection of sequence data generated by performing any one of the steps alone, a combination of two or more steps, or all of the steps. The huge number of sequence data may be, for example, 10,000 to 10 million pieces, 100,000 to 8 million pieces, 1 million to 7 million pieces, or 2 million to 6 million pieces of sequence data.

As used herein, a "unique sequence" refers to unique sequence data that has sequence information with the same sequence order and sequence length within a certain sequence data group. Unique sequence data is defined as, but not limited to, when there are two or more pieces of sequence data having the same sequence information and sequence length in the sequence data group after sequencing, as described below. In the array data group obtained by performing data aggregation processing, it may be the only array data that further includes information regarding the number of aggregated array data pieces in addition to the above-mentioned array arrangement and array length. . In another example, if there is only one piece of sequence data in the sequence data group after sequencing that has sequence information with the same sequence order and sequence length, the above-mentioned sequence order in the sequence data group In addition to the array length, the array data may be the only array data that further includes information that the number of aggregated array data is one.

As used herein, the term "next generation sequencer" refers to a sequencing device that can simultaneously determine the sequence information of a huge number of gene fragments. The huge number of gene fragments includes, but is not limited to, 100,000 or more fragments, such as 300,000 or more fragments, 1 million or more fragments, 3 million or more fragments, 10 million or more fragments, 30 million or more fragments, 100 million or more fragments. For example, the gene fragments may be 200 million fragments or less, 100 million fragments or less, or 50 million fragments or less. The chain length of the gene fragment is not limited, but may be between 30 bp and 1000 bp, such as 900 bp or less, 500 bp or less, or 200 bp or less. Sequence data of antibody fragments whose sequence information has been determined by a next generation sequencer may include, but are not limited to, a quality score.

As used herein, "antibody" refers to all or part of an immunoglobulin (Ig) molecule that has the ability to bind to a predetermined antigen. Antibodies include, but are not limited to, amino acid sequences corresponding to the complementarity determining regions (CDRs) of the variable regions (V regions) of the heavy chains (H chains) and light chains (L chains) that constitute immunoglobulin molecules. It is a protein with Antibodies include, for example, whole antibodies, antibody fragments, and chimeric antibodies. The type of antibody may be, but is not limited to, IgG, IgM, IgA, IgD or IgE. IgG may be, for example, subclasses of IgG1, IgG2, IgG3, and IgG4.

As used herein, "antibody fragment" refers to a portion of an antibody that has the ability to bind to a predetermined antigen. Antibody fragments include, but are not limited to, Fab, F(ab') ₂ , or single chain antibodies (scFv). Antibody fragments can be produced by known methods. Fab can be produced, for example, from a complete antibody by known enzyme treatment.

As used herein, the term "chimeric antibody" refers to a fusion protein in which all or part of two or more proteins are bound together, and has the ability to bind to a predetermined antigen. Chimeric antibodies include, but are not limited to, antibodies in which the variable region (V region) of an antibody derived from a species other than humans and the constant region (C region) of an antibody derived from humans are linked, or antibodies derived from a species other than humans. (also referred to as a "humanized antibody") in which the CDRs of an antibody of 2000 and human antibody parts other than the CDRs are linked. Chimeric antibodies can be produced by known methods. Chimeric antibodies can be produced, for example, by linking genes using genetic engineering techniques.

As used herein, "antigen-specific antibody" refers to an antibody that binds to a target antigen with a predetermined binding force or higher. Binding strength is indicated by, but is not limited to, a dissociation constant (K _D ), an affinity constant (K _A ), an association rate constant (k _on ), or a dissociation rate constant (k _off ). Bonding strength can be measured by a known method or a commercially available measuring device. A commercially available measurement device may be, for example, GE Healthcare's Biacore.

^The predetermined binding ^strength is, but ^is not limited to, a ^dissociation constant (K _D ). The predetermined binding strength is, but is not limited to, the above-mentioned dissociation constant for the antigen of interest, and for substances other than the antigen, 10 ⁻⁵ M or more, preferably 10 ⁻⁴ M or more, or more. Preferably, it may exhibit a dissociation constant of 10 ⁻³ M or more.

For example, an antigen-specific antibody exhibits a dissociation constant of 10 ⁻⁷ M or less with respect to a predetermined antigen, and a dissociation constant of 10 ⁻⁴ M or more with substances other than the antigen. For example, an antigen-specific antibody exhibits a dissociation constant of 10 ⁻⁸ M or less with respect to a predetermined antigen, and a dissociation constant of 10 ⁻³ M or more with substances other than the antigen. The antigen-specific antibody has, for example, a binding rate constant for the antigen of 10 ⁵ M ⁻¹ ·s ⁻¹ or more, preferably 10 ⁶ M ⁻¹ ·s ⁻¹ or more. Antigen-specific antibodies have, for example, a dissociation rate constant from the antigen of 10 ⁻³ s ⁻¹ or less, preferably 10 ⁻⁴ s ^{−1 or} less.

As used herein, the term "variable region (V region)" refers to a portion consisting of approximately 110 amino acid residues on the amino (N) terminal side of an immunoglobulin molecule. The variable region of the heavy chain is designated _VH , and the variable region of the light chain is designated _VL . As used herein, "complementarity determining region (CDR)" means a region that forms a complementary three-dimensional structure with an antigen molecule and determines the specificity of an antibody. Although not limited to, CDRs are present in three separate locations in the H chain and L chain polypeptide chains of antibodies, and are represented as H1, H2, and H3 in the H chain, and L1, L2 in the L chain. and L3.

As used herein, "constant region (C region)" means a portion of an immunoglobulin molecule other than the variable region (V region). As used herein, "V gene" and "C gene" refer to a gene encoding a variable region (V region) and a gene encoding a constant region (C region), respectively.

As used herein, "VJ recombination" refers to a process in which a V gene fragment and a J gene fragment are linked to form the variable region of an antibody during the differentiation process of B cells. As used herein, "VDJ recombination" refers to a process in which a D (diversity) gene fragment is further linked between a V gene fragment and a J gene fragment to form an H chain. As used herein, "V(D)J recombination" means VJ recombination and/or VDJ recombination. In V(D)J recombination, bases are Deletions or additions may occur randomly at their joining sites.

As used herein, "V gene fragment" and "J gene fragment" each refer to a gene encoding a part of the variable region (V region) of an immunoglobulin polypeptide chain. The V region of the immunoglobulin H chain consists of 110 to 120 amino acids, and the gene consists of a part that encodes 90 to 100 amino acids from the N-terminus (V _H ), and a part that encodes the following several amino acids ( It consists of three gene segments: D _H ), and the remaining part (J _H ) encoding more than ten amino acids. The V region of the immunoglobulin L chain does not contain a D gene fragment and is composed of a V gene fragment (V _L ) and a J gene fragment (J _L ).

As used herein, "affinity maturation" refers to a phenomenon in which the affinity of an antibody for an antigen to which it is immunized increases over time after immunization. Affinity maturation can occur, but is not limited to, by point mutations occurring in the V regions of H and L chains during the rapid proliferation of B cells activated by antigen stimulation. Mutations during affinity maturation may introduce, but are not limited to, about one mutation per cell generation in the CDRs of the V region.

As used herein, the "original sequence" (also referred to as "primordial sequence" or "ancestral sequence") refers to the sequence of the antibody and the animal species corresponding to the animal from which the antibody was obtained. This means the sequence with the highest degree of sequence identity with the genome sequence of The sequence with the highest degree of sequence identity means, when a plurality of similar sequences exist, the sequence with the highest degree of sequence identity among the similar sequences. In the context of the original sequence, the antibody sequence is, but is not limited to, one or both of the V and J gene fragments that make up the variable region of the antibody. The original sequence is, for example, homologous in sequence identity between either or both of the V gene fragment and J gene fragment constituting the variable region of the antibody and the genome sequence of the animal species corresponding to the animal from which the antibody was obtained. Among the multiple similar sequences shown in the search, this is the gene fragment alone or in combination that has the highest degree of sequence identity.

As used herein, "homology search" refers to an operation of searching a database for a sequence similar to a query sequence. Homology searches are performed by, but are not limited to, known programs. Such known programs include, for example, the FASTA program, the BLSAT program, and the PSI-BLAST program. Calculation of sequence identity can be performed for either base sequences or amino acid sequences.

As used herein, "sequence identity" refers to an index indicating the degree of sequence identity between two sequences. Sequence identity is, but is not limited to, the number of sequence matches, sequence identity, or a score value based on sequence identity. In other examples, the degree of sequence identity is a score value based on the number of sequence discrepancies, sequence discrepancies, or sequence discrepancies. Sequence identity is calculated using, but not limited to, gap penalties.

The number of sequence matches or the number of sequence mismatches refers to the number of bases or amino acids that are matched or mismatched between the query sequence and a similar sequence. Sequence identity or sequence discrepancy is the ratio of the number of matched or mismatched bases or amino acids to the length of the query sequence (=[number of matched or mismatched bases or amino acids]/[of the query sequence]) length]). The score value means a numerical value obtained by scoring sequence identity or sequence discrepancy.

Sequence identity scoring is performed using, but is not limited to, a scoring matrix. Score matrices for scoring matches and mismatches in base sequences include, for example, Jukes-Cantor69 model, Kimura80 model, Tamura92 model, and Tamura-Nei93 model in which the probability of each base (A, T, G, C) being replaced is constant. There is a general time-reversal (GTR) model. Scoring models for scoring matches and mismatches in amino acid sequences include, for example, PAM matrices (eg, PAM250) and BLOSUM matrices (eg, BLOSUM50, BLOSUM62, BLOSUM80).

As used herein, the term "same sequence" refers to base sequences or amino acid sequences that have the same arrangement and length.

As used herein, "change in sequence identity" means that the degree of sequence identity between the sequence shown in the sequence data of the antibody in the sequence data group and its original sequence changes with the passage of time after antigen administration. do. The change in the degree of sequence identity is, for example, the slope of the degree of sequence identity. The slope of the sequence match may be a change in the sequence match with respect to the elapsed time from the start of the blood collection protocol to the end, or a change in the sequence match per number of blood collections from the first blood collection to the end of blood collection.

As used herein, "change in antibody type" means that the class of antibody changes over time after antigen administration. The change in antibody type is, for example, an antibody class switch. In one example, the index indicating the class switch of antibodies is the ratio of antibody classes in the antibody population (e.g., [type of IgG2]/[type of IgG class in the antibody population]) shown at the time of blood collection; This is the difference between the proportion of antibody classes in the antibody population shown after blood collection.

As used herein, the term "molecular phylogenetic tree" refers to a diagram showing branching relationships regarding antibody sequences. The molecular tree can be, but is not limited to, a star tree or an infinite tree. A molecular phylogenetic tree is created based on, but not limited to, information regarding the amino acid sequence of the antibody or the base sequence encoding the amino acid sequence. A molecular phylogenetic tree can be created by a known method. Methods for creating molecular phylogenetic trees are described, for example, in Bioinformatics: Sequence and Genome Analysis 2nd Edition (David Mount, Cold Spring Harbor Laboratory Press; d edition (August 16, 2004)) can be used.

In one example, a molecular phylogenetic tree can be created based on a distance matrix. When created based on a distance matrix, a molecular phylogenetic tree can be created using a neighbor-joining method, an average distance method, or a minimum evolution method. In other examples, a molecular phylogenetic tree can be created without using a distance matrix. When created without using a distance matrix, a molecular phylogenetic tree can be created using maximum parsimony, maximum likelihood, or Bayesian estimation. A molecular phylogenetic tree can be created using, but is not limited to, known software (e.g., ape package of statistical analysis software "R"), MEGAX (https://www.megasoftware.net/). . Molecular phylogenetic trees can be performed on groups of sequence data. Creation of a molecular phylogenetic tree is performed, for example, on cluster-forming sequence data groups. In other examples, the creation of a molecular phylogenetic tree may be performed on, but is not limited to, groups of sequence data in sequence clusters as described below.

As used herein, the term "sequence cluster" refers to a group (subgroup) that is interconnected in a manner that allows it to be distinguished from others in a molecular phylogenetic tree. Sequence clusters can be found using, without limitation, known methods of searching for subgroups used in the field of network analysis. Methods for finding sequence clusters include, for example, a method of searching based on the number of components connecting each separated sequence, and a method of extracting a portion of a molecular phylogenetic tree with high density of sequence information. These methods can be executed by, for example, known programs. As such known programs, for example, the Sna package and the Igraph package can be used. In one example, a sequence cluster is defined as a sequence cluster between branch points (also referred to as connection points) in a molecular phylogenetic tree when the length of the branch connecting the branch points is greater than or equal to a predetermined value. A region outside the center of the molecular phylogenetic tree from a predetermined position may be extracted as a sequence cluster.

As used herein, "candidate for antibody production" refers to sequence data present in a sequence data group or sequence cluster determined to contain an antigen-specific antibody. In one example, candidates for antibody production may be sequence data present in sequence clusters determined to contain antigen-specific antibodies. In one example, candidates for antibody production may be sequence data present in sequence groups generated after classification and elimination of sequence data that do not form sequence clusters.

As used herein, "antibody production" or "producing an antibody" means producing an antibody. Antibodies can be produced by known methods. As a known method, reference may be made to, for example, International Patent Publication WO99/18212. For example, an antibody can be produced by constructing a plasmid for antibody expression by inserting a DNA fragment encoding the CDR of interest into a known protein expression vector, introducing the plasmid into a predetermined host, and injecting the antibody of interest into the host. It can be produced or produced by expressing it. Antibodies expressed in a host can be recovered and purified by known methods. A DNA fragment encoding a CDR of interest can be optimized depending on the codon usage or codon frequency of the host in which it is expressed.

As used herein, the term "antigen" or "immunogen" refers to a substance that induces humoral immunity or cell-mediated immunity by producing antibodies or sensitized lymphocytes. The antigen may be, but is not limited to, a substance introduced into the animal's body from outside the body, or may be produced within the animal's body. The antigen introduced into the body from outside the body may be, for example, a proteinaceous substance of interest, a bacterium, or a virus. Antigens produced in the body may be, for example, proteinaceous substances produced due to mutations or foreign genes, or proteins denatured by external factors such as radiation exposure.

As used herein, "animals that may have received immune stimulation" include, but are not limited to, animals that have been administered a predetermined antigen, animals that have been irradiated, and animals that may have been infected with a virus. Contains animals. An animal to which a given antigen has been administered is, for example, an animal to which the protein antigen of interest has been administered according to the known administration protocols (number of doses, interval between doses, dose/time) commonly used for that animal species. If there is no known administration protocol for a specific animal species, an administration protocol can be appropriately determined by those skilled in the art with reference to known administration protocols for closely related animal species. The antigen is administered to the animal at least once. The antigen is administered, for example, in the animal the number of times that affinity maturation of antibodies is expected. When an antigen is administered two or more times, the interval between administrations of the antigen may be at regular intervals or at irregular intervals.

Animals that may have been infected with a virus include, for example, animals that are present in areas where viruses such as influenza viruses are endemic. Irradiated animals include, for example, animals that have undergone radiation therapy. Without limitation, in animals that have undergone tumor-directed radiation therapy, immune stimulation may result in antibodies that have undergone affinity maturation. Those skilled in the art can appropriately set the radiotherapy protocol depending on the condition of the region to be irradiated with radiation and the condition of the torso to be irradiated with radiation.

As used herein, "animal" includes, but is not limited to, cartilaginous fish, cetacean artiodactyls, rodents, or primates. Cartilaginous fishes include, for example, sharks, rays, and sharks. Cetacean artiodactyls include, for example, camelids, including alpacas and camels, boars, including pigs, and bovids, including cows, goats, and sheep. The cetacean artiodactyl is preferably an alpaca. Rodents include, for example, lagomorphs, including rabbits, and rodents, including mice and rats. The rodent animal is preferably a rabbit. Primates are, for example, prosimians including lemurs, Cercopithenes including Japanese monkeys, and hominids including humans, chimpanzees, and gorillas. A primate is, for example, a non-human animal or a human. The primate is preferably a human.

As used herein, "animal species" refers to the biological species to which an animal belongs. The genome sequence of the animal species may be, but is not limited to, a genome sequence stored on a known database. As known databases, for example, Genbank (NCBI), DDBJ (DNA Data Bank of Japan), Entrez (NCBI), and Genomic Biology (NCBI) can be used.

Embodiments according to aspects of the present invention will be described below, but these embodiments are illustrative of the present invention and do not limit the invention described in the appended claims in any way.

[Embodiment 1]
A first aspect of the invention provides a method for identifying sequence clusters containing sequence data indicative of antigen-specific antibodies that are candidates for antibody production. One embodiment (hereinafter referred to as "Embodiment 1") of the first aspect of the present invention will be described below.

<Preparation of sequence data group>
The identification method includes the step of preparing a sequence data group containing sequence data of antibodies obtained from an animal that has undergone immune stimulation. The step of preparing the sequence data group includes acquiring a specific sequence data group, and in some cases, performing a process such as cleanup described below on the acquired sequence data group to generate a new sequence data group. It further includes:

(Sequence data group after sequencing)
The step of preparing a sequence data group includes, but is not limited to, obtaining a sequence data group after sequencing. The sequence data group after sequencing is, for example, each antibody of an antibody population obtained i times (i = 1, 2, 3...) at different times from an animal that has been administered with the antigen of interest and subjected to immune stimulation. It is a collection of sequence data (X _i ) having sequence information obtained by determining the sequence of . Obtaining a sequence data group after sequencing includes, for example, loading the sequence data group into a computer. i is, but is not limited to, an integer between 1 and 15. i is, for example, an integer between 3 and 8, 3, 4, 5, 6, 7 or 8. The intervals at which multiple antibody populations are obtained may be at regular intervals or at irregular intervals.

A post-sequencing sequence data group could be obtained, for example, from an alpaca in which the antigen of interest was administered with adjuvant five times at two-week intervals, once before challenge, once during the challenge interval, and once after challenge. It may be a collection of sequence data of each antibody obtained from lymphoid cells in blood samples collected six times in total. The sequence data group after sequencing includes, for example, about 3 million pieces of sequence data.

(Cleanup of sequence data, etc. that does not indicate amino acid sequences)
Antibody sequence data obtained by sequencing from lymphoid cells in animal body fluids (e.g., blood, lymph) contains reading frame shifts due to, for example, the V(D)J recombination process. Sequence data having gene sequence information that does not indicate the amino acid sequence may be included.

The step of preparing a sequence data group includes, but is not limited to, the step of removing (cleaning up) sequence data that does not represent an amino acid sequence from the sequence data group. This cleanup step may be performed, for example, by excluding from the sequence data group nucleotide sequence data in which the nucleotide sequence encoding the variable region portion of the antibody or the region containing the portion is not a multiple of 3. By performing the cleanup process on the sequence data group, a collection of sequence data indicating amino acid sequences (hereinafter also referred to as "cleanup sequence data group") is generated. Additionally, the cleanup process can reduce the number of array data in the array data group.

The cleanup may optionally further include excluding sequence data below a predetermined quality score from the sequence data group based on the quality score. Cleanup based on quality scores can be performed on sequence data having base sequence information or amino acid sequence information.

(Aggregation into unique sequence data)
An antibody population obtained from an animal may include multiple antibodies having the same sequence. Therefore, a collection of sequence data obtained by sequencing the antibody population (sequence data group after sequencing) may include a plurality of sequence data having the same sequence information.

The sequence data group preparation process includes, but is not limited to, a process of aggregating multiple sequence data containing the same sequence information into one sequence data containing the sequence information (hereinafter referred to as "unique sequence data"). include. The unique sequence data also includes information regarding the number of pieces of aggregated sequence data (hereinafter referred to as "frequency of appearance (aggregated)"). By performing the aggregation process on the sequence data group, a collection of unique sequence data (hereinafter referred to as "unique sequence data group") is generated. The aggregation process can reduce the number of sequence data in the sequence data group.

(Consolidation of sequencing errors)
It is known that sequence information obtained by sequencing may contain random errors that occur during the process associated with sequencing. This random error can occur in various processes, including, for example, the process of DNA synthesis by DNA polymerases, including DNA amplification by PCR and DNA replication during sequencing, and/or the process of signal analysis from fluorescent images. Sequence data related to sequence information that contains random errors that occur during the sequencing process is referred to as sequence data (hereinafter referred to as "integrated unique sequence data") that indicates the sequence of the gene to be sequenced (also referred to as the "original sequence"). to be integrated into In addition to the information included in the unique sequence data, the integrated unique sequence data further includes information regarding the total value of the frequency of occurrence (aggregated) included in the integrated sequence data (hereinafter referred to as "frequency of occurrence (integrated)"). . By performing this integration step on the sequence data group, a collection of integrated unique sequence data (hereinafter referred to as "integrated unique sequence data group") is generated. The consolidation process makes it possible to reduce the number of sequence data in a sequence data group.

Preparing a group of sequence data includes, but is not limited to, integrating sequencing errors. Sequencing errors can be integrated, for example, by selecting sequence data (hereinafter referred to as "reference sequence data") with the highest frequency of occurrence (aggregated) in a collection of sequence data having frequencies of occurrence (aggregated), and (hereinafter referred to as "reference sequence"), but has n base substitutions with respect to said reference sequence (hereinafter referred to as "query sequence data"), and the appearance of said reference sequence data The method includes the step of integrating query sequence data in which the ratio of appearance frequency (aggregation) of the query sequence data to frequency (aggregation) is less than a threshold value with the reference sequence data.

The integration step further includes the step of generating a sequence data group for integration by selecting in advance sequence data having a sequence of the same length as the reference sequence from a sequence data group in which sequencing errors are integrated. good.

A case will be described in which integration of sequencing errors is performed on unique sequence data. In the unique sequence data group, the base sequence shown in the unique sequence data indicating the original sequence (or without any sequencing errors) is defined as the "reference sequence". The base sequence of the unique sequence data in the unique sequence data group, which has the same sequence length as the reference sequence and is different from the reference sequence, is referred to as a "query sequence."

When the base sequence of the query sequence is caused by an error in sequencing the base sequence of the reference sequence, the unique sequence data representing the query sequence is integrated with the unique sequence data representing the reference sequence. Whether or not the query sequence is caused by an error in sequencing the reference sequence is determined by calculating the ratio of the probabilities that the query sequence and the reference sequence appear.

For example, let p _n be the probability that a sequence determination error will occur at the nth base of a base sequence of sequence length k. The probability (P ₀ ) that no sequencing error occurs can be expressed by the following formula.

The probability (P _i ) of one error occurring in the i-th base of a base sequence of sequence length k can be expressed by the following formula.

Further, the probability (P _ij ) of two errors occurring in the i-th and j-th bases of a base sequence of sequence length k can be expressed by the following formula.

Similarly, the probability that errors occur in three or more bases between the "reference sequence" and the "query sequence" can also be calculated.

For example, if the i-th base differs by one location between the "reference sequence" and the "query sequence" regarding a base sequence of sequence length k, the base sequence will consist of four types: A, G, C, and T. Therefore, the number of cases where an arbitrary base changes to another base is 3. As mentioned above, when the probability that the reference sequence appears is _P0 , the probability that the query sequence will occur can be expressed by the following formula. Although the probability may change depending on the type of base substitution, we assumed here that all base substitutions occur with the same probability.

Furthermore, if the "reference sequence" and the "query sequence" differ in two places, for example, the i-th base and the j-th base, the probability that the query sequence will occur can be expressed by the following formula.

Similarly, the probability that a query sequence will occur when the "reference sequence" and the "query sequence" differ in three or more bases can also be calculated.

Here, if the appearance frequency (aggregation) of the standard sequence is _F0 , and the sum of the appearance frequencies (aggregation) of sequences derived from the standard sequence is _Ftotal , then the probability _P0 that the "standard sequence" appears is as follows. It can be expressed as

Furthermore, if the appearance frequency (aggregation) of a query sequence is F _query , the probability that a "query sequence" occurs can be expressed as follows.

The following formula is derived from [Math. 6] and [Math. 7].

[Equation 6] to [Equation 8] can be calculated from the appearance frequency (aggregation) of each sequence data in the unique sequence data group.

Once the value of the probability p _i that a sequencing error occurs in the i-th base is known, the formulas shown in [Equation 4] and [Equation 5] can be transformed to ``Probability that the query sequence occurs/Reference sequence is The probability of occurrence (P ₀ ) can be calculated. For example, if there is only one difference in the base sequence at the i-th base, the following equation is derived from [Equation 4] and [Equation 6].

Furthermore, when the base sequences differ in two locations, i.e., the i-th and j-th bases, the following formula is derived from [Math. 5] and [Math. 6].

Similarly, even if there are three or more differences in the nucleotide sequences between the "reference sequence" and the "query sequence", the "probability of the query sequence occurring/probability of the reference sequence occurring (P ₀ )" can be calculated. .

In determining whether or not the base sequence of the query sequence is caused by an error in sequencing the base sequence of the reference sequence, from the calculated value on the right side of [Math. 9] or [Math. 10], it is necessary to A threshold value (hereinafter also referred to as "threshold value (probability)") regarding "probability of occurrence/probability of reference sequence occurring (P ₀ )" can be set. The threshold value (probability) can be appropriately set to a value larger than the calculated value on the right side of [Equation 9] or [Equation 10]. For example, if the calculated value of [Equation 8] that can be calculated from the unique sequence data group is equal to or less than a threshold (probability), the query sequence is integrated into the reference sequence. On the other hand, if the calculated value of [Equation 8] is larger than the threshold (probability), the query array is not integrated into the reference array.

A method for deriving the probability p _i that a sequencing error will occur in the i-th base sequence will be described below. [Math. 1] to [Math. 10] have explained the probability p _i that a sequencing error will occur at the i-th base in a base sequence of sequence length k, and p _i is a value determined for each experiment. .
If there is only one difference in the base sequence between the reference sequence and the query sequence at the n-th base, the probability p _i that a sequencing error will occur at the i-th base is calculated as follows by transforming [Equation 2]. It is expressed by the formula.

Here, if F _i is the sum of the frequencies of appearance (aggregation) of base sequences in which the i-th base differs in only one place, P _i can be expressed by the following formula.

Further, p _i can be expressed by the following formula from [Equation 6], [Equation 11], and [Equation 12].

A method of calculating the probability p _i that a sequencing error will occur in the i-th base derived above will be explained below. First, the unique sequence data in the unique sequence data group are sorted in descending order of appearance frequency (aggregation), and the base sequence of the unique sequence data with the highest appearance frequency (aggregation) is set as a reference sequence. The appearance frequency (aggregation) of the reference array is F ₀ in [Equation 13]. Next, unique sequence data having a base sequence that is the same as the base sequence of the reference sequence but differs only in the i-th base is selected from the unique sequence data group (here, only the i-th base is Since there are four types of bases in different unique sequences, a maximum of three types of unique sequence data can be selected), and the frequency of appearance (aggregation) of these unique sequence data is summed. This total value becomes F _i in [Equation 13]. Since the values of F ₀ and F _i are obtained, p _i can be calculated by substituting these values into [Equation 13].

By repeating the process of integrating sequencing errors at the i-th base of a base sequence with sequence length k until i = 1 to k, the unique sequence with the highest frequency of occurrence (aggregated) in the unique sequence data group is obtained. The values of p _i (i=1 to k) for the data are obtained. Next, in the unique sequence data group, the same calculation as above is performed using the unique sequence data with the second highest frequency of occurrence (aggregated) as the reference sequence, and the p The value of _i (i=1 to k) is obtained. Similarly, calculations are made for the third, fourth, etc. unique sequence data, and the value of p _i is obtained for each unique sequence data set as the reference sequence.

The threshold value (probability) may be set based on the statistical values of many obtained values of p _i . For example, the average value or median value may be obtained from a plurality of p _i values after removing outliers from the many obtained p _i values, and a threshold value (probability) may be calculated from that value, or It may be set as appropriate based on the value. As used herein, "outlier" means a value that deviates significantly from a plurality of values. Outliers can be detected based on, but are not limited to, standard deviation, Smirnov-Grubbs, or Thompson test. In one example, outliers are detected by Smirnov-Grubbs.

(class division)
It is known that antibodies mature through a process of V(D)J recombination followed by a process of affinity maturation. During the process of V(D)J recombination, antibodies of various sequence lengths can be produced. During the process of affinity maturation, different antibodies with the same sequence length but different amino acid sequences can be generated during the process of affinity maturation. In order to track the maturation process of antibodies due to additional immune stimulation, we can classify the sequence data of antibody populations by sequence length and classify the antibody sequence data by V gene region and J gene region. , it may become possible to efficiently find sequence data indicative of antigen-specific antibodies.

The sequence data group preparation process includes, but is not limited to, the sequence data in the sequence data group based on either or both of the sequence length and the original sequence on the genome sequence of the animal species corresponding to the animal. Including the step of classifying. The classification step generates a classified sequence data group composed of classes containing corresponding sequence data.

Sequence data in a sequence data group is classified, for example, based on sequence length and combination of original sequences on the genome sequence of the animal species. Classification based on sequence length includes, for example, assigning each sequence data in a sequence data group to a class related to the corresponding sequence length based on the sequence length indicated in the sequence data. Classification based on the original sequence can be done, for example, by combining sequence data in a sequence data group with one or both of the V gene fragment and J gene fragment constituting the variable region of the antibody shown in the sequence data and corresponding to the animal. The method includes assigning each of the sequence data to a class related to a gene fragment having the highest degree of sequence identity (i.e., the original sequence) based on the degree of sequence identity obtained by a homology search with the genome sequence of the animal species. The degree of sequence identity is, for example, sequence identity or a score value regarding sequence identity.

(Exclusion of sequence data that does not form sequence clusters)
Considering the process of affinity maturation, an antigen-specific antibody that has undergone affinity maturation and exhibits higher affinity will have a sequence similar to that of the antibody when molecular phylogenetic tree analysis is performed, and It is thought that it forms a sequence cluster in the molecular phylogenetic tree together with a series of antibodies that have high affinity for each other. On the other hand, an antibody that has not undergone affinity maturation and exhibits relatively low affinity has few or almost no antibodies with sequences similar to that of the antibody, and when molecular phylogenetic analysis is performed, the antibody has a relatively low affinity. It is thought that no sequence clusters are formed in the phylogenetic tree. Therefore, by excluding sequence data of antibodies that are estimated not to form sequence clusters (hereinafter referred to as "sequence data that do not form sequence clusters") from the sequence data group, sequence data that has sequence information of antigen-specific antibodies can be It may be possible to find it with high efficiency.

The sequence data group preparation step includes, but is not limited to, removing non-sequence cluster forming sequence data from the sequence data group to generate a sequence cluster forming data group. Sequence cluster non-formable sequence data is excluded when the number of sequence data having a predetermined degree of sequence matching with respect to the sequence indicated by specific sequence data in the sequence data group is less than a threshold regarding the number of sequence data. In this case, the method includes the step of excluding the specific sequence data (sequence data that does not form a sequence cluster) from the sequence data group.

In one example, the predetermined degree of sequence match is sequence identity. When the sequence identity is 90% and the threshold for the number of sequence data is 20, the exclusion of sequence data that does not form a sequence cluster is defined as 90% sequence identity with the sequence indicated by specific sequence data in the sequence data group. If the number of sequence data having the same property is less than a threshold (20), the sequence data is excluded from the sequence data group. A sequence data group in which this exclusion process is repeated becomes a collection of sequence data in which there are 20 or more pieces of sequence data having sequences showing 90% sequence identity with a certain sequence. Such sequence data groups are expected to very efficiently contain sequence data indicative of antigen-specific antibodies that have undergone affinity maturation.

In other examples, the predetermined sequence match is a number of sequence mismatches. The number of sequence mismatches may be a value corresponding to 10% of the sequence length shown in the sequence data. For example, if the sequence length is 400 bp, 40, which is 10% of the sequence length, may be set as the number of sequence mismatch degrees. When the number of sequence discrepancies is 40 and the threshold value regarding the number of sequence data is 10 (U40<10), the exclusion of sequence data that does not form a sequence cluster is based on the sequence indicated by specific sequence data in the sequence data group. If the number of sequence data having up to 40 bases or amino acids that are mismatched is less than the threshold (10), that sequence data is excluded from the sequence data group. It is expected that the sequence data group in which this exclusion process is repeated will very efficiently contain sequence data indicative of antigen-specific antibodies that have undergone affinity maturation.

Although the threshold value regarding the number of sequence data is not limited, it may be a constant value, or it may vary depending on the frequency of appearance (aggregation) or frequency of appearance (integration) of the specific sequence data. . In one example, the threshold for the number of array data may be ten. In another example, the threshold for the number of array data may be 10% of the frequency of occurrence (aggregation) or frequency of occurrence (integration). In this example, if the appearance frequency (aggregation) is 100, the threshold for the number of array data is 10.

In Embodiment 1, the steps of preparing a sequence data group include cleaning up, aggregating into unique sequence data, integrating sequencing errors, classifying sequence data that do not form sequence clusters. However, the number and order of the steps included in the preparation step according to the first aspect are not limited thereto. The step of preparing a sequence data group may appropriately include at least one step, at least two steps, or at least three steps selected from the group consisting of the above-described steps. The steps described above can be included in various orders without particular limitation, as long as the step of aggregating into unique sequence data is performed before the step of integrating sequencing errors. The step of aggregating into unique sequence data is preferably performed before the step of excluding sequence data that does not form sequence clusters. In the step of preparing a sequence data group, when a sequence data group in which the above-described steps have been performed is obtained, the sequence data group may include only the step of obtaining the sequence data group.

<Creation of molecular phylogenetic tree>
The identification method according to Embodiment 1 includes the step of creating a molecular phylogenetic tree from the sequence data group. The molecular phylogenetic tree was calculated by calculating a distance matrix from the sequences shown in the sequence data in the sequence data group using known software (the ape package of the statistical analysis software "R"), although it is not limited to this. It can be created from a distance matrix by using the neighbor-joining method.

<Determination of whether a sequence cluster contains sequence data indicating an antigen-specific antibody>
The identification method according to Embodiment 1 is based on a change in the degree of sequence identity between the antibody sequence shown in sequence data in a sequence cluster in a molecular phylogenetic tree and the original sequence of the antibody, or a change in the type of antibody. The method includes a step of determining whether the cluster includes sequence data indicating an antigen-specific antibody that is a candidate for antibody production. Through this determination step, one or more sequence clusters containing sequence data indicating antigen-specific antibodies that are candidates for antibody production are identified.

The original sequence is a sequence that has the highest sequence identity between the antibody sequence shown in the sequence data and the genome sequence of the animal species corresponding to the animal from which the antibody was obtained. Although the original sequence is not limited, the degree of sequence identity between each of the V gene fragment and J gene fragment constituting the variable region of the antibody and the genome sequence of the animal species corresponding to the animal from which the antibody was obtained is determined by BLAST. This is the gene fragment with the highest sequence identity (%) obtained through the search.

The change in the degree of sequence identity in the determination process is, for example, the change in the degree of sequence identity per unit time (= "slope of the degree of sequence identity"), and the slope of the degree of sequence identity is less than the threshold regarding the slope of the degree of sequence identity. In this case, it is determined that the sequence cluster contains sequence data indicating an antigen-specific antibody. The change in the type of antibody in the determination step is, for example, a class switch of the antibody, and if the class switch of the antibody exceeds a predetermined number of times during a certain period, the sequence data indicates that the sequence cluster is an antigen-specific antibody. It is judged that it contains.

The determination step includes, for example, whether the sequence identity of the antibody shown in the sequence data in the sequence cluster in the molecular phylogenetic tree and the original sequence of the antibody before antigen administration is less than a threshold regarding the sequence identity before antigen administration. In this case, the method may include determining that the sequence data is a sequence cluster that does not include a sequence-specific antibody.

The determination step may further include the step of extracting sequence clusters from the molecular phylogenetic tree. In one example, a sequence cluster is defined as a sequence cluster extending outward from the center of the molecular phylogenetic tree from a predetermined position between the branch points when the length of the branch connecting the branch points in the molecular phylogenetic tree is greater than or equal to a predetermined value. The region may be extracted as a sequence cluster.

Although the molecular phylogenetic tree of Embodiment 1 was created for the sequence data group prepared in the sequence data group preparation step, the invention according to the first aspect is not limited to this. The creation of a molecular phylogenetic tree may be performed again for a collection of sequence data included in a sequence cluster determined to include sequence data indicating an antigen-specific antibody that is a candidate for antibody production. A step of determining whether the sequence cluster in this re-created molecular phylogenetic tree includes sequence data indicating an antigen-specific antibody that is a candidate for antibody production may be further performed.

[Embodiment 2]
A second aspect of the present invention provides a determination system for determining whether a sequence cluster includes sequence data indicating an antigen-specific antibody that is a candidate for antibody production. One embodiment of the second aspect of the present invention (hereinafter referred to as "Embodiment 2") will be described below.

The determination system includes a control section and a storage section. The storage unit stores a database including sequence data groups including sequence data of antibodies obtained from animals administered with antigens. The control unit includes a molecular phylogenetic tree creation unit that creates a molecular phylogenetic tree from the sequence data group, and a sequence match between the sequence of the antibody shown in the sequence data in the sequence cluster in the molecular phylogenetic tree and the original sequence of the antibody. and a sequence data determination unit that determines whether the sequence cluster is a sequence cluster containing sequence data indicating an antigen-specific antibody that is a candidate for antibody production, based on a change in degree or a change in type of antibody.

<Control unit and storage unit>
The determination system according to the second embodiment includes a control section and a storage section.
FIG. 7 is a block diagram showing a schematic configuration of the determination system 1 according to the second embodiment. The determination system 1 includes a control unit 11 that controls its overall operation, and a storage unit 13 that stores a database 107. The determination system 1 is connected to an input device 15 through which a user performs input, a display device 17 that performs screen display, and a measuring device 20.

The control unit 11 includes a processing circuit corresponding to a processor such as a CPU (central processing unit), and a memory (main storage device). The processor of the control unit 11 executes a computer program loaded into memory. The control unit 11 implements the functions of a sequence data group preparation unit 101, a molecular phylogenetic tree creation unit 103, a sequence data determination unit 105, etc., which will be described later, by executing a predetermined computer program. The control unit 11 may realize a predetermined function in cooperation with software, or may realize a predetermined function only by hardware.

The storage unit 13 is an auxiliary storage device, and is, for example, a hard disk drive (HDD) or a solid state drive (SSD). For example, a computer program is stored in the storage unit 13. Computer programs are loaded into memory and executed by a processor. The computer programs include an operating system and application programs.

The application programs include a sequence data group preparation program that functions the sequence data group preparation section, a calculation program that calculates various parameters including sequence identity, a molecular phylogenetic tree creation program that functions the molecular phylogenetic tree creation section, and a molecule. It includes a sequence cluster extraction program that extracts sequence clusters from a phylogenetic tree, a sequence data determination program that makes the sequence data determination section function, a threshold acquisition program that acquires a threshold value, and a display program that displays determination results.

In the second embodiment, the database 107 is stored in the storage unit 13.
FIG. 8 is a block diagram showing a schematic configuration of the database 107 according to the second embodiment. The database 107 stores sequence data groups 110, threshold data 120, function data 130, and animal genome sequence data 140.

Sequence data group 110 includes a sequence data group after sequencing. The sequence data of the sequence data group includes information regarding the base sequence encoding the variable region portion of the antibody, information regarding the sequence length, and a quality score (Q score). The sequence data may further include information regarding the type of antibody (IgG2 or IgG3), information regarding the time of blood collection, and information regarding the animal from which the blood was collected. The sequence data group 110 may further include a clean-up sequence data group, a unique sequence data group, an integrated unique sequence data group, a classification sequence data group, and a sequence cluster-forming data group, which will be described later.

The threshold value data 120 includes a threshold value (hereinafter referred to as "threshold value (degree of identity)") regarding a change in the degree of sequence identity of an antibody sequence shown in sequence data in a sequence cluster in a molecular phylogenetic tree described below. The threshold value (degree of matching) includes, for example, a predetermined value regarding the slope of the degree of sequence matching per change in time at which sequence data of each antibody was acquired. The threshold is a threshold related to the quality score of sequencing, and a threshold related to "probability of query sequence occurring/probability of reference sequence occurring (P ₀ )" used when integrating sequencing errors (hereinafter referred to as "threshold (probability)"). ) may further include a threshold regarding the number of sequence data used to exclude sequence data that does not form a sequence cluster.

［式中、Ａは動物のゲノム塩基配列に関する情報を示し、Ｂは配列データの塩基配列に関する情報を示す。］。この例において、ゲノム配列の塩基（Ａ）と対応する配列データの塩基（Ｂ）とが一致する場合（α Ａ＝Ｂ）は１点が付与され、不一致の場合（β Ａ≠Ｂ）は－３が付与される。このスコア関数を以下の表に示す：

The function data 130 includes score function data that scores the degree of sequence matching in a homology search for base sequences or amino acid sequences. A scoring function for scoring matches and mismatches in base sequences is, for example, expressed by the following formula:

[In the formula, A represents information regarding the genome base sequence of the animal, and B represents information regarding the base sequence of the sequence data. ]. In this example, if the base (A) of the genome sequence matches the base (B) of the corresponding sequence data (α A = B), 1 point is given, and if they do not match (β A≠B), - 3 is given. This scoring function is shown in the table below:

A scoring function that scores matches and mismatches in amino acid sequences is, for example, a scoring matrix such as a PAM matrix (eg, PAM250), a BLOSUM matrix (eg, BLOSUM50), and the like. The data included in the function data 130 may further include a probability density function related to Poisson distribution and a function used for regression analysis.

The animal genome sequence data 140 includes genome sequence data regarding animal species including alpacas. Alpaca genome sequence data includes sequences encoding the antibody V and J genes. Animal genome sequence data 140 may further include human genome sequence data.

When the database 107 is stored in a storage medium such as a magnetic disk, an optical disk, a magneto-optical disk, or a flash memory, the storage unit 13 includes a drive device that reads/writes information to and from the storage medium, and a drive device that reads/writes information to and from the storage medium. It may be composed of.

The input device 15 is constituted by a device or device for inputting necessary information by the user (specifying information specifying the address where the array data group is stored) and instruction information to the control unit 11. The input device 15 has a function of outputting input information to the control unit 11 via an interface. Input device 15 may be, for example, a keyboard, a mouse, and a voice recognition device. The types of input information are not limited to these. The input information may include, for example, information regarding the animal from which blood was collected.

The instruction information includes information that instructs the processing steps. The information instructing the processing steps includes, for example, regarding the processing in the sequence data group preparation section described below, from the sequence data group after sequencing, cleanup processing, aggregation processing into unique sequences, integration processing of sequencing errors, It includes instruction information regarding the order in which any or all of the classification process and the sequence data non-forming sequence data exclusion process should be executed.

The control unit 11 acquires sequence data groups after sequencing from the sequence data groups 110 stored in the database 107 in accordance with instructions input from the input device 15, and determines the possibility of forming sequence clusters in the molecular phylogenetic tree described below. A function to prepare a sequence library including generating a collection of sequence data (hereinafter referred to as a "sequence cluster-forming sequence data group"); a function to create a molecular phylogenetic tree from the sequence data group; a function to create a molecular phylogenetic tree from the sequence data group; It has a function of determining whether a sequence cluster of contains an antigen-specific antibody; and a display function of displaying the determination results.

The functions of the control unit 11 are realized by executing a corresponding application program loaded into the memory of the control unit 11 in a processing circuit including a processor of the control unit 11.

The display device 17 is configured by a device that allows the user to perceive the determination result of the control unit 11, and may be a display or a printer, for example.

The measuring device 20 is composed of a device for measuring gene sequences. The measuring device 20 may be, for example, a next-generation DNA sequencer. The measuring device 20 has a function of outputting measurement data to the control unit 11 via an interface.

In FIG. 7, a solid line connecting the measuring device 20 and the control unit 11 is constituted by a transceiver including an interface that functions to transmit and receive data and signals by wire or wirelessly between these elements. In FIG. 7, the same applies to solid lines connecting each element other than these.

In the second embodiment, the database 107 is stored in the storage unit 13 inside the determination system 1, but the location where the database is stored is not limited to this. The database 107 may be stored in a storage device that exists outside the determination system 1, for example. Such a storage device may be configured, for example, by all or part of a storage medium such as an optical disk, and the storage device may be provided in a server connected to the determination system according to the second aspect via a network. .

In the second embodiment, the determination system 1 is connected to an external measuring device 20, but the determination system according to the present invention is not limited to this. For example, the measuring device 20 may be present inside the determination system 1. The measurement data obtained using the measurement device 20 may be read into the control unit 11 and stored in the database 107, for example. The determination system 1 may be configured such that it is not connected to the measuring device 20 and the measurement data is read into the control unit 11.

The operation of the determination system according to the second embodiment will be explained using FIGS. 9 to 11.
A user uses the input device 15 (FIG. 7) to input instructions to the control unit 11 and necessary information (designation information specifying addresses where various data are stored and information regarding the animal from which blood was collected).

FIG. 9 is a flowchart showing determination of whether a sequence cluster in a molecular phylogenetic tree includes sequence data indicating an antigen-specific antibody in the determination system according to the present invention.

<Sequence data group preparation department>
The determination system according to the second embodiment includes a sequence data group preparation unit that generates a predetermined sequence data group.
When the processor of the control unit 11 receives an instruction and necessary information from the user via the input device 15, it prepares an array data group (S11). In array data group preparation S11, the processor loads the array data group preparation program stored in the storage unit 13 into memory and executes it. Thereby, the array data group preparation section 101 is realized in the control section 11 (FIG. 7).

FIG. 10 is a flowchart showing preparation of sequence data groups in sequence library preparation S11. The array data group preparation unit 101 refers to the designation information input by the user and executes array data group acquisition processing and the like.

(Getting array data group)
The sequence data group preparation unit 101 refers to the designation information input by the user and loads into memory the sequence data group after the sequence determination stored in the sequence data group 110 in the database 107 (S31). The sequence data constituting the sequence data group after sequencing includes information regarding the base sequence encoding the variable region portion of the antibody, information regarding the sequence length, information regarding the blood collection period (weeks), and quality score (Q score). include.

The array data group preparation unit 101 refers to the instruction information input by the user, executes the instructed process, and generates an array data group according to the process. In Embodiment 2, the instruction information includes generation of a clean sequence data group by cleanup processing, generation of a unique sequence data group by aggregation processing into unique sequences, and sequencing errors for the sequence data group after sequencing. Instructions to execute in the following order: generation of an integrated unique sequence data group by the integration process, generation of a classified sequence data group by the classification process, and generation of a sequence cluster-forming sequence data group by the process of excluding sequence data that does not form a sequence cluster. , and an instruction to store the array data group generated by each process in the database 107.

(Cleanup processing)
When the sequence data group after sequence determination is acquired, the sequence data group preparation unit 101 executes removal (cleanup) of sequence data that does not indicate an amino acid sequence (S32). As a result, a clean-up sequence data group is generated, and the generated sequence data group is stored in the sequence data group 110 in the database 107.

In cleanup S32, the sequence data group preparation unit 101 acquires information regarding the sequence length of the base sequence encoding the variable region portion of the antibody shown in the sequence data in the sequence data group after sequencing, and determines whether the sequence length is If the number is a multiple of 3 (3k), a selection tag is assigned, and if the number is not a multiple of 3 (3k+1 or 3k+2), an exclusion tag is assigned. The sequence data group preparation unit 101 performs selection/exclusion tagging on all sequence data in the sequence data group after sequencing. When the sequence data group preparation unit 101 finishes tagging all the sequence data in the sequence data group after sequencing, it ends the selection/exclusion tagging.

After the tagging is completed, the sequence data group preparation unit 101 selects the sequence data to which the selection tag has been attached from the sequence data group after sequencing, and generates a clean-up sequence data group composed of the selected sequence data. The generated cleanup sequence data group is stored in sequence data group 110 in database 107.

When the nucleotide sequence encoding the variable region portion is a multiple of 3, the sequence represents an amino acid sequence. The sequence data in the cleanup sequence data group may further include information regarding amino acid sequences.

In Example 3, regarding the cleanup process, a process of removing sequence data that does not indicate an amino acid sequence was described, but the cleanup process is not limited to this. The cleanup process may be, for example, a cleanup process based on a quality score included in the sequence data. Additionally, the cleanup process may be a combination of removal of sequence data that does not represent an amino acid sequence and removal based on quality scores.

In the second embodiment, the cleanup process is executed by selecting array data to which a selection tag is attached, but the contents of the cleanup process are not limited to this. The cleanup process may be performed, for example, by excluding sequence data to which an exclusion tag has been attached from a sequence data group after sequencing. In Example 3, the cleanup process was performed on base sequences, but the type of sequence on which the cleanup process is performed is not limited to this. Cleanup processing can also be performed on amino acid sequences.

(Aggregation into unique sequence data)
When the cleanup sequence data group is generated, the sequence data group preparation unit 101 executes a process of aggregating multiple sequence data indicating the same sequence into one sequence data (= "unique sequence data") (S33 ). As a result, a unique sequence data group is generated, and the generated sequence data group is stored in the sequence data group 110 in the database 107.

In the process of aggregation into unique sequence data, the sequence data group preparation unit 101 assigns an aggregation tag (α) to one sequence data in the cleanup sequence data group, and acquires information regarding the sequence indicated by the sequence data. . The sequence data group preparation unit 101 assigns the same aggregate tag (α) to other sequence data indicating the same sequence as the acquired sequence. When the sequence data group preparation unit 101 finishes attaching the aggregate tag (α) to all the sequence data indicating the same sequence as the aforementioned sequence in the cleanup sequence data group, it finishes tagging with the aggregate tag (α). .

The sequence data group preparation unit 101 assigns an aggregate tag (β) different from the aggregate tag (α) to one piece of sequence data indicating a sequence different from the sequence indicated by the sequence data to which the aggregate tag (α) has been attached. However, similarly to the above, the tagging with the aggregate tag (β) ends when all of the sequence data showing the same sequence as the above sequence in the cleanup sequence data group is tagged. Further, the sequence data group preparation unit 101 sequentially tags array data indicating a different sequence from the sequence indicated by the sequence data to which the aggregate tag (β) has been added with another aggregate tag (γ, . . . ). . When the number of array data in the cleanup array data group matches the number of array data to which aggregate tags have been added, the array data group preparation unit 101 ends the assignment of aggregate tags.

The sequence data group preparation unit 101 then converts the plurality of sequence data tagged with the same type of aggregation tag into one sequence data (for example, any sequence data tagged with the aggregation tag or a newly generated sequence). data) to generate unique array data. In addition to the information included in the sequence data, the unique sequence data further includes information regarding frequency of appearance (aggregation).

In Example 3, the aggregation process into unique sequence data was performed on base sequences, but the type of sequence on which unique sequence data is processed is not limited to this. Aggregation processing into unique sequence data can be performed on amino acid sequences.

(Consolidation of sequencing errors)
When the unique sequence data group is generated, the sequence data group preparation unit 101 executes processing to integrate sequence determination errors (S34). Through this process, an integrated unique sequence data group is generated, and the generated sequence data group is stored in the sequence data group 110 in the database 107.

FIG. 11 is a flowchart illustrating sequencing error consolidation steps.
(a) Generation of sequence data group for integration The sequence data group preparation unit 101 acquires information regarding frequency of occurrence (aggregation) from all unique sequence data in the unique sequence data group, and selects the frequency of occurrence (aggregation) with the highest frequency of occurrence (aggregation). When the appearance frequency (aggregation) of unique sequence data is 2 or more (S41; YES), the sequence data group preparation unit 101 executes generation of a sequence data group for integration (S42). As a result, a sequence data group for sequence integration is generated, and the sequence data used to generate the sequence data group for sequence integration is excluded from the unique sequence data group.

In step S42 of generating an integrated sequence data group, the sequence data group preparation unit 101 selects unique sequence data (hereinafter referred to as "reference sequence data α") that exhibits the highest frequency of appearance (aggregation).
The sequence data group preparation unit 101 loads the calculation program stored in the storage unit 13 into memory and realizes a function of generating and calculating various parameters including sequence matching (hereinafter referred to as “generation & calculation function”). . The sequence data group preparation unit 101 that realizes the generation and calculation function generates a base sequence of unique sequence data (hereinafter also referred to as "query sequence") with the same sequence length as the base sequence of standard sequence data α (hereinafter also referred to as "standard sequence α"). ), and generates information on the number of bases that have mismatched bases, information on the types of base substitutions, and information on the number of base substitutions, and generates information on the sequence compared with the reference sequence data α. Assigned to data (also referred to as "query sequence data"). The array data group preparation unit 101 executes the information generation and provision process for all array data in the unique array data group.

から算出することができる。 Next, the sequence data group preparation unit 101 selects sequence data having the number of base substitutions that could have occurred due to an error in sequencing from the reference sequence α. The sequence data group preparation unit 101 acquires information regarding the appearance frequency (aggregation) of the reference sequence data α and the probability density function regarding the Poisson distribution stored in the function data 130 in the database 107, and calculates the obtained appearance frequency (aggregation) and The minimum number of base substitutions (m) for which the appearance frequency (aggregate) is 1 or less is calculated from the probability density function related to Poisson distribution. For example, if the expected value of the number of base substitutions is 0.3, the probabilities that the number of base substitutions will be 0, 1, 2, 3, and 4 are 0.74, 0.22, 0.03, and 0.003, respectively. , 0.0003. In this case, if the appearance frequency (aggregate) of the standard sequence data α is 100, the appearance frequencies of sequences with 0, 1, 2, 3, and 4 base substitutions are 100, 30, 4.5, and 0.5, respectively. , 0.03, and m is set to 3. Similarly, when the appearance frequency (aggregation) of the reference array data α is 1000, m is set to 4. The expected value of the number of base substitutions is calculated using the above-mentioned formula regarding the probability p _i that a sequence determination error will occur at the i-th base:

It can be calculated from

When the number of base substitutions (m) is set, the sequence data group preparation unit 101 that realizes the generation and calculation function refers to the information regarding the number of base substitutions added to the unique sequence data, and selects up to m bases. Sequence data having substitutions (number of base substitutions n=0, 1, . . . m) are selected to generate a sequence data group for integration. The sequence data group preparation unit 101 excludes the unique sequence data selected for generation of the integrated sequence data group from the unique sequence data group.

(b) Generation of integrated unique sequence data When a sequence data group for integration is generated, the sequence data group preparation unit 101 generates sequence data to be integrated (hereinafter referred to as "integrated sequence data") from the sequence data in the integrated sequence data group. (S43).
FIG. 12 is a flowchart showing selection S43 of array data for integration. The sequence data group preparation unit 101, which realizes the generation and calculation function, collects information added to the unique sequence data (information regarding the number of the base that has become a mismatch, information regarding the type of base substitution, and information regarding the base substitution). (information regarding the number of occurrences), and calculates the ratio of "probability of occurrence of query sequence/probability of occurrence of reference sequence (P ₀ )" detailed in Embodiment 1 (S51). The calculated ratio may be added to the unique sequence data.

When the ratio is calculated, the array data group preparation unit 101 loads the threshold value acquisition program stored in the storage unit 13 into the memory to realize a function of acquiring the threshold value. The sequence data group preparation unit 101 that realizes the threshold value acquisition function obtains a threshold value (hereinafter referred to as "threshold value (probability)") related to "probability that the query sequence occurs/probability that the reference sequence occurs (P ₀ )" stored in the threshold data 120. (S52).

When the threshold value (probability) is acquired, the array data group preparation unit 101 loads the array data determination program stored in the storage unit into the memory to implement a function of determining array data. The sequence data group preparation unit 101 that realizes the sequence data determination function selects the unique sequence data to be integrated when the ratio given to the unique sequence data is less than the acquired threshold (probability) (S53; YES). An integrated tag indicating that the unique sequence data is present is added to the unique sequence data (S54).

The sequence data group preparation unit 101 that realizes the sequence data determination function determines that the unique sequence data to be integrated is A non-integrated tag indicating that the unique sequence data is not integrated is added to the unique sequence data (S55). When the sequence data group preparation unit 101 finishes tagging all the unique sequence data in the sequence data group for integration, it ends the assignment of integration/non-integration tags.

When the sequence data group preparation unit 101 finishes attaching the integrated/non-integrated tags, the sequence data group preparation unit 101 returns to the flowchart of FIG. With reference to the tag, the unique sequence data to which the integration tag has been added is integrated into the reference sequence data α (S44). As a result, integrated unique sequence data regarding the reference sequence α is generated, and the generated sequence data is stored in the database 107. The integrated unique sequence data related to the reference sequence α includes, in addition to the information included in the reference sequence data α, information regarding the sum of the frequencies of occurrence (aggregated) included in the integrated unique sequence data (hereinafter referred to as "frequency of occurrence (integrated)"). ).

When the generation of integrated unique sequence data is completed, the sequence data group preparation unit 101 refers to the non-integrated tag attached to the unique sequence data in the integrated sequence data group, and generates the unique sequence data ( That is, the unique sequence data that was not used to generate the integrated unique sequence data in S44 is returned to the unique sequence data group (S45).

The sequence data group preparation unit 101 acquires information regarding frequency of occurrence (aggregation) from all unique sequence data in the unique sequence data group excluding the unique sequence data used to generate the integrated unique sequence data regarding the reference sequence α. However, if the appearance frequency (aggregation) of the unique sequence data with the highest appearance frequency (aggregation) is 2 or more (S41; YES), S42 to S45 are repeated.

In the selection of sequence data for integration in the second embodiment, S52 of obtaining a threshold value (probability) is executed following S51 of calculating a ratio, but the order of steps is not limited to this. S51 and S52 may be executed in any order, or may be executed simultaneously, as long as they are executed before S53, which compares the calculated ratio and the acquired threshold (probability), is executed. good.

(c) Generation of integrated unique sequence data group The sequence data group preparation unit 101 generates the frequency of appearance ( If the appearance frequency (aggregate) of the unique sequence data with the highest frequency of occurrence (aggregate) is less than 2 (S41; NO), the integrated unique sequence data and integrated unique sequence stored in the database are obtained. All unique sequence data in the unique sequence data group from which unique sequence data not used for data generation have been excluded is loaded into memory to generate an integrated unique sequence data group (S46). As a result, an integrated unique sequence data group is generated, and the generated integrated unique sequence group is stored in the sequence data group 110 in the database 107.

In the integration of sequencing errors in the second embodiment, generation S46 of an integrated unique sequence data group is executed after it is determined that the largest number of occurrences (aggregation) in the unique sequence data group is less than 2 (S41; NO). However, the order of steps is not limited to this. S46 may be executed simultaneously with S44 of generating integrated unique sequence data, for example, after it is determined that the largest number of occurrences (aggregation) in the unique sequence data group is 2 or more (S41; YES).

Although S45 of returning the unique sequence data that was not used in the generation of the integrated unique sequence data to the unique sequence data group in the second embodiment was executed after the generation of the integrated unique sequence data S44, the order of the steps is not limited to this. S45 may be executed after S43 of selecting array data for integration and before S41 for determining whether to repeat S42 to S45 is executed. S45 may be executed, for example, before the generation of integrated unique sequence data S44, or at the same time as S44.

In the integration of sequencing errors in Embodiment 2, the generation of integrated unique sequence data S44 was performed after the selection of sequence data for integration S43, but the order of the steps is not limited to this. S43 may be executed, for example, after it is determined that the largest number of occurrences (aggregation) in the unique sequence data group is less than 2 (S41; NO). In this case, the integration tag attached to the unique sequence data is attached with information regarding the type of reference sequence compared (for example, reference sequence data α), and in the integration of unique sequence data, the attached reference sequence Unique sequence data with the same type information are integrated.

Although S41 of determining whether the largest number of occurrences (aggregation) in the unique array data group of the second embodiment is two or more is performed prior to S42 to S45, the order of the steps is not limited to this. S41 may be executed after S45 is executed, for example, in order to determine whether to repeat S42 to S45. That is, S41 does not need to be executed before S42, which generates the first integration array data group.

In the integration of sequencing errors in the second embodiment, generation S42 of the sequence data group for integration is executed, but the present invention is not limited to this. In one example, in the integration of sequencing errors, the generation of the sequence data group for integration S42 may not be performed. In this example, in the integration of sequencing errors, the reference sequence data is sequence data that has a sequence with the same sequence length as the reference sequence of the reference sequence data.

(class division)
Returning to the flowchart showing the preparation of sequence data groups in FIG. 10, when the integrated unique sequence data group is generated, the sequence data group preparation unit 101 stores information regarding the sequence length, original V gene fragment, and original J gene fragment. Classification based on this is executed (S35).

(a) Classification based on sequence length In classification based on sequence length, the sequence data group preparation unit 101 acquires information regarding the sequence length of each sequence data in the integrated unique sequence data group, and A sequence length classification tag related to the sequence length is assigned to each sequence data indicating the sequence.
When the number of array data in the array data group matches the number of times the array length classification tag has been added, the array data group preparation unit 101 finishes adding the sequence length classification tag.

(b) Classification based on the original V gene fragment When the sequence data group preparation unit 101 finishes assigning sequence length classification tags, the sequence data group preparation unit 101 configures the antibody variable region shown in each sequence data in the integrated unique sequence data group. A V gene fragment (="original V gene fragment") on the genome sequence corresponding to the V gene fragment to be used is estimated, and a tag related to the original V gene fragment is added to the sequence data.

In estimating the original V gene fragment, the sequence data group preparation unit 101 loads the calculation program stored in the storage unit 13 into memory to realize a function of calculating the degree of sequence matching. The sequence data group preparation unit 101 that realizes the generation and calculation function refers to the information about the animal whose blood was collected from the input information input by the user, and selects the corresponding animal species stored in the animal genome sequence data 140 in the database 107. Obtain genome information.

Animal genome sequence data includes regions encoding multiple types of V gene fragments, depending on the animal species. In estimating the original V gene fragment, the sequence data group preparation unit 101 that realizes the generation and calculation function encodes sequence information including the V gene region of the antibody shown in the sequence data and each V gene fragment on the genome sequence. Calculate the degree of sequence matching with multiple regions. The degree of sequence identity is calculated, for example, by subtracting the value obtained by dividing the number of mismatches by the sequence length from 1 ([degree of sequence identity]=1-[number of mismatches]/[sequence length]).

The sequence data group preparation unit 101 determines the V gene fragment shown in the region on the genome sequence showing the highest degree of sequence identity among the calculated degrees of sequence identity as the original V gene fragment. The sequence data group preparation unit 101 adds a V gene fragment classification tag regarding the original V gene fragment to the sequence data. When the number of sequence data in the sequence data group matches the number of times V gene fragment classification tags have been added, the sequence data group preparation unit 101 finishes adding V gene fragment classification tags.

(c) Classification based on the original J gene fragment When the sequence data group preparation unit 101 finishes assigning the V gene fragment classification tags, the sequence data group preparation unit 101 assigns the classification tags based on the original J gene fragments in the same manner as described above. A J gene fragment (="original J gene fragment") on the genome sequence is identified, and a J gene fragment classification tag is attached to the sequence data. When the number of sequence data in the sequence data group matches the number of times J gene fragment classification tags have been added, the sequence data group preparation unit 101 finishes adding J gene fragment classification tags.

(d) Generation of Classified Sequence Data Group When the assignment of classification tags is completed, the sequence data group preparation unit 101 generates a class corresponding to the combination of sequence length, type of original V gene fragment, and type of original J gene fragment. generate. The array data group preparation unit 101 assigns array data to which various classification tags are attached to corresponding classes, and generates a classified array data group. Thereby, the generated classification sequence data group is stored in the sequence data group 110 in the database 107. Each class includes sequence data indicating sequences having a sequence length corresponding to the class and derived from the corresponding original V gene fragment and original J gene fragment.

In the classification of Embodiment 2, classification tags were assigned in the order of sequence length-based classification, original V gene fragment-based classification, and J gene fragment-based classification, but the classification tags were The order in which the assignments are performed is not limited to this. For example, these classification tags may be added at the same time. The assignment of the classification tag is performed by assigning at least one classification tag selected from the group consisting of assignment of a sequence length classification tag, assignment of a V gene fragment classification tag, and assignment of a J gene fragment classification tag. It's fine.

(Generation of sequence cluster-forming sequence data groups)
When the classified data group is generated, the sequence data group preparation unit 101 excludes sequence data that does not form a sequence cluster from the sequence data groups in each class (S36). As a result, a sequence cluster-forming sequence data group is generated, and the generated sequence data group is stored in the sequence data group 110 in the database 107.

In S36 for excluding sequence data that does not form a sequence cluster, the sequence data group preparation unit 101 refers to the specification information input by the user and sets a threshold value regarding the number of base substitutions (for example, 30 ) and a threshold value (for example, 50) regarding the number of array data.

When the acquisition of the threshold value is completed, the array data group preparation unit 101 acquires the array information of each array data in an arbitrary class (for example, a class containing a large number of array data) in the classified array data group. The array data group preparation unit 101 realizes the generation and calculation function described above, selects one array data in the class (for example, array data with a high frequency of appearance (integration)), and combines the selected array data with the array. The number of base substitutions with other sequence data in the class is calculated.

The sequence data group preparation unit 101 counts the number of sequence data indicating the number of base substitutions that is less than the threshold regarding the number of base substitutions, and determines whether the counted number of sequence data is less than or equal to the threshold regarding the number of sequence data. Determine whether When the number of counted sequence data is less than the threshold regarding the number of sequence data, the sequence data group preparation unit 101 excludes the sequence data from the class as sequence data that does not form a sequence cluster.

The sequence data group preparation unit 101 executes the exclusion of the sequence data that does not form a sequence cluster for each sequence data in the class. If the number of exclusion processes for sequence data that does not form a sequence cluster in one class matches the number of sequence data in the class, the sequence data group preparation unit 101 removes the sequence data that does not form a sequence cluster for that class. End the exclusion process.

Sequence data constituting a class for which the process of excluding sequence data that does not form a sequence cluster is a collection of sequence data that may form a sequence cluster in the molecular phylogenetic tree described below ("sequence cluster forming sequence data"). ), and is a sequence cluster-forming sequence data group containing these sequence data. When the sequence data group preparation unit 101 finishes excluding sequence data that does not form a sequence cluster for one class of the classified sequence data group, the sequence data group preparation unit 101 similarly removes sequence data that does not form a sequence cluster for other classes of the classified sequence data group. Exclude sex sequence data.

In Embodiment 2, the sequence data group preparation unit 101 determines that when the number of sequence data included in a class in a classified sequence data group is less than the threshold regarding the number of sequence data, the class is classified as sequence cluster-forming sequence data. is excluded from the classification sequence data group as it does not contain it.

Although the cleanup process S32 in the second embodiment is executed on the sequence data group after sequencing, the present invention is not limited thereto. The cleanup process S32 may or may not be performed on a unique sequence data group, an integrated unique sequence data group, a classified sequence data group, or a sequence cluster-forming sequence data group.

Although the aggregation process S33 into unique sequence data in the second embodiment is performed on the sequence data group after cleanup, the present invention is not limited thereto. The aggregation process S33 into unique sequence data may or may not be performed on a sequence data group after sequencing, an integrated unique sequence data group, a classified sequence data group, or a sequence cluster-forming sequence data group. You don't have to.

Although the sequencing error integration S34 in the second embodiment was performed immediately after the unique sequence data group generation step, the present invention is not limited thereto. Sequencing error integration S34 may be performed in any order or may not be performed after the generation of the unique sequence data group has been performed.

Although the classification S35 in the second embodiment is performed on the integrated unique sequence data group, the classification is not limited thereto. The classification step S35 may or may not be performed on a post-sequencing sequence data group, a clean-up sequence data group, a unique sequence data group, an integrated unique sequence data group, a sequence cluster-forming sequence data group. You don't have to.

Although the generation S36 of the sequence cluster-forming sequence data group in the second embodiment was performed on the classification data group, the present invention is not limited thereto. Generation of sequence cluster-forming sequence data groups S36 may or may not be performed on sequence data groups after sequencing, clean-up sequence data groups, unique sequence data groups, and integrated unique sequence data groups. It's okay.

<Molecular phylogenetic tree creation department>
The determination system according to the second embodiment includes a molecular phylogenetic tree creation unit that creates a molecular phylogenetic tree from a sequence data group.
When the sequence data group is prepared, the processor of the control unit 11 executes creation of a molecular phylogenetic tree shown in FIG. 9 (S12). In molecular phylogenetic tree creation S12, the processor of the control unit 11 loads the molecular phylogenetic tree creation program stored in the storage unit 13 into memory and executes it. Thereby, a molecular phylogenetic tree creation section 103 is realized in the control section 11.

The molecular phylogenetic tree creation unit 103 performs multiple alignment on sequence data in the prepared sequence cluster-forming sequence data group, and calculates a distance matrix from the sequence data that has undergone multiple alignment, thereby creating a molecular phylogenetic tree. Create.

Although the molecular phylogenetic tree of the second embodiment was created by calculating a distance matrix, the method of creating the molecular phylogenetic tree is not limited to this. The molecular phylogenetic tree may be created using, for example, a maximum parsimony method based on minimizing the sum of the lengths of the branches of the phylogenetic tree.
The creation of the molecular phylogenetic tree in Embodiment 2 was performed on sequence cluster-forming sequence data groups, but is not limited thereto. Creation of a molecular phylogenetic tree may be performed on an integrated unique sequence data group or a classified sequence data group.

<Sequence data determination section>
The determination system according to Embodiment 2 determines the sequence based on a change in the degree of sequence identity between the sequence of an antibody shown in sequence data in a sequence cluster in a molecular phylogenetic tree and the original sequence of the antibody, or a change in the type of antibody. It includes a sequence data determining unit that determines whether the cluster is a sequence cluster containing sequence data indicating an antigen-specific antibody that is a candidate for antibody production.

(Sequence cluster extraction from molecular phylogenetic tree)
When the molecular phylogenetic tree is created, the processor of the control unit 11 loads into memory and executes a sequence data determination program that causes the sequence data determination unit stored in the storage unit 13 to function. Thereby, the array data determination section 105 is realized in the control section 11.

The array data determination unit 105 loads the array cluster extraction program stored in the storage unit 13 into memory and executes it. This realizes a sequence cluster extraction function. The sequence data determination unit 105, which has realized the sequence cluster extraction function using sequence data, extracts sequence clusters from the created molecular phylogenetic tree (S13). This provides sequence data in sequence clusters extracted from the molecular phylogenetic tree. The sequence data determination unit 105 that realizes the sequence cluster extraction function extracts a portion with a high density of sequence information in the created molecular phylogenetic tree as a sequence cluster.

(Calculating the slope of sequence matching of sequence data in a sequence cluster)
When a sequence cluster is extracted, the sequence data determination unit 105 calculates a change in sequence information of sequence data in the sequence cluster (for example, the slope of the sequence coincidence degree per unit time) (S14). In calculation S14 of the slope of the degree of sequence matching, the sequence data determination unit 105 loads the calculation program stored in the storage unit 13 into the memory and executes it. Thereby, the function of calculating the slope of the degree of sequence matching is realized in the sequence data determination unit 105. The sequence data determination unit 105 that realizes the generation and calculation function calculates the degree of sequence matching between the sequence information of each sequence data in the sequence cluster and the sequence information of the original V and J gene fragments included in the sequence data.

The sequence data determination unit 105 acquires information regarding the blood sampling period (weeks) from each sequence data in the sequence cluster, and uses the sequence matching degree calculated above to calculate the slope of the sequence matching degree over time.

(Determination of whether it is a sequence cluster containing antigen-specific antibodies)
When the slope of the sequence matching degree is calculated, the array data determination unit 105 loads the threshold value acquisition program stored in the storage unit 13 into the memory, and realizes a function of acquiring the threshold value. The array data determination unit 105 that has implemented the threshold value acquisition function acquires a threshold value (hereinafter referred to as "threshold value (slope)") regarding the slope of the sequence matching degree stored in the threshold value data 120 (S15).

When the threshold value (slope) is acquired, the sequence data determination unit 105 compares the calculated slope of the sequence matching degree with the acquired threshold value (slope), and determines if the slope of the sequence matching degree is smaller than the threshold value (slope). (S16; YES), a display tag indicating that sequence data indicating an antigen-specific antibody is included is attached to the sequence cluster showing the slope of the sequence identity (S17).

If the slope of the degree of sequence identity is equal to or greater than the threshold value (slope) (S16; NO), the sequence data determination unit 105 determines sequence data indicating an antigen-specific antibody for the sequence cluster showing the slope of the degree of sequence identity. A display tag indicating that it is not included is added (S18).

<Judgment result display>
When the display tag is assigned, the processor of the control unit 11 displays the determination result (S19). In determination result display S19, the processor executes the display program loaded into the memory from the storage unit 13 and outputs a video signal according to the display tag to the display device 17. The display device 17 may be, for example, a display. In this case, the display displays the determination result based on the input video signal.

In the second embodiment, the molecular phylogenetic tree creation S12 and the sequence cluster extraction S13 are executed once each, but the number of times they are executed is not limited to this. For example, for the molecular phylogenetic tree created by executing the molecular phylogenetic tree creation S12' again on the collection of sequence data in the sequence cluster extracted in the sequence cluster extraction S13, the sequence cluster extraction S13' is performed. may be executed.

In the second embodiment, calculation of the slope of the degree of sequence matching of the sequence data in the sequence cluster S14 and acquisition of the threshold value (slope) S15 are performed in this order, but the order of the steps is not limited to this. S14 and S15 need only be executed before the comparison S16 of the slope of the sequence matching degree and the threshold value (slope), and S14 and S15 may be executed in any order or may be executed simultaneously.

[Embodiment 3]
A third aspect of the invention provides a method of producing antigen-specific antibodies. One embodiment of the third aspect of the present invention (hereinafter referred to as "Embodiment 3") will be described below.

The production method according to Embodiment 3 includes the steps of: preparing a sequence data group containing sequence data of an antibody obtained from an animal that may have received immune stimulation; creating a molecular phylogenetic tree from the sequence data group; The sequence cluster becomes a candidate for antibody production based on a change in the degree of sequence identity between the sequence of the antibody shown in the sequence data in the sequence cluster in the molecular phylogenetic tree and the original sequence of the antibody or a change in the type of antibody. Step of determining whether sequence data indicating an antigen-specific antibody is included; selecting sequence data indicating an antigen-specific antibody that is a candidate for antibody production from sequence clusters determined to include sequence data indicating an antigen-specific antibody and producing an antibody having the amino acid sequence shown in the selected sequence data.

In the production method according to Embodiment 3, the step of preparing a sequence data group, the step of creating a molecular phylogenetic tree, and the step of determining whether a sequence cluster includes sequence data indicating an antigen-specific antibody includes the following steps: The features described in the identification method according to Form 1 and the features described in this specification are applied as appropriate.

<Selection of sequence data showing antigen-specific antibodies>
The production method according to Embodiment 3 includes the step of selecting sequence data indicating an antigen-specific antibody that is a candidate for antibody production from a sequence cluster determined to include sequence data indicating an antigen-specific antibody. Sequence data is, for example, sequence data in a sequence cluster that has been determined to contain sequence data indicating an antigen-specific antibody, and the sequence identity between the sequence indicated by the sequence data and its original sequence is below a threshold regarding sequence identity. may be selected if In other examples, the sequence data is the one that has the least sequence identity between the sequence indicated by the sequence data and its original sequence, or the sequence data for which the antibody was collected the latest and the sequence indicated by the sequence data. The sequence identity with the original sequence may be less than or equal to a threshold for sequence identity.

<Production of antibody having the amino acid sequence shown in the sequence data>
The production method according to Embodiment 3 includes the step of producing an antibody having an amino acid sequence shown in sequence data in a sequence cluster determined to include sequence data indicating an antigen-specific antibody. Although the antibody is not limited to, a plasmid is constructed by inserting a DNA fragment having a sequence in which the CDR of the selected sequence data is optimized according to the codon frequency of E. coli into a protein expression vector for E. coli. , is produced by introducing the plasmid into E. coli and expressing it.

[Embodiment 4]
A fourth aspect of the invention provides a method for selecting sequence data indicative of antigen-specific antibodies that are candidates for antibody production. One embodiment of the fourth aspect of the present invention (hereinafter referred to as "Embodiment 4") will be described below. Although the selection method according to Embodiment 4 is different from the identification method according to Embodiment 1 described above, since there are common parts in specific steps, here, the differences from the identification method according to Embodiment 1 will be explained. I will mainly explain the points.

The selection method according to Embodiment 4 differs from the identification method according to Embodiment 1 in that it includes the step of creating a molecular phylogenetic tree and determining whether sequence data indicating an antigen-specific antibody that is a candidate for antibody production is included. The process is not included as an essential process. The selection method according to the fourth embodiment includes, as an essential step, "exclusion of sequence data that does not form a sequence cluster", which is detailed in the identification method according to the first embodiment. The selection method according to Embodiment 4 further includes a step of selecting sequence data indicating an antigen-specific antibody from the sequence data group as an essential step.

The step of selecting the sequence data described above includes the step of comparing the degree of sequence matching between the sequence of the antibody shown in the sequence data in the sequence data group and the original sequence of the antibody with a threshold value. Although the above-described comparing step is not limited to, the degree of sequence identity (e.g., 88% identity) between the antibody sequence shown in the sequence data in the sequence data group and the original sequence of the antibody is determined by the threshold value. (for example, 90% identity), this includes selecting sequence data indicating an antigen-specific antibody that is a candidate for antibody production. The above-mentioned comparing step may further include, for example, comparing the time when the antibody was obtained (for example, the 5th time) and the threshold value (for example, the 4th time) in addition to comparing the degree of sequence identity and the threshold value. . In this example, in the selection process, if the time when the antibody was acquired is equal to or greater than the threshold value, and if the degree of sequence identity is less than the threshold value, the sequence data is used as sequence data indicating an antigen-specific antibody that is a candidate for antibody production. Including being elected as.

The selection method according to Embodiment 4 includes "cleaning up sequence data etc. that do not indicate an amino acid sequence", "aggregation into unique sequences", and "integration of sequencing errors" detailed in the selection method according to Embodiment 1. , and “classification”. The selection method according to the fourth aspect of the invention may further include a step of creating a molecular phylogenetic tree, and a step of determining whether or not sequence data indicating an antigen-specific antibody that is a candidate for antibody production is included. .

[Embodiment 5]
A fifth aspect of the present invention provides a selection system for selecting sequence data indicating antigen-specific antibodies that are candidates for antibody production. One embodiment (hereinafter referred to as "Embodiment 5") of the second aspect of the present invention will be described below. Although the selection method according to Embodiment 5 is a different system from the determination system according to Embodiment 2 described above, there are common parts in specific data processing, so here, the selection method according to Embodiment 2 will be explained. The differences will be mainly explained.

The selection system according to the fifth embodiment differs from the determination system according to the second embodiment in that it does not include a molecular phylogenetic tree creation section and a sequence data determination section as essential sections. The selection system according to the fifth embodiment includes, as an essential process, "generation of a sequence data group capable of forming a sequence cluster" in the "sequence data preparation unit" detailed in the determination system according to the second embodiment. The selection system according to the fifth embodiment further includes an array data selection section as an essential section.

The sequence data selection unit described above includes executing a process of comparing the degree of sequence matching between the sequence of the antibody shown in the sequence data in the sequence data group and the original sequence of the antibody with a threshold value. In the comparison process described above, the sequence data selection unit determines, but is not limited to, the degree of sequence identity between the antibody sequence shown in the sequence data in the sequence data group and the original sequence of the antibody (for example, 88%). If the degree of match) is lower than the threshold value (e.g., 90% match), a selection tag indicating that the sequence data represents an antigen-specific antibody that is a candidate for antibody production is added to the sequence data. include. For example, in addition to comparing the degree of sequence identity and a threshold value, the above-described comparing process further includes performing a process of comparing the time when the antibody was obtained (for example, the 5th time) and the threshold value (for example, the 4th time). That's fine. In this example, the sequence data selection unit uses the sequence data to identify antigen-specific antibodies that are candidates for antibody production when the time when the antibody was acquired is equal to or greater than the threshold and the degree of sequence identity is less than the threshold. Includes processing to add a selection tag indicating that it is array data.

The selection system according to the fifth embodiment performs the "cleanup processing", "aggregation processing into unique sequences", "integration processing of sequence determination errors", and "classification processing" detailed in the determination system according to the second embodiment. "including. The selection system according to the fifth aspect of the invention may further include a molecular phylogenetic tree creation section and a sequence data determination section.

[Embodiment 6]
A sixth aspect of the invention provides a method of producing antigen-specific antibodies. One embodiment of the sixth aspect of the present invention (hereinafter referred to as "Embodiment 6") will be described below. Although the manufacturing method according to Embodiment 6 is different from the manufacturing method according to Embodiment 3 described above, since there are common parts in specific steps, here, the differences from the specifying method according to Embodiment 3 will be explained. I will mainly explain the points.

The production method according to Embodiment 6 differs from the production method according to Embodiment 3 in that it includes the step of creating a molecular phylogenetic tree and determining whether sequence data indicating an antigen-specific antibody that is a candidate for antibody production is included. The process is not included as an essential process. The production method according to the sixth embodiment includes, as an essential step, "exclusion of sequence data that does not form a sequence cluster", which is detailed in the production method according to the third embodiment. The production method according to Embodiment 6 further includes a step of selecting sequence data indicating an antigen-specific antibody from the sequence data group as an essential step. The features described in the process of selecting sequence data in Embodiment 4 and the features described in this specification are applied as appropriate to the process of selecting sequence data described above.

The features described in the step of preparing a sequence data group according to Embodiment 1 and the features described in this specification are applied as appropriate to the step of preparing a sequence data group in the production method according to Embodiment 6. The features described in the step of producing an antibody according to Embodiment 3 and the features described in the present specification are applied as appropriate to the step of producing the antibody according to Embodiment 6.

Embodiments according to the first or fourth aspect of the invention may be, for example, but not limited to, those described below:
[Item 1] A method for identifying a sequence cluster containing sequence data indicating an antigen-specific antibody that is a candidate for antibody production, the sequence containing sequence data of an antibody obtained from an animal that may have received immune stimulation. preparing a data group; creating a molecular phylogenetic tree from the sequence data group; and determining the degree of sequence identity between the antibody sequence shown in the sequence data in the sequence cluster in the molecular phylogenetic tree and the original sequence of the antibody. determining whether the sequence cluster contains sequence data indicative of an antigen-specific antibody that is a candidate for antibody production based on a change or a change in the type of antibody; A method for specifying a sequence that has the highest degree of sequence identity between the sequence of the antibody obtained and the genome sequence of the animal species corresponding to the animal from which the antibody was obtained.
[Section 2-1] The preparation step includes a step of obtaining a sequence data group, the sequence data group obtained in the preparation step is a sequence data group after sequencing, and the preparation step includes the step of obtaining an amino acid sequence. Item 2. The identification method according to item 1, comprising the step of cleaning up sequence data that does not show.
[Section 2-2] The preparation step includes a step of aggregating a plurality of sequence data indicating the same sequence into one sequence data (hereinafter referred to as "unique sequence data"), and the unique sequence data is the sequence data The identification method according to item 2-1, which includes information regarding the sequence shown in , and information regarding the number of aggregated sequence data (hereinafter referred to as "frequency of appearance (aggregated)").
[Section 2-3] The preparation step includes a step of obtaining a sequence data group, the sequence data group obtained in the preparation step is a sequence data group after sequencing, and the preparation step includes the step of obtaining a sequence data group, and the preparation step Item 1, comprising the step of aggregating a plurality of sequence data indicating sequences into unique sequence data, wherein the unique sequence data includes information regarding the sequence shown in the sequence data and information regarding frequency of occurrence (aggregation). Specific method.
[Item 2-4] The identification method according to Item 2-3, wherein the preparation step includes a step of cleaning up sequence data that does not indicate an amino acid sequence.
[Section 2-5] The preparation step includes a step of acquiring a sequence data group, and the preparation step includes comparing information regarding the quality of sequencing included in the sequence data with a threshold value, The identification method according to any one of Items 1 to 2-4, comprising the step of cleaning up sequence data below.

[Item 3] The sequence data group is a unique sequence data group composed of the unique sequence data, the unique sequence data is nucleotide sequence data, and the sequence data group has the highest frequency of occurrence (aggregation) among the unique sequence data groups. The large unique array data is set as reference array data (n=0), and the length is the same as the array of the reference array data (hereinafter referred to as "reference array"), but there are n pieces (n=1, ... m) unique sequence data having a sequence having a base substitution is used as query sequence data, and the preparation step is performed so that the ratio of the frequency of occurrence (aggregated) of the query sequence to the frequency of occurrence (aggregated) of the reference sequence data is The identification method according to any one of Items 2-2 to 2-5, including the step of integrating a query sequence that is less than a threshold into the reference sequence data.
[Section 4-1] The preparation step classifies the sequence data in the sequence data group based on either or both of the sequence length and the original sequence on the genome sequence of the animal species corresponding to the animal. The identification method according to any one of Items 1 to 3, comprising the step.
[Section 4-2] In the preparation step, the number of sequence data regarding sequences having a predetermined degree of sequence matching with the sequence indicated by specific sequence data in the sequence data group is less than a threshold regarding the number of sequence data. Item 3. The identification method according to any one of Items 1 to 3, including the step of excluding the specific sequence data from the sequence data group.
[Section 4-3] The preparation step classifies the sequence data in the sequence data group based on either or both of the sequence length and the original sequence on the genome sequence of the animal species corresponding to the animal. and when the number of pieces of sequence data regarding a sequence having a predetermined degree of sequence matching with the sequence indicated by the specific sequence data in the sequence data group is less than a threshold regarding the number of pieces of sequence data, the specific sequence Item 3. The identification method according to any one of Items 1 to 3, including the step of excluding data from the sequence data group.

[Item 5-1] The identification method according to Item 4-3, wherein the preparation step includes performing the integration step and the classification step in this order.
[Item 5-2] The identification method according to Item 4-3, wherein the preparation step includes performing the integration step and the exclusion step in this order.
[Item 5-3] The identification method according to Item 4-3, wherein the preparation step includes performing the integration step, the classification step, and the exclusion step in this order.
[Section 6] The classification based on the original sequence corresponds to either or both of the V gene fragment and J gene fragment constituting the variable region of the antibody shown in the sequence data in the sequence data group and the animal. Item 4-1, comprising a step of calculating the degree of sequence matching obtained by a homology search with the genome sequence of an animal species, and a step of assigning the sequence data to a class related to a gene fragment showing the highest degree of sequence matching; The identification method described in Section 4-3, Section 5-1, or Section 5-3.
[Item 7-1] According to any one of Items 1 to 6, wherein the change in sequence identity is a change in sequence identity per unit time, and the change in antibody type is an antibody class switch. How to identify.
[Section 7-2] The sequence data group has sequence information obtained by sequencing each antibody of an antibody population obtained i times at different times from animals that may have received immune stimulation. The identification method according to any one of Items 1 to 6 and 7-1, which is a collection.

[Item 10] A method for selecting sequence data indicating antigen-specific antibodies that are candidates for antibody production, the method comprising preparing a sequence data group containing sequence data of antibodies obtained from animals that may have received immune stimulation. and a step of selecting sequence data indicative of an antigen-specific antibody from a sequence data group, wherein the preparation step has a predetermined degree of sequence identity with a sequence indicated by specific sequence data in the sequence data group. If the number of sequence data related to the sequence is less than a threshold value regarding the number of sequence data, the specific sequence data is excluded from the sequence data group, and the selection step includes the step of excluding the specific sequence data from the sequence data group. The degree of sequence identity between the antibody sequence shown in the sequence data and the original sequence of the antibody is compared with a threshold value, and the original sequence is compared with the antibody sequence shown in the sequence data and the animal from which the antibody was obtained. A selection method in which the sequence has the highest sequence identity with the genome sequence of the corresponding animal species.
[Section 13-1] The preparation step includes a step of obtaining a sequence data group, the sequence data group obtained in the preparation step is a sequence data group after sequencing, and the preparation step 11. The selection method according to item 10, comprising the step of cleaning up sequence data that does not show.
[Section 13-2] The preparation step includes a step of aggregating a plurality of sequence data indicating the same sequence into unique sequence data, and the unique sequence data includes information regarding the sequence shown in the sequence data and frequency of occurrence. The selection method described in Section 13-1, including information regarding (aggregation).
[Section 13-3] The preparation step includes the step of obtaining a sequence data group, the sequence data group obtained in the preparation step is a sequence data group after sequencing, and the preparation step includes the step of obtaining a sequence data group, and the preparation step 11. The method according to item 10, comprising the step of aggregating a plurality of sequence data indicating sequences into unique sequence data, wherein the unique sequence data includes information regarding the sequence shown in the sequence data and information regarding frequency of occurrence (aggregation). Selection method.
[Item 13-4] The selection method according to Item 13-3, wherein the preparation step includes a step of cleaning up sequence data that does not indicate an amino acid sequence.
[Item 13-5] The selection method according to Item 13-1 or 13-2, wherein the preparation step includes performing a cleanup step and an exclusion step in this order.
[Item 13-6] The selection method according to Item 13-3 or Item 13-4, wherein the preparation step includes performing an aggregation step into unique sequence data and an exclusion step in this order.
[Item 13-7] The selection method according to Item 13-2 or Item 13-4, wherein the preparation step includes performing a cleanup step, an aggregation step into unique sequence data, and an exclusion step in this order.
[Section 13-8] The preparation step includes a step of acquiring a sequence data group, and the preparation step includes comparing information regarding the quality of sequencing included in the sequence data with a threshold value, 10. The selection method according to any one of Items 10 and 13-1 to 13-7, which comprises the step of cleaning up sequence data of less than or equal to.

[Item 14] The sequence data group is a unique sequence data group composed of the unique sequence data, the unique sequence data is nucleotide sequence data, and the sequence data group has the highest frequency of appearance (aggregation) among the unique sequence data groups. The large unique array data is set as reference array data (n=0), and the length is the same as the array of the reference array data (hereinafter referred to as "reference array"), but there are n pieces (n=1, ... m) unique sequence data having a sequence having a base substitution is used as query sequence data, and the preparation step is performed so that the ratio of the frequency of occurrence (aggregated) of the query sequence to the frequency of occurrence (aggregated) of the reference sequence data is The selection method according to any one of Items 13-2 to 13-7, including the step of integrating inquiry sequences that are less than a threshold into the reference sequence data.
[Section 15] The preparation step includes a step of classifying the sequence data in the sequence data group based on either or both of the sequence length and the original sequence on the genome sequence of the animal species corresponding to the animal. The selection method according to any one of Items 10 and 13 to 14, comprising:
[Section 16-1] The selection method according to Item 14, wherein the preparation step includes performing the integration step and the exclusion step in this order.
[Section 16-2] The selection method according to Item 15, wherein the preparation step includes performing the integration step, the classification step, and the exclusion step in this order.

[Section 17] The classification based on the original sequence corresponds to either or both of the V gene fragment and J gene fragment constituting the variable region of the antibody shown in the sequence data in the sequence data group and the animal. Item 15 or Item 16, comprising the step of calculating the degree of sequence identity obtained by a homology search with the genome sequence of an animal species, and the step of assigning the sequence data to a class regarding the gene fragment showing the highest degree of sequence identity. -2. Selection method described in 2.
[Item 18] Any one of Items 10 and 13 to 17, wherein the change in sequence identity is a change in sequence identity per unit time, and the change in antibody type is an antibody class switch. Selection method described.
[Section 19] The sequence data group is a collection of sequence data having sequence information obtained by sequencing each antibody of an antibody population obtained i times at different times from animals that may have received immune stimulation. The selection method according to any one of Items 10 and 13 to 18.

[Section 20] A step of creating a molecular phylogenetic tree from the sequence data group, and a change in the degree of sequence identity between the sequence of the antibody shown in the sequence data in the sequence cluster in the molecular phylogenetic tree and the original sequence of the antibody, or The method further includes a step of determining whether or not the sequence cluster includes sequence data indicating an antigen-specific antibody that is a candidate for antibody production, based on a change in the type of antibody, and the selection step The selection according to any one of Items 10 and 13 to 19, which is replaced with the step of selecting sequence data indicating an antigen-specific antibody from a sequence cluster determined to include sequence data indicating a candidate antigen-specific antibody. Method.
[Item 21] The selection method according to Item 20, wherein the preparation step, the creation step of a molecular phylogenetic tree, the determination step, and the selection step are performed in this order.

The features mentioned above for the embodiments according to the first aspect and the fourth aspect correspond to the embodiments according to the second aspect, third aspect, fifth aspect and sixth aspect of the present invention. It can be applied to processes and treatments as appropriate.

Hereinafter, specific examples will be described, but these examples indicate preferred embodiments of the present invention, and are not intended to limit the invention described in the appended claims in any way.

[Example 1]
The purpose of this example is to efficiently select sequence data indicating antigen-specific antibodies from a sequence data group obtained by sequencing an antibody population derived from an animal to which an antigen has been administered.

(1) Antigen administration and blood collection protocol 2 mg/dose of human IgG antibody Fab was administered as an antigen together with an adjuvant to one alpaca. Antigen was administered six times at two-week intervals. Blood (50 mL/time) was collected from the alpaca a total of 15 times, once before antigen administration and 14 times every week from the start of antigen administration. A summary of the antigen administration and blood collection protocols is provided below.

(2) Sequencing of antibody genes Human lymphocyte separation solution (1.077 g/mL) and EDTA (final concentration 0.1%) were added to the blood samples collected at each time, and Leukosep (centrifuge tube) was added. Peripheral blood mononuclear cells (PBMCs) were separated by centrifugation using a PBMC. The PBMCs of the separated blood samples at each time period were washed with PBS, and then the mRNA of each blood sample was extracted using RNAisoPlus (Takara Bio Inc.).

cDNA was synthesized from the extracted mRNA using SuperScript (trademark) III First-Strand Synthesis System (Invitrogen). DNA fragments (sequence lengths of about 300 bp to about 500 bp) encoding the variable regions of alpaca IgG2 and IgG3 antibodies were obtained by PCR using the synthesized cDNA and the primer set shown in the table below.

A DNA fragment library for sequencing was prepared by adding an adapter sequence for Miseq and a barcode sequence to the DNA fragments using Nextera XT Index kit (Illumina). The base sequence of the library was determined by the paired-end method using Illumina's next-generation sequencer (Miseq).

(3) Sequence data group
(A) Sequence data group after sequencing Sequence data for each antibody in the antibody population in blood samples collected at each period was collected to form a sequence data group after sequencing. Antibody sequence data includes information on the base sequence encoding the variable region portion of the antibody, information on the type of antibody (IgG2 or IgG3), information on the timing of blood collection, information on the animal from which the blood was collected, and quality score (Q score). ).
The number (A) of sequence data regarding antibodies (IgG2 and IgG3) in blood samples collected at each time in the sequence data group after sequencing is shown in Table a below.

In this example, the number of sequence data in the sequence data group to be analyzed was approximately 4.95 million (Table a).
The purpose of this example is to efficiently select sequence data indicating antigen-specific antibodies from the sequence data group. For this purpose, in this example, a molecular phylogenetic tree is created for the sequence data in the sequence data group. Generally, when creating a molecular phylogenetic tree, the amount of calculation increases exponentially as the number of data to be analyzed increases. For this reason, when a molecular phylogenetic tree is created for an excessive amount of data, the calculation requires an enormous amount of time, which is not practical. Alternatively, depending on the performance of the computer, it may not be possible to create an appropriate molecular phylogenetic tree. In order to achieve the purpose of this example, the sequence data group was subjected to the processes (B) to (F) described below to reduce the number of data, a molecular phylogenetic tree was created, and the created molecules We decided to select antigen-specific antibodies from the phylogenetic tree.

(B) Aggregation into unique sequence data Among the above sequence data groups, sequence data showing the same base sequence was aggregated into one sequence data (="unique sequence data"). Unique sequence data was generated for the entire sequence data group to generate a unique sequence data group. Through this unique sequence data aggregation process, for example, regarding the IgG2 antibody from the blood sample (first time), the number of sequence data in the sequence library decreased from about 210,000 pieces (A) to about 90,000 pieces (B). (Table a). Therefore, the amount of calculation for creating a molecular phylogenetic tree can be reduced by aggregating data into unique sequence data.
In addition to the above-mentioned information included in the sequence data, the unique sequence data further includes information regarding the number of aggregated base sequence data (="frequency of appearance (aggregated)").

(C) Exclusion of sequence data, etc. that does not indicate the amino acid sequence (cleanup)
If the sequence length of the unique sequence of the unique sequence data is not a multiple of 3, it means that the unique sequence data does not have information regarding the amino acid that becomes the antibody. Unique sequence data that does not indicate an amino acid sequence was removed (cleaned up) from the unique sequence data group. This cleanup process was performed on all sequence data in the unique sequence data group to generate a clean-up sequence data group composed of unique sequence data whose sequence length was a multiple of 3.
Through this cleanup process, for example, regarding the IgG2 antibody of the blood sample (first time), the number of sequence data in the sequence library of unique sequence data will go from about 90,000 pieces (B) to about 69,000 pieces (C). (Table a). Therefore, the amount of calculation for creating a molecular phylogenetic tree can be reduced by the cleanup process.

(D) Integration of Sequencing Errors It is known that sequence information obtained by sequencing can contain random errors that occur during the processes associated with sequencing. Sequence data with random errors may have sequence information in which a single base substitution, a double base substitution, . . . n base substitutions have been introduced with respect to the original sequence. Sequence data having such base substitutions is randomly generated, so the appearance rate of the sequence data follows a Poisson distribution. If the appearance rate of a certain unique sequence data and the appearance rate of sequence data showing a base sequence that differs from that unique sequence by 1, 2, ... n bases follow a Poisson distribution, then a group of sequence data Since it is considered to be a collection of sequence data representing sequences with random errors in the sequence, the group of sequence data is integrated into unique sequence data representing the original sequence.

In this example, the integration of sequencing errors was performed as follows:
(a) Generation of sequence data group for integration regarding reference sequence α
Among the cleanup sequence data groups, the unique sequence data with the highest frequency of occurrence (aggregation) is set as the standard sequence data α, and the sequence length is the same as the base sequence shown in the standard sequence data α (hereinafter also referred to as “standard sequence α”). However, unique sequence data that shows a sequence that differs by one base from the standard sequence α, sequence data that shows a sequence that differs by two bases, ... sequence data that shows a sequence that differs by m bases (hereinafter referred to as "query sequence data (n=m )") were collected to generate a sequence data group for integration regarding the reference sequence α.

In the integrated sequence data group related to the reference sequence α, the query sequence data (n = 1) that shows a base sequence that differs by one base from the reference sequence α has one base substitution in common, but the location of the base substitution is different. Unique sequence data exists for various base sequences. In the integrated sequence data group for the query sequence α, m in the query sequence data (n = m) is determined from the appearance frequency (aggregate) of the reference sequence data α and the expected value of base substitutions to be 1 for the first time. The values were set as below.

(b) Calculation of probability density In the integrated sequence data group regarding the standard sequence α, the appearance frequency (aggregation) of the standard sequence data α and the appearance frequency (aggregation) of each query sequence data (n=1, 2...m) The probability density of the standard sequence data α was calculated by dividing the frequency of appearance (aggregation) of the standard sequence data α by the sum of . Similarly, the probability density of each query sequence data was calculated by dividing the frequency of appearance (aggregation) of each query sequence data (n=1, 2, . . . m) by the total sum. Here, the query sequence data (n=1) includes various unique sequence data that differ by one base from the reference sequence α. Therefore, the probability density of the query sequence data (n=1) is the sum of the probability densities of the individual unique sequence data that differ by one base from the reference sequence α.

(c-1) When the histogram related to the reference sequence follows a Poisson distribution For the integration sequence data group related to the reference sequence α, a histogram related to the reference sequence α was created, arranged in order of the number of base substitutions with respect to the reference sequence α (Figure 1a) . In the histogram related to the reference sequence α, if the distribution of the probability density of the reference sequence data α and the probability density of the query sequence data (n = 1 to 5) follows a Poisson distribution (Fig. 1b), in the integration sequence data group regarding the reference sequence α. The query sequence data (n=1 to 5) were integrated with the reference sequence data α (n=0)) to generate combined unique sequence data regarding the reference sequence α (Fig. 1c).

(c-2) When the histogram related to the reference sequence does not follow a Poisson distribution For the integration sequence data group related to the reference sequence β, a histogram related to the reference sequence β was created, arranged in order of the number of base substitutions with respect to the reference sequence β (Fig. 1d). In the histogram regarding the reference sequence β, the distribution of the probability density of the reference sequence data β and the probability density of the query sequence data (n=1 to 5) did not follow a Poisson distribution, so the following processing was performed.

Occurrence rate of specific unique sequence data included in specific query sequence data (for example, n = 2) (= [frequency of occurrence (aggregated) of that unique sequence] / [frequency of occurrence (aggregated) of reference sequence data β and each query If the total appearance frequency (aggregation) of sequence data (n=1, 2...m)]) is judged to be an outlier from the perspective of probability, the unique sequence data is different from the standard sequence β. Estimated to be a unique sequence. Based on this estimation, the unique sequence data indicating the other unique sequence (hereinafter referred to as "reference sequence data γ") was excluded from the integration sequence data group regarding reference sequence β and returned to the original cleanup sequence data group. .

In addition, in the integrated sequence data group for the standard sequence β excluding the standard sequence data γ, the standard sequence that has the same base substitution as the base sequence of the standard sequence data γ (the same type of base substitution at the same base number) Sequence data showing a base sequence different from γ by one base, two bases, etc. was excluded from the integration sequence data group regarding the reference sequence β and returned to the original cleanup sequence data group. This process is performed because sequence data having the same base substitution as the base sequence of the standard sequence data γ is considered to be sequence data indicating a sequence in which a random error has occurred in the base sequence of the standard sequence data γ.

If there is unique sequence data that exhibits a probability density that is judged to be an outlier from a probability perspective in the integration sequence data group related to the standard sequence β from which sequence data related to the sequence of the standard sequence data γ has been excluded, the above exclusion process is performed. was repeated.
In the integrated sequence data group regarding the standard sequence β after excluding the sequence data indicating a unique sequence different from the standard sequence β, the probability density of the standard sequence data β and its query sequence data (n=1, 2...m ), the query sequence data (n=1, 2...m) were integrated into the reference sequence data β if the distribution of the probability density of ) followed a Poisson distribution (Fig. 1e).

Sequence data used for integrating sequencing errors is excluded from the cleanup sequence data group, and the above-described sequencing error integration process is repeated for the excluded cleanup sequence data group. While repeating the sequencing error integration process, the sequencing errors were also integrated for the reference sequence data δ (FIG. 1f).

(d) Generation of sequence data group after integration of sequencing errors In the sequence data group excluding the sequence data used for integration of sequencing errors, the integration process of the above sequencing errors is performed to This process was repeated until the total (aggregation) became 1. When the highest frequency of occurrence (aggregation) was 1, the integrated unique sequence data generated above was collected to generate an integrated unique sequence data group.

Through this sequencing error integration process, the number of sequence data in a sequence data group increases from approximately 69,000 (C) to approximately 0,200,000 (D) for IgG2 antibodies in a blood sample (first time). ) (Table a). Therefore, by creating a molecular phylogenetic tree for a sequence data group containing a reduced number of sequence data due to the sequence determination error integration process, the amount of calculation can be reduced.

In addition to the above information included in the unique sequence data, the unique sequence data after integration further includes information (frequency of appearance (integration)) regarding the number of sequence data integrated into the unique sequence data.

(E) Generation of classification sequence data groups by classification Antibodies undergo a process of V(D)J recombination, then an affinity maturation process, known to mature.
In the process of V(D)J recombination, antibodies are produced by combining one each of multiple types of V gene fragments, multiple types of D gene fragments, and multiple types of J gene fragments that exist on the genome sequence. Ru. At this time, deletions or additions of bases may occur randomly at the connection sites. Thus, antibodies of various sequence lengths with various combinations of V(D)J gene fragments can be generated during the process of V(D)J recombination.
During the process of affinity maturation, antibodies undergo a high frequency of point mutations in their variable regions. These point mutations are mainly base substitution mutations, and base deletions or insertions are rare. Thus, various antibodies with the same sequence length but different amino acid sequences can be generated during the process of affinity maturation.

To track the maturation process of specific antibodies over time due to additional administration of antigen, the antibody population in the blood sample can be classified according to the antibodies produced during the process of V(D)J recombination. Next, we considered it rational to track mutations (base substitutions) that occurred in the variable regions of the V gene region and J gene region of antibodies in the above group.

(E-1) Classification based on sequence length Considering the process of V(D)J recombination described above, the sequence data in the sequence data group is classified based on the sequence length (Figure 2a), and each class A sequence length class sequence data group consisting of sequence data indicating the sequence length corresponding to is generated.

As mentioned above, affinity maturation of antibodies mainly results in base substitutions, and usually there is no change in the sequence length. Therefore, each class in the sequence length class sequence data group is expected to contain a variety of antibodies produced by V(D)J recombination and a series of antibodies that have undergone subsequent affinity maturation. be done. Therefore, by creating a molecular phylogenetic tree for each sequence length class sequence data group obtained through the classification process based on sequence length, we can reduce the amount of calculations and at the same time It becomes possible to efficiently select antibody sequence data.

(E-2) Classification based on original sequence Base substitutions during the affinity maturation process are concentrated in the portions of the antibody V and J gene fragments that correspond to the complementarity determining regions (CDRs). In the V gene fragment and the J gene fragment, the base sequences other than the CDR sequences are almost identical to the base sequences of the V gene fragment and the J gene fragment on the genome sequence. Therefore, from the nucleotide sequences of the V gene fragment and J gene fragment of the antibody, we thought that it would be possible to estimate which combination of V gene fragment and J gene fragment on the genome sequence produced the antibody.

In fact, when the nucleotide sequences of the V gene fragment and J gene fragment of the antibody were searched for sequence identity with the alpaca genome sequence using BLAST, they were found to correspond to the sequences of the V gene fragment and J gene fragment of the antibody. It was possible to estimate the sequences of each of the V gene fragment and J gene fragment on the genome sequence.

The sequence data groups are further divided into classes based on the original sequences of each of the antibody V gene fragment and J gene fragment. For example, a sequence length class sequence data group in which the base sequence of the variable region of an antibody is 354 bp was classified based on the original sequences of each of the V gene fragment and J gene fragment of the antibody (FIG. 2b). Classification based on this original sequence was performed for each sequence length class sequence data group to generate classified sequence data groups.

Each sequence data classified based on the original sequence further includes information regarding its original V gene fragment and J gene fragment in addition to the information contained in the integrated unique sequence data.

As described above, during antibody affinity maturation, base substitutions occur mainly in the CDR sequences of the antibody, and other base sequences are usually not changed. Therefore, each class in the original sequence data group is an efficient collection of specific antibodies produced by V(D)J recombination and a series of antibodies that have undergone subsequent affinity maturation. Be expected. Therefore, by creating a molecular phylogenetic tree for each class obtained through classification processing based on the original sequence, it is possible to reduce the amount of calculation and more efficiently select antigen-specific antibodies from the molecular phylogenetic tree. becomes.

(F) Generation of sequence cluster-forming sequence data groups In the process of affinity maturation, B cells exhibiting antibodies with relatively high affinity for the antigen preferentially receive stimulation from the antigen, and the B cells are selectively stimulated by the antigen. This will lead to an increase in numbers. During the proliferation process of B cells, mutations are introduced into the variable region of the antibody at a high mutation rate, and antibodies exhibiting higher affinity can be produced. Considering this process of affinity maturation, the nucleotide sequence of the antigen-specific antibody that enables it to preferentially receive stimulation from the antigen until an antigen-specific antibody exhibiting higher affinity is produced. There must have been a series of antibodies similar to this.
Therefore, an antigen-specific antibody that undergoes affinity maturation and exhibits higher affinity has a base sequence similar to that of the antibody and a relatively high affinity when molecular phylogenetic tree analysis is performed. Together with a series of antibodies, it is thought to form a sequence cluster in molecular phylogenetic tree analysis.

On the contrary, B cells exhibiting antibodies with relatively low affinity for the antigen cannot preferentially receive stimulation from the antigen, and the B cells cannot selectively proliferate. As a result, mutations do not accumulate in antibody genes that exhibit antibodies with such relatively low affinity.
In this way, an antibody that has not undergone affinity maturation and exhibits a relatively low affinity can be found in molecular phylogenetic tree analysis because there are few or almost no antibodies with a base sequence similar to that of the antibody. It is thought that no sequence clusters are formed in phylogenetic tree analysis.

Based on the above, in molecular phylogenetic tree analysis, by excluding in advance antibody sequence data that is estimated not to form sequence clusters (hereinafter referred to as "sequence data that does not form sequence clusters") from sequence data groups, molecular phylogenetic trees can be created. This makes it possible to reduce the amount of calculation required.

In this example, since the sequence length of the sequenced antibody base sequence was approximately 300 bp to 500 bp, "40" was set as the number of base substitutions that would result in approximately 90% sequence identity.
With respect to the sequences of specific sequence data in any class within the classified sequence data group based on the sequence length and original sequence described above, the total number of sequence data showing sequences having up to 40 base substitutions is 10 in the class. (hereinafter referred to as "U40<10", etc.), the specific sequence data is evaluated as sequence data related to an antibody that has not undergone affinity maturation (i.e., sequence data that does not form a sequence cluster), and is excluded from the class. did.

By excluding this sequence data that does not form a sequence cluster from all sequence data included in the class, there are at least 10 sequence data showing sequences with up to 40 base substitutions in that class. Sequence data (hereinafter referred to as "sequence cluster-forming sequence data") will remain. Such sequence cluster-forming sequence data is presumed to be sequence data indicative of antigen-specific antibodies that have undergone affinity maturation. (also referred to as "formative sequence data group") will very efficiently contain sequence data indicative of antigen-specific antibodies.

By the above-described process of excluding sequence data that does not form a sequence cluster, the number of sequence data included in that class can be reduced, so the amount of calculation when creating a molecular phylogenetic tree can be reduced. Furthermore, it is possible to select antigen-specific antibodies very efficiently from a molecular phylogenetic tree created for a sequence cluster-forming sequence data group.

(4) Creation of molecular phylogenetic tree A distance matrix was calculated from the sequence data group using Jukes-Cantor 69, and a molecular phylogenetic tree was obtained from the distance matrix using the neighbor-joining method. For these calculations, the ape package of the statistical analysis software "R" was used.

In this example, first, without excluding sequence data that do not form sequence clusters, we created a molecular phylogenetic tree (U40≧0) for a classification sequence data group with a sequence length of 354 bp and an original sequence (vhh3-ighJ-4). was created (Figure 3a). Furthermore, we excluded sequence data that did not form sequence clusters (U40<10, U40<50), and created molecular phylogenetic trees for each group of sequence data that formed sequence clusters (Figure 3b; U40≧10, and Figure 3c; U40≧50). Here, in this example, "U40≧10" and "U40≧50" are the number of sequence data (= "U40") indicating a sequence having up to 40 base substitutions in the base sequence shown in certain sequence data. is 10 or more and 50 or more, respectively.

By pre-excluding sequence data that do not form sequence clusters from sequence data groups, it is possible to not only reduce the amount of calculation for creating a molecular phylogenetic tree, but also to reduce the complexity of the created molecular phylogenetic tree (e.g., branching points (number of nodes, number of branches) decreased. This facilitated the extraction of sequence clusters from the molecular phylogenetic tree described below.

(5) Extraction of sequence clusters from the molecular phylogenetic tree In order to extract sequence clusters containing sequence data representing a series of antibody groups expected to have undergone affinity maturation, the following process was performed.
In the created molecular phylogenetic tree, if the length of a branch extending from a node between branches is greater than or equal to the threshold regarding branch length, the branch extending from that node is further away (outside) from the branch root than the position corresponding to the threshold. was extracted as a sequence cluster (Fig. 4a). Furthermore, in the molecular phylogenetic tree after sequence cluster extraction, if the number of nodes within a predetermined range was greater than or equal to the threshold for the number of nodes, that predetermined region range was also extracted as a sequence cluster (b2 in Figure 4b). Through this sequence cluster extraction, nine sequence clusters (b1 to b9) were extracted from the molecular phylogenetic tree shown in FIG. 4a (FIG. 4b).

(6) Determining whether the sequence cluster contains sequence data indicating an antigen-specific antibody. Determine whether the sequence cluster extracted as above includes sequence data indicating an antigen-specific antibody as follows. did.
During antibody affinity maturation, point mutations occur frequently in the variable region of antibodies, and as a result, antibodies mature into antigen-specific antibodies with high affinity for antigens. Antibody sequences that have undergone affinity maturation accumulate numerous point mutations over time after antigen administration. For this reason, it is expected that the score value (a positive score is given when the sequences match) in a sequence identity or homology search for that nucleotide sequence will become smaller when compared to its original sequence. Ru.

A molecular phylogenetic tree (FIGS. 5a and 5b) was created from the sequence data group as described in (4) and (5) (U40≧10). FIG. 5b is a molecular phylogenetic tree created from the sequence data group included in the sequence cluster b1 (FIG. 4b) of the phylogenetic tree shown in FIG. 4 (FIG. 4a). Regarding the sequence data of antibodies in the sequence clusters (the area surrounded by the middle red circle (circled by a curved line) in Figure 5a and the area surrounded by the middle red circle (circled by a curved line) in Figure 5b) in the created molecular phylogenetic tree, the sequence of each antibody is The degree of sequence identity between the antibody and the original sequence of the antibody was calculated. Specifically, the identity between the sequence data of the V gene region and J gene region of each antibody and the original sequence corresponding to the gene region estimated by BLAST search was evaluated using a bit score (score value).

In order to show the change in bit score according to the time of blood collection for the sequence data of antibodies in each sequence cluster surrounded by red circles (circled with curves) in Figures 5a and 5b, the horizontal axis is the timing of blood collection after antigen administration (weeks). ), and a scatter diagram was created with the vertical axis representing the bit score (FIGS. 5c and 5d). In FIGS. 5c and 5d, circle plots (◯) indicate sequence data of IgG2, and triangular plots (△) indicate sequence data of IgG3. The size of each plot reflects the number (frequency of appearance (total)) of each unique sequence data, and the larger the plot size, the higher the frequency of appearance (total). The color (shading) of each plot indicates when the antibody sequence first appeared (insets in Figures 5c and d).

(Slope of sequence matching)
Figure 5c shows that there is almost no change between the bit score calculated from the original sequence of the antibody before antigen administration (first time) and the bit score calculated from the antibody sequence data from the 4th to 15th blood sampling period. shows. This means that with the protocol of this example, which includes six antigen administrations, mutations are not accumulated in the antibodies shown in the sequence data in this sequence cluster (red circle (area circled by a curve) in Figure 5a). ie, the sequence cluster does not contain sequence data indicative of antibodies specific to the administered antigen. Therefore, this sequence cluster is determined not to be a sequence cluster containing antigen-specific antibodies.

Figure 5d shows that compared to the bit score calculated from the original sequence of the antibody before antigen administration (first time), the bit score calculated from the antibody sequence data from the 4th to 15th blood collection period decreases at a certain rate. did. This is due to the sequential accumulation of mutations in the antibodies shown in the sequence data in this sequence cluster (red circle (area circled by a curve) in Figure 5b) according to the protocol of this example, which includes six antigen administrations. This suggests that the sequence cluster contains sequence data indicative of antigen-specific antibodies. Therefore, this sequence cluster is determined to be a sequence cluster containing antigen-specific antibodies.

(Diversification of antibody types over time)
The color (shading) of the plot shown in Figure 5d changes from yellow (light color "3") to light blue (relatively dark color "9") to red as the number of weeks after antigen administration increases. (Dark color "15") (Fig. 5d inset). This change in color (shade) indicates that the number of antibody types has increased through additional administration of antigen. On the contrary, the color (shade) of the plot shown in FIG. 5c remained green (light color "4-7"). The fact that this color (shade) does not change indicates that there was no change in the type of antibody through additional administration of the antigen.
These results indicate that if the number of types of sequence data in a sequence cluster exceeds a threshold at a certain point after antigen administration or increases over time, that sequence cluster is a sequence cluster containing antigen-specific antibodies. suggests that. Therefore, such a sequence cluster is determined to be a sequence cluster containing antigen-specific antibodies.

(class switch)
In FIG. 5d, plots of circles and triangles are included, and their appearance ratio changes significantly over time. This indicates that this sequence cluster contains IgG2 antibodies and IgG3 antibodies, and that the antibody classes have been switched between IgG2 and IgG3 (that is, antibody class switching has occurred). . On the contrary, in Fig. 5c, the appearance ratio of round and triangular plots was almost constant. This indicates that almost no class switching has occurred within this sequence cluster.
These results suggest that when the antibody class shown in the sequence data within a sequence cluster changes over time, the sequence cluster is a sequence cluster containing antigen-specific antibodies. Therefore, such a sequence cluster is determined to be a sequence cluster containing antigen-specific antibodies.

(Sequence identity before antigen administration)
The bit score before challenge in Figure 5d was approximately 430. On the other hand, the bit score before challenge in Figure 5c was approximately 350. In this example, the bit score indicates a larger numerical value as the degree of sequence matching with the original sequence is higher.
The reason for the small bit score before antigen administration is that the antibody has already undergone affinity maturation for an antigen different from the antigen administered in this example. Antibodies that have undergone such affinity maturation do not biologically normally undergo affinity maturation for the antigen administered in this example. In fact, it was concluded that the sequence cluster in Figure 5c containing antibodies with a small ratio of pre-antigen bit score to its maximum value was not an antigen-specific antibody against the antigen administered in this example, as described above. Therefore, a sequence cluster containing an antibody whose bit score before antigen administration has a small ratio to its maximum value is determined not to contain an antigen-specific antibody.

(7) Measurement of binding strength of antibodies reconstructed from sequence data in sequence clusters Bases of sequence data in sequence clusters determined to contain sequence data of antigen-specific antibodies as described in (6) above Based on the sequences, antibodies containing the corresponding variable regions were reconstructed.
Specifically, based on the information on the base sequence encoding the variable region of an antigen-specific antibody shown in the sequence data, synthesize a DNA having a base sequence with optimized base codons for expression in E. coli, The synthesized DNA was inserted into an E. coli expression vector (pAED4 or pCold3) to construct a VHH antibody expression plasmid. E. coli (BL21(DE3)pLysS) transformed using the plasmid is cultured under predetermined conditions to produce a VHH antibody having the corresponding variable region from the DNA, and the VHH antibody is extracted from E. coli according to a known method. was recovered.

The binding strength (dissociation constant) of the collected antibody to the administered antigen (Fab of human IgG antibody) was measured using Biacore 2000 (GE Healthcare Japan). In the binding force measurement, the collected antibodies (concentrations: 150 nM, 100 nM, 75 nM, 50 nM, 25 nM, 12.5 nM) were measured in a manner that they were brought into contact with the antigen Fab immobilized on a CM5 sensor chip (Figure 5f).

As a result, antibodies having variable regions shown in the sequence data in sequence clusters determined to include sequence data of antigen-specific antibodies bind to antigen Fab with a dissociation constant of 0.5 nM. It has been shown. From the results of Example 1, the sequence cluster determined to contain sequence data indicating an antigen-specific antibody actually contains base sequence data of an antibody (antigen-specific antibody) that binds to the antigen with relatively strong binding force. It was shown that

Similarly, in (6) above, based on the base sequence of the sequence data in the sequence cluster determined not to contain antigen-specific antibody sequence data, an antibody containing the corresponding variable region is reconstructed, and the antibody containing the corresponding variable region is reconstructed. The binding strength of the antibody to the antigen was measured (Fig. 5e). As a result, it was not possible to measure the binding strength to the antigen, and it was shown that the data was not actually the base sequence data of an antigen-specific antibody.

[Example 2]
We measured the binding strength of antibodies reconstructed from various sequence data in sequence clusters determined to contain antigen-specific antibody sequence data.
A molecular phylogenetic tree was created in the same manner as in Example 1 for a classification sequence data group (sequence length: 381 bp, V gene fragment: VHH3-S8, and J gene fragment: ighJ-6) different from that in Figure 5b ( Figure 6a). Regarding the antibody sequence data in the area surrounded by a red circle (circled by a curve) in Fig. 6a, the horizontal axis is the timing (weeks) of blood collection after antigen administration, and the vertical axis is the bit score, in the same manner as described above. A scatter plot was created (Figure 6b). From the slope of the plot and the change in color (shading), this sequence cluster was determined to be a sequence cluster containing an antigen-specific antibody.

Next, a molecular phylogenetic tree was created using the sequence data of the antibodies contained in the sequence cluster shown in the red circle (circled by a curved line) in FIG. 6a (FIG. 6c). Using substantially the same method as described in Example 1, each antibody was reconstructed from the sequence data of (c1, c2, c3) indicated by the dashed arrow in FIG. 6c. The binding strength of the reconstituted antibody was measured in the same manner as in Example 1.

The antibody reconstructed from the sequence data indicated by c1 in FIG. 6c had a binding constant to the antigen of 370 nM, and was characterized by rapid dissociation from the antigen or a high dissociation rate constant [nM/s]. The two antibodies reconstructed from the sequence data shown as c2 and c3 in Figure 6c have antigen binding constants of 68 nM and 90 nM, respectively, which are both relatively high compared to the antibody c1 in Figure 6c. It bound to the administered antigen by avidity. Dissociation from antigen was also relatively slow.

The sequence data indicated by c1 in FIG. 6c was obtained at the timing of blood collection in the 6th week. The sequence data for c2 and c3 in FIG. 6c, which showed stronger binding strength than c1 in FIG. 6c, was obtained at the timing of blood sampling at the 13th week.

Example 2 showed that all sequence data in sequence clusters determined to contain sequence data indicative of antigen-specific antibodies contained nucleotide sequences encoding variable regions of antibodies with relatively high avidity. . Furthermore, Example 2 showed that sequence data having nucleotide sequence information encoding variable regions of antibodies having various characteristics also existed within the sequence cluster.

Example 2 shows sequence data of antibodies for which the antigen has been administered more times and more weeks have elapsed since the antigen administration, among the sequence data in the sequence cluster determined to include sequence data indicative of antigen-specific antibodies. showed the possibility that it is an antigen-specific antibody with higher binding strength.

Claims (11)

In the control section of the computer,
preparing a sequence data group containing sequence data of antibodies obtained from animals that may have received immune stimulation;
creating a molecular phylogenetic tree from the sequence data group; and a change in the degree of sequence identity between the antibody sequence shown in the sequence data in the sequence cluster in the molecular phylogenetic tree and the original sequence of the antibody, or a change in the type of antibody. determining whether the sequence cluster contains sequence data indicating an antigen-specific antibody that is a candidate for antibody production, based on;
A program for executing a method for identifying a sequence cluster containing sequence data indicating an antigen-specific antibody that is a candidate for antibody production, the program comprising:
The program, wherein the original sequence has the highest sequence identity between the antibody sequence shown in the sequence data and the genome sequence of the animal species corresponding to the animal from which the antibody was obtained. The sequence data group obtained in the preparation step is a sequence data group after sequencing,
The preparation step includes a step of aggregating a plurality of sequence data indicating the same sequence into one sequence data (hereinafter referred to as "unique sequence data"),
2. The program according to claim 1, wherein the unique sequence data includes information regarding the sequence shown in the sequence data and information regarding the number of aggregated sequence data (hereinafter referred to as "frequency of appearance (aggregated)"). The sequence data group is a unique sequence data group composed of the unique sequence data,
The unique sequence data is base sequence data,
Among the unique sequence data groups, the unique sequence data with the highest frequency of appearance (aggregation) is set as standard sequence data (n=0), and the length is the same as the sequence of the standard sequence data (hereinafter referred to as "standard sequence"). , unique sequence data having a sequence having n base substitutions (n=1, . . . m) with respect to the reference sequence is set as query sequence data,
The preparatory step is a step of integrating query sequences in which the ratio of the appearance frequency (aggregation) of the query sequence to the appearance frequency (aggregation) of the reference sequence data is less than a threshold value into the reference sequence data (hereinafter referred to as the "integration step"). The program according to claim 2, comprising: The preparation step is
Classifying the sequence data in the sequence data group based on either or both of the sequence length and the original sequence on the genome sequence of the animal species corresponding to the animal (hereinafter referred to as the "classification step"); /or When the number of pieces of sequence data regarding a sequence having a predetermined degree of sequence identity with the sequence indicated by the specific sequence data in the sequence data group is less than a threshold regarding the number of pieces of sequence data, the specific sequence data is a step of excluding from the sequence data group (hereinafter referred to as "exclusion step");
The program according to any one of claims 1 to 3, comprising either one or both of the following. The program according to claim 4, which refers to claim 3,
The program, wherein the preparation step includes the integration step, the classification step, and the exclusion step, and includes performing the integration step, the classification step, and the exclusion step in this order. The classification based on the original sequence is
Obtained by homology search between either or both of the V gene fragment and J gene fragment constituting the variable region of the antibody shown in the sequence data in the sequence data group and the genome sequence of the animal species corresponding to the animal. 6. The program according to claim 4, comprising the steps of: calculating the degree of sequence identity based on the sequence identity; and assigning the sequence data to a class related to a gene fragment showing the highest degree of sequence identity. 7. The program according to claim 1, wherein the change in sequence identity is a change in sequence identity per unit time, and the change in antibody type is an antibody class switch. causing the control unit to execute a step (hereinafter referred to as "selection step") of selecting sequence data indicating an antigen-specific antibody that is a candidate for antibody production from sequence clusters determined to include sequence data indicating an antigen-specific antibody; further including;
The selection step includes a step of comparing the degree of sequence identity between the sequence of the antibody shown in the sequence data in the sequence data group and the original sequence of the antibody with a threshold value,
Any one of claims 1 to 7, wherein the original sequence has the highest degree of sequence identity between the antibody sequence shown in the sequence data and the genome sequence of the animal species corresponding to the animal from which the antibody was obtained. The program described in paragraph 1. A determination system for determining whether a sequence cluster includes sequence data indicating an antigen-specific antibody that is a candidate for antibody production,
Comprising a control unit and a storage unit,
The storage unit stores a database containing sequence data groups containing sequence data of antibodies obtained from animals that may have received immune stimulation,
The program according to any one of claims 1 to 7 is loaded into the control unit, and the control unit executes a method for identifying a sequence cluster containing sequence data indicating an antigen-specific antibody that is a candidate for antibody production. Judgment system. A step in which the program selects sequence data indicative of an antigen-specific antibody that is a candidate for antibody production from sequence clusters determined by the control unit to include sequence data indicative of an antigen-specific antibody (hereinafter referred to as "selection step"). The selecting step includes a step of comparing the degree of sequence identity between the sequence of the antibody shown in the sequence data in the sequence data group and the original sequence of the antibody with a threshold value,
The determination system according to claim 9, wherein the original sequence is a sequence that has the highest degree of sequence identity between the sequence of the antibody shown in the sequence data and the genome sequence of the animal species corresponding to the animal from which the antibody was obtained. . A method for producing an antigen-specific antibody, the method comprising:
A production method comprising the steps of selecting sequence data using the determination system according to claim 10 and producing an antibody having the amino acid sequence shown in the sequence data.