JP2009219366A - Method for judging haplotype - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method for efficiently improving a fixing rate of diplotype by selecting an SNP pair for carrying out haplotyping, and combining the SNP pair with SNP typing. <P>SOLUTION: The method for judging haplotype includes carrying out the haplotyping of a part of a plurality of the SNPs constituting the haplotype of the subject gene, carrying out the SNP typing of a part or the whole of the remainder, and carrying out the judgement of the haplotype of the subject gene from the result of both typings. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  This is related to the field of detecting gene polymorphism and searching for related genes using it. In addition, regarding polymorphic markers that have been proved to be relevant, they are also used in clinical predictions of phenotypes such as side effects.

  Attempts to screen polymorphisms that serve as markers of disease susceptibility and side effects by associating gene polymorphisms with phenotypes have been widely carried out with recent advances in polymorphism detection techniques. In particular, SNP (Single Nucleotide Polymorphism), which exists in the human genome in a large amount of 3 to 10 million, and is relatively easy to type, is an essential polymorphism for genome-wide association analysis. .

  In recent years, multiple human genome sequences have been obtained, and their polymorphisms have been examined in detail. In particular, HAPMAP, published in 2005, provides linkage disequilibrium maps in the genome, greatly affecting association analysis using SNPs.

  Since the human genome is diploid, the result of SNP typing gives two alleles. For example, a wild type of a certain SNP is A, and a mutant type is G (hereinafter referred to as A> G). Here, the wild type is an allele with a high frequency in the population, and the mutant type is an allele with a frequency of 1% or more, usually called a polymorphism. At this time, there are three types of SNP typing results: AA, AG, and GG. AA and GG are called homo and AG is called hetero. The result obtained by typing SNP in this way is called genotype.

  Think about multiple SNPs. For example, assume that two SNPs are SNP1 and SNP2, and A> G and C> T, respectively. It is assumed that the result of SNP typing at two locations (genotype) is that SNP1 is AG heterogeneous and SNP2 is CT heterozygous. When two SNPs are on the same chromosome, there can be A-C / G-T and A-T / G-C depending on what allele is physically linked by SNP1 and SNP2. A combination of SNPs on a single chromosome that is physically linked in this way is called a haplotype. Here, A-C / G-T has A-C and G-T haplotypes, and A-T / G-C has A-T and G-C haplotypes. A pair of two haplotypes is called a diplotype. Here, the diplotype is A-C / G-T, or the diplotype is A-T / G-C.

  The diplotype is complete information. If the diplotype is known, the genotype can be known, but the diplotype may not be known from the genotype. For example, in the case of the above SNP1: AG hetero, SNP2: CT heterogeneity, it cannot be determined whether the diplotype is A-C / G-T or A-T / G-C. However, the genotype is obtained by many SNP typing methods known in the world, and therefore information on the diplotype or its constituent haplotypes may not be obtained.

  In correlation analysis that associates gene information with phenotypes, it is desirable to know the diplotype, which is complete information. However, as shown above, the diplotype may not be known from the SNP typing result. A haplotype estimation algorithm that is generally used at this time is a haplotype frequency that is statistically estimated from the genotype results of a plurality of persons (Non-Patent Document 1).

  Usually, a person is analyzed as having the most probable diplotype as a result of estimation, but in this case, the first type of error (false negative) in the analysis result may become large. This is not a very appropriate method.

  As a countermeasure to such a problem, instead of selecting one diplotype with the highest probability of estimation as described above, all possible diplotypes are added with a weight of probability. There has been proposed a method of performing association analysis simultaneously with estimation (Non-patent Document 2, Patent Document 1). Specifically, an algorithm for simultaneously estimating the penetration rate based on the diplotype in addition to the haplotype frequency and the diplotype form is provided. Even if the diplotype type of each individual is not determined by this method, the maximum likelihood estimation of the haplotype frequency of the population, the diplotype distribution and the penetrance of each individual is given given the genotype and phenotype of the population. It became possible to do.

  Using the above method, even if the diplotype of each individual cannot be determined, correlation analysis within the population can be performed, and polymorphic markers (haplotypes) suspected to be related to disease or drug response are searched. can do. However, when the actually obtained polymorphic marker is applied in clinical practice, it is necessary to be able to determine the diplotype of each individual.

  For example, when it is known that a haplotype is associated with side effects, the genotyping results indicate that both diplotypes with and without the haplotype are possible, for example, with side effects. Even if the probability of a diplotype including a haplotype is low, it is dangerous to disregard and administer it. On the other hand, if there is a risk, if no medication is given, there will be no side effects, and the treatment will be taken away from those who are expected to be effective.

  However, when different diplotypes give equal genotypes, the method of statistical analysis based on the commonly used SNP detection results can only obtain a posterior probability distribution of the diplotypes that can be realized. The type cannot be fixed to one.

  When such different diplotypes give equal genotypes, several methods for determining the diplotype are known. The most common method is using pedigree information, where the diplotype of the child may be determined from the parents' SNP typing results. However, recombination may occur between the two SNPs, and the diplotype of the child may not be uniquely determined depending on the genotype of the parents. Also, the parents' genomes are not always available.

  On the other hand, development of so-called haplotyping for directly detecting diplotypes directly from the genome is underway. Haplotyping can be broadly divided into two types. The first type is to obtain information from a haploid by diluting the genome in stages (Non-patent Document 3). The second type intends to detect using the fact that two alleles forming a haplotype are physically linked in an amplification product by PCR (Patent Document 2). However, the former method is still in the development stage, and the latter method requires a larger number of processes than SNP typing. If the haplotypes of multiple SNPs are to be determined only by the latter method, it is necessary to perform haplotyping for all SNP pairs. In the case of n SNPs, the following number of detections are required. Become.

Excoffier L, Slatkin M: Molecular Biology of Evolution Vol12 921-927, 1995 `` Maximum-likelihood estimation of molecular haplotype frequencies in a diploid population '' Shibata K, Ito T, Kitamura Y, Iwasaki N, Tanaka H, Kamatani N: Genetics Vol168 525-539, 2004 `` Simultaneous estimation of haplotype frequencies and quantitative trait parameters: applications to the test of association between phenotype and diplotype configuration '' Ding C and Cantor C: PNAS Vol100 7449-7453,2003 `` Direct molecular haplotyping of long-range genomic DNA with M1-PCR '' JP 2004-354373 A JP 2002-272482 A

  When considering a plurality of SNPs included in a target region, genotypes obtained as a result of SNP typing may be equal for a plurality of different diplotypes. In such a case, the diplotype cannot be uniquely determined only by statistically analyzing the result of SNP typing.

  Various methods are currently under development for so-called haplotyping methods for directly detecting haplotypes, but generally require a larger number of steps than SNP typing. Further, in order to determine a haplotype by a plurality of SNPs only by haplotyping, it is necessary to perform haplotyping between all SNP pairs, and a large number of detections are required.

  Therefore, an object of the present invention is to provide a technique for efficiently improving the diplotype determination rate by selecting an SNP pair for haplotyping and combining it with SNP typing.

  The haplotype determination method of the present invention performs haplotyping on a part of a plurality of SNPs constituting a haplotype of a target gene, performs SNP typing on the remaining part or all of the target gene, and determines the target based on the result of both typings. It is a haplotype determination method characterized by determining a haplotype of a gene.

  According to the present invention, when a plurality of different diplotypes give equal genotypes, the rate of determination is improved with less haplotyping in order to improve the rate of determination of diplotypes by combining SNP typing with haplotyping. Can be made.

  In the haplotype determination method of the present invention, haplotyping is performed on a part of a plurality of SNPs constituting the haplotype of the target gene, SNP typing is performed on the remaining part or all of the target gene, and the target gene is determined from the result of both typings. Determine the haplotype.

  The SNP to be haplotyping is preferably selected from SNP pairs included in a plurality of diplotypes having the same genotype based on frequency information. Moreover, it is preferable that SNP used as the object of SNP typing is selected from SNP other than SNP used as haplotyping object. Furthermore, it is preferable that the SNP to be haplotyping is selected based on information on linkage disequilibrium between SNPs.

Below, based on FIGS. 1-5, embodiment of this invention is described in detail. FIG. 1 is a diagram showing algorithms for selecting pairs for SNP typing and haplotyping, respectively. FIG. 2 is a determination flowchart for determining the duplotype. FIG. 3 shows 10 SNPs as an example. Among them, SNPs with Δ ² = 1 are grouped together into one group, and a haplotype obtained by collecting only representative SNPs of each group is simplified. It is a figure explaining the case of calling. FIG. 4 is a diagram for explaining a case where a linkage disequilibrium coefficient D ′ is calculated for all pairs between SNPs constituting a simplified haplotype and only a pair of D ′ ≠ 1 is selected. FIG. 5 is a schematic diagram of SNPs that perform haplotyping and SNPs that perform SNP typing, which are selected by the present algorithm among 10 SNPs.

  As described above, when different diplotypes give the same genotype, the genotype cannot be uniquely determined only by SNP detection. Therefore, consider combining the haplotyping described above. Here, many simple methods exist for SNP typing, whereas haplotyping methods are complicated, and no simple method has yet been established. In addition, in the method using the feature that they exist on the same chromosome and are physically connected, it is difficult to perform haplotyping when the distance between SNPs is too far away. It is limited to a length that allows PCR amplification with primers. Therefore, when combining haplotyping with SNP typing, it is desirable to be able to determine the diplotype with as few haplotypings as possible. That is, it is necessary to select an SNP pair for which haplotyping is effective.

  As this selection criterion, it is preferable to select an SNP pair for which haplotyping is effective based on information on linkage disequilibrium between SNPs.

  Further, based on the frequency information of the diplotypes configured in each SNP and the linkage disequilibrium information between the SNPs, as will be described later, the improvement rate of the diplotype determination rate for the SNP pair is calculated and determined by haplotyping. It is preferable to select a combination that improves. As frequency information of the diplotype configured by each SNP, for example, known information can be used.

  A method for combining SNP typing and haplotyping, which is effective for uniquely determining a diplotype when different diplotypes give the same genotype, will be described below with specific examples.

  Assume that there are 10 SNPs in the target area. For the haplotypes composed of these SNPs, frequency information is given. For all 10 SNP pairs, consider linkage between SNPs.

  When multiple SNPs are present at close positions on the same chromosome, linkage disequilibrium exists between these SNPs. Linkage disequilibrium is an exception to Mendel's independent law, because when two SNPs are on the same chromosome, the distribution of alleles at different loci is not independent of each other. In other words, when two SNPs on the same chromosome are located at different loci, but one of the alleles is distributed to the child, the other physically linked allele is simultaneously This is because the probability of being distributed to becomes higher.

Various measures of linkage disequilibrium have been proposed, and D 'and Δ ² are particularly frequently used. To illustrate these specific definitions, consider two SNPs, namely SNP1 and SNP2. One SNP1 has alleles a and b, and the other SNP2 has alleles c and d. The allele frequencies are Pa, Pb (= 1-Pa), Pc, and Pd (= 1-Pc), respectively. Next, there are four types of haplotypes that are combinations of two SNPs: ac, ad, bc, and bd, and the frequency of each is Pac, Pad, Pbc, and Pbd. Here, these variables satisfy the following relationship.

  Table 1 shows the relationship between the above variables.

  The linkage disequilibrium coefficient D is defined as follows.

In the absence of linkage disequilibrium, the haplotype frequency is given by the product of the allele frequencies of each SNP, so D = 0. Using this D, D ′ and Δ ² are defined as follows.

D ′ is a standardization of D so that it takes a value between 0 and 1, Δ ² is an X ² statistic representing the independence of a 2 × 2 contingency table,

There is a relationship. Here, n represents the total number of chromosomes.

The SNP selection method using the above-described linkage disequilibrium coefficient (D, Δ ² ) will be described below. However, there are various scales of linkage disequilibrium, and it may be performed based on such information.

Next, consider the 10 SNPs shown in FIG. First, an algorithm for selecting one to be SNP-typed and one to be haplotyped from the 10 SNPs shown in FIG. 3 using the above-described linkage disequilibrium coefficient is shown below. If there is no SNP group having a relationship of Δ ² = 1, the process proceeds to a step of selecting a diplotype having the same genotype from all the next SNPs.

First, consider the linkage disequilibrium coefficient Δ ² . Δ ² is 0 when the two SNPs are independent, and 1 when the SNP is perfectly linked (Perfect LD). Here, a perfect chain is a state in which if the SNP1 allele is known, the SNP2 allele is determined and vice versa. In such a case, since the information given by the typing results of SNP1 and SNP2 is equal, there is no need to type both. Therefore, the SNP groups having the relationship of Δ ² = 1 may be combined into one, and one of them may be typed as a representative. FIG. 3 schematically shows this state.

The method of selecting the representative SNP is arbitrary as long as Δ ² = 1, but for example, the following indicators are worth considering.
1. Suspected direct relationship with phenotype (eg, with amino acid mutation in exon region). 2. Advantageous to the SNP typing method used (large ΔTm, GC content 40% to 60%, no other polymorphism in the vicinity of ~ 20bp from SNP position, etc.).

In FIG. 3, the groups are limited to those adjacent to each other, but in reality, there are cases where SNPs at distant positions have a relationship of Δ ² = 1, so all SNPs existing in the considered region It is necessary to calculate Δ ² between them and perform grouping. From the grouped SNPs, only one representative SNP is considered to constitute a simplified haplotype. FIG. 3 shows that 10 SNPs have decreased to 4 locations. This process corresponds to step S101 in FIG. Hereinafter, the three-digit number indicated by S indicates each step in FIG. If there is no Δ ² = 1, the process proceeds to the following step as it is.

Then, using the simplified haplotypes, find all the diplotypes that combine them. (S102) In principle, 2 ⁴ = 16 types of haplotypes are possible from the ^four SNPs, but the frequency of the given haplotypes determines the number of haplotypes to be examined below. Although haplotypes with a certain frequency or higher will be considered, the threshold value is determined based on the accuracy desired to be finally obtained and the accuracy of the frequency distribution itself.

  Among the diplotypes obtained in this way, combinations that give the same genotype are extracted. The haplotypes from the four SNPs can be listed, for example, as shown in Table 2 (A) below, and the diplotypes obtained therefrom can be listed, for example, as shown in Table 2 (B) below.

  There are 21 diplotypes generated from the six haplotypes, of which there is one combination with the same genotype. It gives the same genotype (A / G, T / C, GG, GG) when the diplotype is d1 = (ATGG, GCGG) and d2 = (ACGG, GTGG).

For those whose genotype and diplotype are determined one-to-one, the total frequency I ₀ is calculated.

Here, A is a set of diplotypes when there are a plurality of diplotypes that give it to one genotype. (D1, d2∈A). I ₀ represents the ratio at which the diplotype can be determined only with the genotype. Here, α represents a coefficient that is 1 when i = j and 2 when i ≠ j.

Next, consider the linkage disequilibrium coefficient D ′. As described above, D ′ is a standardized version of D. However, similarly to Δ ² , D ′ becomes 0 when two SNPs are independent. When Δ ² is 1, D ′ is also 1. However, even when Δ ² is not 1, D ′ may be 1. That is when one of the four haplotype frequencies Pac, Pad, Pbc, and Pbd is zero. This is achieved when there is no continuous mutation at any of the two SNP sites and no recombination between sites has occurred.

  In the case of D ′ = 1, since the haplotype can be determined only from the result of genotyping, there is no additional information obtained by haplotyping. D ′ is calculated for all pairs of SNPs that make up the simplified haplotype, and only pairs with D ′ ≠ 1 are selected. (S103) Here, there are four SNPs, and if D ′ = 1 in 2-3 and 3-4, D ′ ≠ 1, and 1-2, 1-3, 1-4, 2-4 Four SNP pairs are selected (Figure 4).

In the process so far,
(1) diplotypes with equal genotypes (d1, d2∈A), and (2) SNP pairs (1-2, 1-3, 1-4, 2-4 with linkage disequilibrium coefficient D ′ ≠ 1) ∈B)
Is selected. When haplotyping is performed between two SNP pairs between the two groups, it is determined whether one diplotype can be discriminated. If not possible, x is written, and if possible, the frequency of the diplotype is written (S105).

  Here, whether or not the diplotype can be determined will be described in detail (S104). As shown in the table, the genotypes of SNP pair 1-2 are equal to A / G and T / C. However, due to haplotyping of SNP pair 1-2, d1 is (AT, GC) and d2 is (AC, GT ) Therefore, if haplotyping of SNP pair 1-2 is performed, it is possible to determine whether it is d1 or d2. On the other hand, in the information obtained from the haplotyping of the SNP pair 1-3, d1 is (A-G, G-G), d2 is (A-G, G-G), and they are equal. That is, d1 and d2 cannot be determined in order to give the same result for haplotyping of SNP pair 1-3 for both d1 and d2. The genotypes of d1 and d3 are A / G, GG, and homozygous for SNP3. In other words, when one is homozygous, it is not possible to obtain more information than genotype even if haplotyping is performed. Even if both are heterogeneous, it may not be possible to distinguish by haplotyping, so extract the relevant part of diplotypes d1 and d2 as described above (in the case of SNP pair 1-2 (AT, GC) and AC, GT)), it is necessary to check whether or not discrimination is possible.

  If the determination is possible, the ratio (the frequency of the diplotype) determined thereby is obtained as ΔId1 (1,2). When all the combinations are determined to be unsuccessful, the target SNP pair cannot receive the effect of the combination of haplotyping and SNP typing, so the selected SNP group is changed. After all determinations are completed, ΔI in each SNP pair is calculated as follows. (S106)

  ΔI (i, j) indicates how much the diplotype determination rate is increased by haplotyping of the SNP pair (i, j). here,

It is. Therefore, ΔI (1,3) = ΔI (1,4) = ΔI (2,4) = 0, indicating that only haplotyping between SNP pairs 1-2 is effective.

  Next, the efficiency of haplotyping between the SNP pairs 1-2 validated as described above is determined. There are several different approaches to haplotyping, but many take advantage of the fact that physically linked states are preserved by PCR. In this case, the distance between SNPs in the genome or mRNA needs to be a distance that can be amplified by one PCR. Therefore, the criteria for determining the efficiency of haplotyping depend on the efficiency of the processing method in which the physical distance between SNP1 and SNP2 is the same strand in the genome or mRNA after sample processing.

As shown in FIG. 3, the SNPs constituting the SNP pair 1-2 are representative SNPs in which three SNPs and two SNPs are grouped by Δ ² = 1, respectively. As described above, if Δ ² = 1, it is arbitrary to select which SNP from the group, but in addition to this, “efficiency of haplotyping” is also examined. In general, a physical distance is more advantageous for haplotyping, and when the physical distance is several hundred kbp, it is difficult to obtain an amplification product having two SNPs on the same strand. It should be noted that the preferred threshold value is 5 kbp as to whether haplotyping is possible, but this value is a parameter that can be changed depending on the amplification method used.

With respect to SNPs that can be selected within a group, haplotyping is possible regardless of which representative SNP pair is selected as long as the physical distance on the gene is about 500 bp. In this case, the efficiency of haplotyping depends on the typing performance of each SNP. In general, when using the hybridization method to discriminate between wild type and mutant type, the difference in ΔTm between the wild type target binding to the wild type target and the mutant type probe binding to the wild type target is important. It is known to become. Therefore, if there are multiple SNP pairs with a physical distance of 500 kbp or less, calculate ΔTm assuming a 10 bp left and right probe (total 21 bp) centered on the SNP point, and ΔTm is the most Select a pair that grows.
The SNP for haplotyping may be selected by performing the evaluation based on the selection criteria as described above. However, as a specific method, between pairs having the largest ΔI (i ′, j ′) as shown in FIG. First, the possibility of haplotyping is evaluated (S107). If the physical distance of the selected SNP pair exceeds the threshold (currently 5 kbp), the SNP pair that can be haplotyped for the corresponding (i, j) can be selected. Can not.

  In this case, (i, j) is removed from the elements of B (for example, the rows in Table 3) (S108), and the pair (i ', j') with the next largest ΔI (i ', j') is selected. You can repeat the same process.

FIG. 5 shows a schematic diagram of the selection result. From the group constituting the representative SNP 1-2, haplotyping is performed by selecting two SNPs (a, b) at which the distance between the corresponding SNPs is closest. Since there is no SNP grouped by Δ ² = 1 in the representative SNP3 (c), SNP typing is performed as it is, and one SNP4 (d) is selected from the group and SNP typing is performed.

  In this example, since only d1 and d2 have a set of diplotypes with the same genotype, the algorithm ends here. However, if there are many dk, A to d1, d2 and B from (i, j) are excluded (S109, S110), and ΔI is calculated again. The same process is repeated until there is no A element or B element (S106 to S109). When all the elements of A (for example, the column in Table 3) disappear, the selected SNP is determined for haplotyping (S111), and all diplotypes can be determined. If the B element disappears even though the A element remains, there is a diplotype whose phase cannot be determined. This corresponds to the case where the physical distances of all SNP pairs that can be selected for a certain (i, j) are equal to or greater than the threshold and haplotyping cannot be performed for (i, j). Again, this threshold is a parameter that varies depending on the sample processing method used, and in the following, 5 kbp is set as the threshold.

  Next, a configuration is shown in which SNP typing and haplotyping are simultaneously performed on the SNP selected by the above algorithm.

  In order to perform SNP typing using a DNA chip, a DNA chip is prepared by immobilizing on a substrate two types of probes that fully match both wild-type and mutant-type probes that contain SNP sites. Design a primer containing the SNP site of the sample and perform amplification using the PCR method. At the same time, the amplification product is labeled with a fluorescent label (eg, Cy3). By performing hybridization between the amplification product thus generated and the DNA chip, a difference in hybridization intensity between the full match and mismatch can be determined as a difference in signal intensity of the label (FIG. 6).

  For haplotyping, for example, a method using allele-specific PCR and a DNA chip can be used. In this method, allele-specific primers labeled with different dyes (Cy3, Cy5) are set for one of the SNP pairs (a, b) for which haplotyping is desired. Perform PCR with one SNP in the amplification product. A DNA chip in which two types of wild-type and mutant-type probes corresponding to the other SNP are immobilized on a substrate is prepared, and a hybridization reaction is carried out with the product subjected to the allele-specific PCR. As a result of fluorescence detection, the haplotype is identified from the type of dye and the position where hybridization occurs (FIG. 7). In addition, a method of detecting a haplotype by mixing probes (US Pat. No. 6,063,463 B1) can be used.

  Both the SNP typing and haplotyping methods can be performed using a DNA chip. Therefore, the probe designed for SNP typing and the probe designed for haplotyping can be immobilized on the same substrate and simultaneously hybridized on a DNA chip to detect fluorescence (FIG. 8).

  In this way, by selecting polymorphism and combining SNP typing and haplotyping, it is possible to discriminate two diplotypes d1 and d2 that give equal genotypes, which could not be discriminated by the conventional method using only SNP typing. It becomes possible.

  Here, it is important that the determination of d1 and d2 cannot be achieved even if the number of typing is increased (even if the number is 4 to 10) in the method of performing only SNP typing.

  Next, with reference to FIG. 9, a computer system that implements an algorithm for selecting the most efficient polymorphism to enable discrimination of a plurality of diplotypes that give equal genotypes according to the present invention will be described.

  FIG. 9 is a block diagram illustrating a configuration of an information processing apparatus to which haplotype estimation according to the present embodiment is applied. The haplotype estimation method of this embodiment is implemented in a device in which a central processing unit (CPU) 91, a storage device 92, a RAM 93, and an input / output device 94 are connected by a bus 95. That is, it can be mounted on a general personal computer, a workstation or the like.

In FIG. 9, a central processing unit (CPU) 91 temporarily stores the program of the present embodiment stored in the storage device 92, data necessary for program execution of the present embodiment, and the like on the RAM 93, The program of the embodiment is executed. The input / output device 94 includes a display, a keyboard, a pointing device, a printing device, a network interface, and the like, and performs interaction with the user when executing the program of the present embodiment. In many cases, the execution of the program of the present embodiment is triggered by the user via the input / output device 94. Also, the input / output device 94 is used for referring to the execution result of the user and for parameter control during program execution.
Do through.

  FIG. 10 is a flowchart for explaining a program for performing haplotype estimation according to the present embodiment. In each step, a program stored in the storage device 92 shown in FIG. 9 is expanded on the RAM 93 and executed by the central processing unit (CPU) 94. Data input / output and the like are performed through the input / output device 94 as appropriate.

  101 is a step of inputting a haplotype from the input / output device 94. Reference numeral 102 denotes a step of constructing a simplified haplotype using an index such as a linkage disequilibrium coefficient. 203 is a step of determining a haplotyping location from the simplified haplotype constituted by 102 and outputting the result.

Next, examples of the present invention will be described.
Example 1
The usefulness of the technique proposed by the present invention will be described using haplotype data relating to the SAA gene used in Patent Document 1. However, in the above document, the SAA gene haplotype is considered to span two genes, SAA1 and SAA2, but in the present invention, only the haplotypes of the SAA1 gene based on five SNPs are used.

The five SNPs in SAA1 are 1. -61C> G, 2.-13T> C, 3. -2G> A, 4. 2995C> T, 5. 3010C> T, and the haplotype frequency is given as follows: It is done. Here, with respect to the 5SNP, there was no SNP between which Δ ² = 1.

  Next, all diplotypes generated from the 10 types of haplotypes (cumulative haplotype frequency 99.8%) are obtained. There are 55 diplotypes, of which diplotypes that give equal genotypes are listed below.

  In SAA1, since a plurality of diplotypes correspond to 5 types of genotypes, 11 types of diplotypes cannot be distinguished (d1 to d11). For example, d1 and d2 both give a genotype of (CC, T / C, GG, C / T, CC), and d3, d4, d5 are all (C / G, T / C, GG, CC, C / T) gives the genotype.

  As described above, when a plurality of diplotypes cannot be determined in correspondence with one genotype, the diplotype having the higher frequency is generally represented. Then, d1, d3, d6, d8, and d10 are adopted, and the values of d2, d4, d5, d7, d9, and d11 are not reflected. In this case, the percentage of diplotypes that actually exist but are not recognized and discarded is 0.954%, which is one in 100 people.

  Next, a linkage disequilibrium coefficient D ′ is calculated for all pairs between the five SNPs of SAA1. The total number of pairs is 10 (1-2, 1-3, 1-4, 1-5, 2-3, 2-4, 2-5, 3-4, 3-5, 4-5) It can be seen that the calculation result D ′ = 1 is the SNP pairs 1-4, 2-3, and 4-5. Therefore, there are seven pairs (1-2, 1-3, 1-5, 2-4, 2-5, 3-4, 3-5) that do not satisfy D ′ = 1.

  Whether diplotype can be determined when haplotyping is performed between each SNP pair for 11 diplotypes corresponding to the same genotype and 7 SNP pairs with D '≠ 1 A table summarizing these is shown below.

  The table shows that 2-4 haplotyping is most effective. Phases d1, d2, d6, d7, d8, and d9 are determined by performing haplotyping of 2-4.

  Next, a table created by excluding the above 6 confirmed diplotypes (d1, d2, d6, d7, d8, d9) from A and excluding SNP pair 2-4 performing haplotyping from B is shown below. .

  The table shows that 2-5 haplotyping is the next most effective. This establishes the d3 phase.

  Hereinafter, similarly, a table created by excluding A to d3 and excluding B to 2-5 is shown.

  From the table, it can be seen that haplotyping of SNP pairs 1-3, 3-4, and 3-5 has the same effect to give equal ΔI. Here, since the distance between SNP pairs is <500 bp and higher efficiency is obtained than other pairs, 1-3 haplotyping is performed and the phases d10 and d11 are determined. Similarly, even if typing is performed with SNP pairs 1-2 and 1-5, the phases d4 and d5 can be determined, but 1-2 is selected such that the distance between the SNP pairs is <500 bp.

From the above, as a configuration for obtaining the diplotype of SAA1,
`` 2-4 (-13T> C: 2995C> T), 2-5 (-13T> C: 3010C> T), 1-2 (-61C> T: -13T> C), 1-3 (-61C > T: -2G> A) Haplotyping "
All phases can be confirmed by performing. Regarding the five SNPs of SAA1 examined this time, there was no SNP that required only SNP typing.

  Next, a method for actually detecting the haplotype of SAA1 using a DNA microarray will be described. The method shown here uses the method shown in US Pat. No. 6,036,463, but is not limited to this method. First, the entire SAA1 region is amplified (including 5 SNPs). The primer used in that case is shown below.

  Here, conditions for calculating Tm are shown below.

  Moreover, the probe corresponding to each SNP was designed as follows.

  Next, a haplotyping probe is created using the probe. Hereinafter, the case of haplotyping by the method of US Pat. No. 6,306,463 will be specifically described with respect to the SNP pair 2-4.

In this example, the synthesis on the substrate was not performed as in the above specification, and 4 types of 5 'terminal all-labeled oligo probes purified in the liquid phase and then purified (-13C> T wild type, mutant type, and 2995C> T wild type and mutant type) are mixed in equal amounts in the following combinations.
(1) -13C> T wild type + 2995C> T wild type (2) -13C> T wild type + 2995C> T mutant type (3) -13C> T mutant type + 2995C> T wild type (4) -13C > T variant + 2995C> T The variant mixed solution is discharged onto the substrate and fixed by the method disclosed in JP-A-11-187900.

  In another method, the same combination of sequences as described above is used, but instead of mixing the probes, a probe having both sequences in succession is synthesized. For example, in the above example 1, a 37 bp probe (5 ′ terminal all-label) having a sequence of −13C> T wild type and having a sequence of 2995C> T wild type was synthesized, and JP-A-11-187900 It discharges and fixes on a board | substrate by the method shown by gazette.

  Samples were purchased from the Human Science Research Resource Bank (HSRRB) extracted DNA from PSC (Pharma SNP Consortium). When used in this example, the sequence of the product obtained by amplifying the extracted genome using the SAA1 Forward and Reverse primers was obtained by a sequencer (ABI Prism 3100 Genetic Analyzer). The genotype results by the sequencer and the haplotype results determined by the five SNPs for 10 samples are shown below. As shown in the table, haplotypes of 5 samples out of 10 samples could not be determined, and it was found that there is a possibility of the above two to three diplotypes.

  In this embodiment, haplotyping is performed on these specimens.

  Hereinafter, a series of flow from preparation of a DNA chip to detection will be described in more detail. Here, an example using a DNA microarray in which a probe nucleic acid is immobilized on a substrate carrier by an ink jet method (Japanese Patent Laid-Open No. 11-187900) will be described, but the present invention is not limited to this method.

(Configuration of microarray)
FIG. 6 shows how the probe is fixed on the microarray. As described in detail in Japanese Patent Application Laid-Open No. 11-187900, the probe is immobilized using a method of discharging oligo DNA having a 5 ′ end thiolated by inkjet onto a surface-treated substrate. Here, the DNA used as a probe has a length of about 25 bases and was purchased from Bex Corporation.

(Preparing the target)
An example of amplification reaction (PCR) of nucleic acid derived from a specimen is shown below. An example of the amplification reaction solution composition is shown below. Here, the forward / reverse primer sequence is as shown above, but the combination of 5'-end Cy3-labeled F Primer + 5'-end phosphorylated R Primer and 5'-end Cy3-labeled R Primer + 5'-end phosphorylated F Primer PCR was performed with two types of combinations. As a result, only the Cy3-labeled chain remains in the subsequent single-stranding treatment, and the phosphorylated chain is decomposed to perform a hybridization reaction with the single-stranded target.

--PCR solution composition--
Takara LA Taq: 0.25μl
Genome DNA (50ng / μl): 1μl
Forward / Reverse Primer (1μM): 3 μl
dNTP (2.5 mM): 4.5 μl
buffer I: 12.5 μl
H ₂ O: 0.75 μl
Total: 25μl
The reaction solution having the above composition was subjected to an amplification reaction using a thermal cycler according to the temperature cycle protocol shown in FIG. After completion of the reaction, the amplification product is quantified by electrophoresis (BioAnalyzer: Agilent).

(Single-strand processing)
Using the amplified PCR product described above, a single-stranded target is produced by performing a single-stranded treatment. The reaction is adjusted so that 50 ng of amplification product is contained in the solution with reference to the above quantitative results. As a control, instead of Strandase λ Exonuclease, a solution obtained by diluting Strandase λ Exonuclease 100 times is added to carry out the reaction.

--Stranding reaction solution composition--
PCR product 50ng + H ₂ O: 8μl
10xStrandase Buffer: 1μl
Strandase λ Exonuclease: 1μl
Total: 10μl
After holding the reaction solution having the above composition at 37 ° C. for 20 minutes, primers and the like are removed using a purification column (QUIAGEN QIAquick PCR Purification Kit: manufactured by QUIGEN). After purification, the product is quantified by electrophoresis (BioAnalyzer: Agilent). At this time, no signal is observed when the single-stranded reaction is completed. Therefore, if an equal amount of PCR product is added and a solution in which the enzyme is diluted 100-fold instead of the enzyme is added, and an amount equal to the amount of product (molar concentration) of the control reaction is present in a single-stranded form, Perform the hybridization reaction.

(Hybridization)
The drained DNA microarray is set in a hybridization apparatus (Genomic Solutions Inc. Hybridization Station), and a hybridization reaction is performed with the following hybridization solution and conditions. The reaction may be performed manually using a slide glass and a hybridization chamber without using a hybridization apparatus.

(Hybridization solution)
An example of the composition of the hybridization solution is shown below.
“6 × SSPE / 10% Formamide / Target (Nucleic acid derived from an unknown sample) (Product obtained by single strand after PCR 0.5 nM) /0.05% SDS”
Dissolve 0.5 nM of the product that has been single-stranded after amplification in buffer (SSPE), and add Formamide to a final concentration of 10%. An SDS solution is added to this solution so that the final concentration is 0.05% to obtain a hybridization solution. Note that the concentration of the buffer (SSPE) is calculated in advance so as to be 6 × SSPE in the final solution state.

  The hybridization solution was heated to 92 ° C. and held for 2 minutes, and further held at 60 ° C. for 4 hours. Thereafter, washing was performed at 50 ° C. using 2 × SSC and 0.1% SDS. Furthermore, it wash | cleaned at 20 degreeC using 2 * SSC, and rinsed with the pure water according to the normal manual as needed, and drained with the spin dryer.

(Fluorescence measurement)
The above-described DNA microarray was subjected to fluorescence measurement under the following conditions using a fluorescence detection apparatus for DNA microarray (Axon, GenePix 4000B). The fluorescence measurement wavelengths were set to Cy3 and Cy5 measurement wavelengths, and the intensity of excitation light was adjusted so that the fluorescence measurement value was 30000 or less.

(Spot analysis)
The image of the fluorescence measurement result was analyzed with the data analysis software ArrayPro (manufactured by Media Cybernetics) for microarray, and the brightness value data for each spot was obtained.

(result)
The results of 10 specimens whose phases have been determined by the haplotyping are shown below. # 348, # 493, # 484 are identified by haplotyping of SNP pair 2-4, # 418 is identified by haplotyping of SNP pair 2-5, and # 317 diplotype is identified by haplotyping of SNP pair 1-3. I was able to.

(Example 2)
In the first embodiment, as a haplotyping method, a method that can cope with even an SNP separated by about 5 kbp is used, but such a method may not be used. In that case, it is necessary to select a combination of SNP typing and haplotyping that is more effective within the possible range. In the second embodiment, an optimum SNP selection state is shown for a case where haplotyping is effective between SNPs of 500 bp or less (within 500 bp). As a haplotyping method, a method using allele-specific PCR and hybridization on a substrate is used.

  The target is SNP in five locations of SAA1 as in Example 1. Find a diplotype that gives an equal genotype based on a given haplotype frequency, and pair that does not have D ′ = 1 (1-2,1-3, 1-5,2-4,2-5,3-4, The process is the same as in Example 1 until 3-6) and Table 6 for diplotypes (d1 to d11) that cannot be determined are prepared (the same table is shown as Table 14 below).

  However, in the haplotyping method used in this embodiment, it is assumed that the distance between SNP pairs that can be accurately typed is 500 bp or less. In that case, since haplotyping of 2-4 and 2-5 is not possible, haplotyping is performed only with 1-2 and 1-3. At this time, determination of d5 to d11 is possible.

From the above, as a configuration for obtaining the diplotype of SAA1,
(1) SNP typing of 2995C> T, 3010C> T (2) Haplotyping of -61C> G, -13T> C, -2G> A may be performed. By doing so, it becomes possible to determine the diplotypes of d5 to d11. Since d1 to d4 cannot be determined, the values of d2 and d4 are not reflected when represented by the diplotype with the higher probability. Therefore, the percentage of diplotypes that are truncated in this case is 0.037%, which is one in 2500 people. When only SNP typing was performed, one person per 100 people could be reduced to one person per 2500 people by combining one haplotyping.

  However, no matter how low the frequency, if d2 or d4 is an allele that correlates with a disease or a side effect, it is necessary to detect d2 or d4. In such a case, it is necessary to select a haplotyping method corresponding to a longer distance SNP pair or to amplify from mRNA. In the present embodiment, since it is not assumed that any of the haplotypes has a correlation with the phenotype, the ratio of the diplotype frequency that can be determined is focused. However, the algorithm according to the present invention can also be applied when a phenotype correlation haplotype is specified.

  Next, a method for simultaneously performing haplotyping and SNP typing in the present embodiment will be specifically described. Set the probes and primers for SNP typing as follows. Since 4 (2995C> T) and 5 (3010C> T) are close to each other, primers were designed to include both (PCR product length 516 bp).

  However, here, the following values were used as conditions for Tm calculation.

  For haplotyping, to design to haplotype 1-2 and 1-3 simultaneously, set an allele-specific primer at 1 (-61C> G) and 2 (-13T> C) and 3 ( Reverse Primer was set so that -2G> A) was included in the amplification product (PCR product length 486 bp). For 2 (-13T> C) and 3 (-2G> A), we designed a probe to be fixed on the substrate.

  FIG. 11 shows the positional relationship between the SNP site and the designed primers and probes.

  In the following examples, the template is created by performing amplification of the entire SAA1 region before performing amplification for SNP or haplotyping. The primer used in that case is shown below.

  Here, conditions for calculating Tm are shown below.

  As in Example 1, 10 extracted genome samples derived from the PSC strain were used. The genotype data relating to the 5 SNPs of the 10 specimens used are as shown in Example 1.

  Hereinafter, a series of flow from preparation of a DNA chip to detection will be described in more detail. Here, an example using a DNA microarray in which a probe nucleic acid is immobilized on a substrate carrier by an ink jet method (Japanese Patent Laid-Open No. 11-187900) will be described, but the present invention is not limited to this method.

(Configuration of microarray)
FIG. 6 shows how the probe is fixed on the microarray. As described in detail in Japanese Patent Application Laid-Open No. 11-187900, the probe is immobilized using a method of discharging oligo DNA having a 3 ′ end thiolated by inkjet onto a surface-treated substrate. Here, the DNA used as a probe has a length of about 25 bases and was purchased from Bex Corporation.

(Preparing the target)
An example of amplification reaction (PCR) of nucleic acid derived from a specimen is shown below. An example of the amplification reaction solution composition is shown below.

--PCR solution composition--
Takara LA Taq: 0.25μl
Genome DNA (50ng / μl): 1μl
Forward / Reverse Primer (1uM): 3μl
dNTP (2.5 mM): 4.5 μl
buffer I: 12.5 μl
H ₂ O: 0.75 μl
Total: 25μl
The reaction solution having the above composition was subjected to an amplification reaction using a thermal cycler according to the temperature cycle protocol shown in FIG.

  After completion of the reaction, the primer is removed using a purification column (QUIAGEN QIAquick PCR Purification Kit: manufactured by QUIGEN), and the amplification product is quantified by electrophoresis (BioAnalyzer: manufactured by Agilent).

(Sample processing for SNP typing)
Using the amplified PCR product described above, PCR is performed on the region containing the SNP site. For amplification, a Cy3-labeled Primer is used. The protocol at this time is shown below.
--PCR solution composition--
AmpliTaq Gold (Applied Biosystems): 0.2 μl
Template Genome DNA: 4ng
Forward / Reverse Primer: 1μM each
dNTP mix: 0.2 mM each
10xbuffer: 2.5μl
Total: 25μl
The reaction solution having the above composition was subjected to an amplification reaction using a thermal cycler according to the temperature cycle protocol shown in FIG.

(Haplotyping specimen processing: allele-specific PCR)
Allele-specific PCR is performed using the amplified PCR product described above. For amplification, PCR is performed using a Cy3-labeled and Cy5-labeled Forward Primer. The protocol at this time is shown below.

--PCR solution composition--
AmpliTaq Gold (Applied Biosystems): 0.2 μl
Template Genome DNA: 4 ng
Forward / Reverse Primer: 0.06μM each
dNTP mix: 0.2 mM each
10xbuffer: 2.5μl
Total: 25μl
The reaction solution having the above composition was subjected to an amplification reaction using a thermal cycler according to the temperature cycle protocol shown in FIG.

(Hybridization)
The drained DNA microarray is set in a hybridization apparatus (Genomic Solutions Inc. Hybridization Station), and a hybridization reaction is performed with the following hybridization solution and conditions. The reaction may be performed manually using a slide glass and a hybridization chamber without using a hybridization apparatus.

(Hybridization solution)
An example of the composition of the hybridization solution is shown below.
“6 × SSPE / 10% Formamide / Target (Nucleic acid derived nucleic acid) (PCR product 100 ng) /0.05% SDS”
The nucleic acid equivalent to 100 ng of nucleic acid derived from the above-mentioned amplified unknown sample is dissolved in a buffer (SSPE), and Formamide is added so that the final concentration becomes 10%. An SDS solution is added to this solution so that the final concentration is 0.05% to obtain a hybridization solution. Note that the concentration of the buffer (SSPE) is calculated in advance so as to be 6 × SSPE in the final solution state.

  The hybridization solution was heated to 92 ° C. and held for 2 minutes, and further held at 50 ° C. for 4 hours. Thereafter, washing was performed at 40 ° C. using 2 × SSC and 0.1% SDS. Furthermore, it wash | cleaned at 20 degreeC using 2 * SSC, and rinsed with the pure water according to the normal manual as needed, and drained with the spin dryer.

(Fluorescence measurement)
The above-described DNA microarray was subjected to fluorescence measurement under the following conditions using a fluorescence detection apparatus for DNA microarray (Axon, GenePix 4000B). The fluorescence measurement wavelengths were set to Cy3 and Cy5 measurement wavelengths, and the intensity of excitation light was adjusted so that the fluorescence measurement value was 30000 or less.

(Spot analysis)
The image of the fluorescence measurement result was analyzed with the data analysis software ArrayPro (manufactured by Media Cybernetics) for microarray, and the brightness value data for each spot was obtained.

(result)
The results of 10 specimens whose phases are determined by haplotyping according to this example are shown below. Because the haplotyping of SNP pair 1-3 determines the phase of # 317 and the haplotyping of SNP pairs 1-2 and 1-3 indicates that it is not d4 or d5, the phase of # 418 is determined to be d3. I was able to.

(Example 3)
In this example, the usefulness of the technique proposed by the present invention is shown using haplotype data relating to the human ALDH2 gene.
[1. Acquisition of haplotype data]
Nine SNPs of ALDH2 published by the HapMap project (http://www.hapmap.org/) were obtained. The nine SNPs are as shown in Table 21 below and FIG. The column of “rsSNPid” in Table 21 indicates the ID of each SNP. In the column “alleles”, allele at the SNP position appears. When “A / G” is shown, “A” is a wild type and “G” is a mutant type. The column “MAF” represents the frequency of Minor Allele (ie, mutant allele).

[2. Calculation of linkage disequilibrium coefficient]
The linkage disequilibrium coefficient between each SNP was calculated using HaploView (http://www.broad.mit.edu/mpg/haploview/). Table 22 shows the calculation results.

[3. Creating simplified haplotypes]
From Table 22, it was found that there was an SNP with a value of 1. Table 23 below shows combinations of SNPs with a value of 1.

  FIG. 13 is a diagram in which SNP positions having a value of 1 are grouped by connecting with a line.

  As described above, the haplotype is simplified as shown in Table 24 and FIG.

  From the combinations of four SNPs shown in Table 24 above, possible haplotype candidates are assembled as follows.

[4. Create haplotype candidates]
The haplotypes shown in Table 25 below and their frequencies were determined by the function of HaploView.

  Table 26 shows the combinations of Table 25 with the simplified haplotypes created earlier.

[5. Create Genotype Candidate]
When all possible genotypes composed of the haplotypes shown in Table 26 are created, Table 27 below is obtained.

[6. Policy decision for haplotype determination]
From the results of Table 27 above, it can be seen that in order to determine the haplotype consisting of the nine SNPs targeted in this example, genotyping should be performed at the four SNP positions shown in Table 28. It was.

It is a figure which shows the algorithm in this invention. It is a determination flowchart. It is a figure explaining the case where SNP which becomes (DELTA) ² = 1 is put together into one group among 10 SNPs, and what collected only the representative SNP of each group is called the simplified haplotype. It is a figure for demonstrating the case where the linkage disequilibrium coefficient D 'is calculated about all the pairs between SNPs which comprise the simplified haplotype, and only the pair of D' ≠ 1 is selected. It is the schematic diagram of SNP which performs haplotyping, and SNP which performs SNP typing selected by this algorithm in ten SNPs. It is a schematic diagram of SNP typing using a DNA chip. It is a schematic diagram of haplotyping using a DNA chip. It is a figure which shows the structure which performs SNP typing and haplotyping simultaneously with a DNA chip. SNPa and b show haplotyping probes, and c and d show SNP typing probes fixed. Wild type (W) and mutant type (M) probes are immobilized on each of the 4 SNPs. It is a block diagram which shows the structure of the information processing apparatus which can apply the program which performs the haplotyping location determination by embodiment. It is a flowchart of the computer system by embodiment. FIG. 4 is a diagram showing the positions of primers and probes in Example 2. It is the figure which showed the SNP position of ALDH2. It is the figure which showed the position of SNP used as the linkage disequilibrium coefficient by SNP of ALDH2. FIG. 4 is a simplified haplotype diagram of ALDH2. It is a figure which shows the temperature cycle of PCR reaction. It is a figure which shows the temperature cycle of PCR reaction. It is a figure which shows the temperature cycle of PCR reaction.

Explanation of symbols

91 Central processing unit 92 Storage device 93 RAM
94 Input / Output Device 95 Bus 101 Step 102 for Inputting Haplotype Step 102 for Constructing Simplified Haplotype 103 Step for Outputting Location for Haplotyping

Claims (7)

  Perform haplotyping on a part of the plurality of SNPs constituting the haplotype of the target gene, perform SNP typing on the remaining part or all, and determine the haplotype of the target gene from the result of both typings Haplotype determination method characterized by   The haplotype determination method according to claim 1, wherein the SNP to be haplotyping is selected based on frequency information from SNP pairs included in a plurality of diplotypes having the same genotype.   The haplotype determination method according to claim 2, wherein the SNP that is a target of the SNP typing is selected from SNPs other than the SNP that is the target of the haplotyping.   The haplotype determination method according to claim 1 or 2, wherein the SNP to be haplotyping is selected based on information on linkage disequilibrium between SNPs. 2. The haplotype determination method according to claim 1, wherein the SNP to be haplotyping is selected from the following steps.
1) calculating a linkage disequilibrium coefficient Δ ² between SNPs by the following formula (1), grouping SNPs where Δ ² = 1, and selecting a representative SNP in the group;
(However, when one SNP has alleles a and b and the other SNP has alleles c and d, the allele frequencies are Pa, Pb (= 1-Pa), Pc, Pd (= 1-Pc), (Haplotype frequency is Pac, Pad, Pbc, Pbd.)
2) selecting a diplotype having the same genotype from all the combinations of the representative SNPs;
3) Calculate the linkage disequilibrium coefficient D ′ according to the following equation (2) for all pairs of the representative SNPs,
D '≠ 1
Selecting a pair to be
4) A step of narrowing down representative SNP pairs that can be haplotyped based on the frequency of representative SNP pairs in the selected diplotype.   6. The haplotype determination method according to claim 5, wherein the representative SNP pair capable of being haplotyped in step 4) has a position on the gene of 500 bp or less.   2. The haplotype determination method according to claim 1, wherein the SNP typing and the haplotyping are performed simultaneously using a DNA chip.