Method for analyzing mixed DNA by using gamma model

Abstract

The invention discloses a method for analyzing mixed DNA by using a gamma model, which comprises the following steps: S1, preprocessing data to obtain gene loci, allelic genes and allelic gene fluorescence intensity peak height information in a mixed DNA sample; s2, evidence weight analysis; s3, calculating the evidence probability of the genotype; and S4, calculating likelihood ratios under the two hypotheses. According to the method, the gamma distribution model is introduced, so that each donor genotype in the mixed DNA sample can be more accurately calculated and identified, and particularly, higher resolution can be provided under the condition that the peak heights of the genotypes are overlapped. According to the method, through optimization algorithm design, unnecessary calculation processes in sample analysis are reduced, and the processing speed of the whole DNA mixed sample analysis process is increased. According to the method, the influence of the peak height variability in the DNA sample is considered, a more accurate result can be provided in wider practical application, and particularly, the error rate is reduced when high-variability and complex mixed samples are processed.

Description

Technical Field

[0001] The present invention relates to the field of biological detection technology, and in particular to a method for analyzing mixed DNA using a gamma model. Background Art

[0002] Mixed DNA samples are samples mixed from multiple donors and are common in forensic identification and genetic analysis. Existing methods for analyzing mixed DNA samples usually rely on empirical assumptions and have difficulty accurately distinguishing the genotypes contributed by multiple donors. Especially in the case of high overlapping peaks, it is difficult for traditional methods to reliably distinguish the contributions of different donors. In addition, existing methods mostly rely on complex calculation processes, such as permutation and combination calculations, which results in long calculation times when analyzing large-scale samples and cannot meet the needs of rapid analysis.

[0003] Traditional analysis methods based on simple statistical distributions (such as normal distribution) cannot adapt well to the asymmetry of real data and the variability of peak height when dealing with complex DNA mixed samples, resulting in insufficient accuracy and stability of the results. Some existing technologies fail to effectively consider the variability of DNA peak heights and ignore the fluctuation of sample peak heights contributed by different donors, resulting in insufficient accuracy of the model when dealing with highly variable samples. Summary of the invention

[0004] In view of the deficiencies of the prior art, the present invention aims to provide a method for analyzing mixed DNA using a gamma model.

[0005] In order to achieve the above object, the present invention adopts the following technical solution:

[0006] A method for analyzing mixed DNA using a gamma model comprises the following steps:

[0007] S1, data preprocessing, obtaining loci, alleles and fluorescence intensity peak height information of alleles in mixed DNA samples;

[0008] S2. Weight of evidence analysis:

[0009] S2.1. Determine the donor genotype:

[0010] First, according to the number of alleles analyzed in step S1, the locus with the largest number of alleles is found, and the minimum number of donors is evaluated according to the number of alleles of the locus; after the number of donors n is determined, the genotype of the donor is obtained according to the permutation and combination of alleles;

[0011] S2.2, Peak Height Modeling:

[0012] The gamma distribution model is shown as follows:

[0013]

[0014] Among them, γ is the observed peak height, α and β are the shape parameter and scale parameter to be determined respectively; the shape parameter α of the gamma distribution model is determined by the mixing ratio and the peak height variation coefficient, and the scale parameter β of the gamma distribution model is determined by the peak height variation coefficient and the peak height expectation value; after determining α and β, the observed peak height γ of a peak is directly substituted into the gamma distribution model to obtain the weight of the corresponding allele;

[0015] S2.3, calculation parameters:

[0016] After initializing the mixing ratio, peak height variation coefficient, and expected peak height of the gamma distribution model, α and β are derived according to the peak height variation coefficient ω;

[0017] Mean μ and variance σ of the gamma distribution 2 The calculation of is shown in formula (2) and formula (3):

[0018] μ=α·β (2)

[0019] σ 2 =α·β 2 (3)

[0020] The coefficient of variation of peak height ω is defined as the ratio of the standard deviation to the mean of the peak height; where the standard deviation is the standard deviation of the observed peak height and the expected peak height, and the mean of the peak height is the expected peak height, then:

[0021]

[0022] Substituting in the mean and variance:

[0023]

[0024] So the relationship between the peak height variation coefficient and shape parameter is as follows:

[0025]

[0026] Because of formula (2), the relationship between the peak height variation coefficient and the scale parameter is as follows:

[0027] β=μ·ω 2 (7)

[0028] In summary, after assigning the initial values ​​of the peak height variation coefficient ω and the peak height mean μ, the scale parameter β of the gamma distribution model can be obtained according to formula (7);

[0029] For the gamma distribution model of mixed DNA samples, the mixing ratio of each donor needs to be considered. Each donor has its own shape parameter, while the scale parameter is a global parameter that is the same for all donors.

[0030] For a specific allele, the shape parameters of each donor are calculated by the mixing ratio and known parameters; according to formula (6), we have:

[0031]

[0032] Among them, α i is the shape parameter of donor i, m i is the mixing ratio of donor i;

[0033] Since β is determined by two global parameters μ and ω, the scale parameter β remains unchanged.

[0034] The observed peak height γ and shape parameter α i Substituting the scale parameter β into formula (1) directly obtains the weight of a single allele, and then multiplying the weights of each allele in a genotype to obtain the evidence weight of the genotype;

[0035] S3. Calculate the probability of evidence for genotype:

[0036] After obtaining the evidence weight of the genotype, the evidence weight of the genotype and the genotype frequency are multiplied to obtain the evidence probability of the corresponding genotype. The genotype frequency is obtained through the frequency library;

[0037] S4. Calculate the likelihood ratio under the two hypotheses:

[0038] The likelihood ratio calculation is based on two assumptions, the prosecution assumption and the defense assumption. The prosecution accuses someone, claiming that someone with a certain genotype has DNA in a mixed DNA sample, and the defense argues that the genotype is a random individual in the population. The prosecution hopes to control the parameters to maximize the possibility of their assumptions under their assumptions, while the defense hopes to control the parameters to maximize the possibility of DNA evidence under their assumptions. Then, on the basis of both parties doing their best, the ratio between the two is fixed, which is the absolute evidentiary force of the evidence. This is the likelihood ratio, and its calculation is shown in the following formula:

[0039]

[0040] Among them, P(E|Hp) is the strength of evidence under the prosecution hypothesis, and P(E|Hd) is the strength of evidence under the defense hypothesis; P(E|G,Hp) is the weight of evidence of genotype under the prosecution hypothesis, and P(G|Hp) is the probability of evidence of genotype under the prosecution hypothesis; P(E|G,Hd) is the weight of evidence of genotype under the defense hypothesis, and P(G|Hd) is the probability of evidence of genotype under the defense hypothesis; Σ summation is to consider the product of the probability of evidence of each genotype obtained by permutation and combination and the weight of evidence under two assumptions, and then sum up all the products after obtaining them.

[0041] Furthermore, the specific process of step S1 is:

[0042] After obtaining the mixed DNA sample, PCR amplification is performed first, and the amplified products are separated by capillary electrophoresis. In capillary electrophoresis, fluorescently labeled probes are used to identify DNA fragments of different lengths. The genetic analyzer uses laser excitation to capture and record the fluorescence signal of each DNA fragment. The fluorescence signal is converted into data, which contains the genotype information of each sample on different markers and the migration time and fluorescence intensity of each DNA fragment. The peak of fluorescence intensity is then analyzed to identify the alleles at each site.

[0043] Furthermore, in step S4, if Hp and Hd under the maximum likelihood estimation calculated according to formula (10) are not optimal, μ and m in step S2.3 are i ,ω will be updated respectively according to the feedback results;

[0044] For donor i, update the mixing ratio m i After that, its expected peak height is updated by:

[0045] μ i =α i ×β (9)

[0046] Each update μ i After that, combined with the updated mixing ratio m i The observed peak height γ and the expected peak height μ under i The standard deviation between them updates the coefficient of variation ω, and then further obtains the updated α of donor i according to equations (8) and (7): i and β, and continue the iterative calculation of steps S3 and S4 until the optimal result is obtained in step S4.

[0047] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program implements the above method when executed by a processor.

[0048] The present invention also provides a computer device, comprising a processor and a memory, wherein the memory is used to store a computer program; and when the processor is used to execute the computer program, the above method is implemented.

[0049] The beneficial effects of the present invention are:

[0050] (1) Improving the accuracy of genotype identification: By introducing the gamma distribution model, the present invention can more accurately calculate and identify the genotypes of each donor in a mixed DNA sample, especially when the genotype peaks overlap highly, and can provide higher resolution.

[0051] (2) Improving processing speed and computing efficiency: The present invention reduces unnecessary computing processes in sample analysis by optimizing algorithm design, thereby improving the processing speed of the entire DNA mixed sample analysis process.

[0052] (3) Enhanced model accuracy and stability: The present invention takes into account the impact of peak height variability in DNA samples and can provide more accurate results in a wider range of practical applications, especially reducing the error rate when dealing with high variability and complex mixed samples. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 The genotyping result spectrum (blue, green and yellow channels) of the mixed DNA sample obtained in Example 2 of the present invention;

[0054] Figure 2 This is a graph of the typing results of the mixed DNA sample obtained in Example 2 of the present invention (red and orange channels). DETAILED DESCRIPTION

[0055] The present invention will be further described below in conjunction with the accompanying drawings. It should be noted that this embodiment is based on the technical solution and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to this embodiment.

[0056] Example 1

[0057] This embodiment provides a method for analyzing mixed DNA using a gamma model, comprising the following steps:

[0058] S1. Data preprocessing.

[0059] After obtaining the mixed DNA sample, PCR amplification is first performed. For the multiple donor DNAs contained in the mixed DNA sample, specific microsatellite markers (such as STR loci) are usually selected for amplification, and the target DNA region is amplified using designed primers.

[0060] After the PCR amplification process is completed, the amplified products are separated by capillary electrophoresis. In capillary electrophoresis, fluorescently labeled probes are used to identify DNA fragments of different lengths. The genetic analyzer uses laser excitation to capture and record the fluorescence signal of each DNA fragment. The fluorescence signal is converted into data and a .fsa format file is generated. This file contains the genotype information of the mixed DNA sample at different markers and data such as the migration time and fluorescence intensity of each DNA fragment. The .fsa file records the migration time and peak intensity of each DNA fragment during the electrophoresis process, and this information can be used for further analysis. In the analysis stage, these peaks are parsed by the analysis software GeneMapper to identify the alleles at each site. After preliminary processing of the mixed DNA sample map in the analysis software, data containing sample name, locus name, allele name and allele peak height are exported.

[0061] S2. Weight of evidence analysis

[0062] For mixed DNA samples, the peak height of each allele at each locus is all the quantitative information that can be used, and these quantitative information need to be analyzed to obtain the weight of evidence. On the premise of preliminarily determining how many donors' DNA samples the mixed DNA sample is mixed from, all genotype combinations are generated, and each combination is the genotype of each single donor. The mixing ratio of each donor and the possibility of each genotype under this sample fact are calculated. This means that the method of this embodiment does not discard any genotype, but only obtains how much weight a donor has for a certain genotype, which is the meaning of the weight of evidence. Specifically comprising the following steps:

[0063] S2.1. Determine the donor genotype:

[0064] First, according to the number of alleles analyzed in step S1, the locus with the largest number of alleles is found, and the minimum number of donors is evaluated based on the number of alleles of the locus. This analysis is subjective and can be modified again. In fact, since the two alleles of the donor may be the same or different, and the alleles of different donors may overlap, the number of donors cannot be determined from the analyst's perspective. However, the minimum number of donors can be evaluated based on the number of the largest allele.

[0065] For example, if we observe that there are six alleles on a certain locus, which is the locus with the largest number of alleles in the entire sample file, we will take this locus as the research object. Rename the alleles in the order of molecular weight from small to large according to international conventions. The codes are named A, B, C, D, E, and F respectively, and all possible allele type combinations can be generated according to the codes. After determining the number of donors n, the genotype of the donor can be obtained based on the permutations and combinations. There are 15 genotypes here. If the permutations of the donors are considered, there will be 90 cases, as shown in Table 1.

[0066] Table 1 Possible genotype combinations

[0067]

[0068] S2.2, Peak Height Modeling:

[0069] Qualitatively, there are certainly many conjectures about an observed allele peak, such as: what proportion of the height of this peak is contributed by multiple donors? If this peak is not due to the overlap and occlusion of the height caused by the participation of different donors, what is its actual height? When estimating the height of the observed peak, what distribution do the errors of these estimates follow? What is the error between multiple estimates and the true peak height, and how to measure this error.

[0070] Quantitatively, the residuals of the observed and expected peak heights for a given allele follow a gamma distribution, which is a quantitative continuous model with skewness. The gamma distribution has two parameters, a shape parameter that determines whether the curve is stretched or compressed, and a scale parameter that determines the scale of the curve. There is also an implicit coefficient of variation in the height of the gamma distribution model, which is the ratio of the standard deviation of the peak to the expected peak height. The gamma distribution model for allele peak heights is expressed as follows:

[0071]

[0072] As shown in formula (1), γ is the observed peak height, α and β are the shape parameter and scale parameter to be determined, respectively. Γ(α) represents the gamma function, which is used to describe the distribution of residuals, especially when the residuals have positive skewness, it is used to model non-negative variables. Here it is the normalization constant of the entire gamma distribution, ensuring that the integral of the entire probability density function is 1.

[0073] The contribution of each donor to the peak height of the allele determines the shape parameter α, while the scale parameter is constant for all loci in the mixed DNA sample. A gamma distribution model of allele peak height is constructed, and the shape parameter α of the gamma distribution model is determined by the mixing ratio and the peak height variation coefficient, and the scale parameter β of the gamma distribution model is determined by the peak height variation coefficient and the peak height expectation value. After determining α and β, the observed peak height γ of a peak is directly substituted into the gamma distribution model to obtain the weight of the corresponding allele.

[0074] S2.3, calculation parameters:

[0075] After initializing the mixing ratio, peak height variation coefficient, and expected peak height of the gamma distribution model, α and β are derived according to the peak height variation coefficient ω.

[0076] Mean μ and variance σ of the gamma distribution 2 The calculation of is shown in formula (2) and formula (3):

[0077] μ=α·β (2)

[0078] σ 2 =α·β 2 (3)

[0079] The coefficient of variation of peak height ω is defined as the ratio of standard deviation to mean peak height. The standard deviation is the standard deviation of the observed peak height and the expected peak height, and the mean peak height is the expected peak height. Then:

[0080]

[0081] Substituting in the mean and variance:

[0082]

[0083] So the relationship between the peak height variation coefficient and shape parameter is as follows:

[0084]

[0085] Because of formula (2), the relationship between the peak height variation coefficient and the scale parameter is as follows:

[0086] β=μ·ω 2 (7)

[0087] In summary, after assigning the initial values ​​of the peak height variation coefficient ω and the peak height mean μ, the scale parameter β of the gamma distribution model can be obtained according to formula (7).

[0088] For the gamma distribution model of mixed DNA samples, the mixing ratio of each donor needs to be considered. Each donor has its own shape parameter, while the scale parameter is a global parameter that is the same for all donors.

[0089] For a specific allele, the shape parameter α is the sum of the relative contributions of all donors, so the shape parameter of each donor can be calculated by the mixing ratio and known parameters. According to formula (6), we have:

[0090]

[0091] Among them, α i is the shape parameter of donor i, m i is the mixing ratio of donor i.

[0092] Since β is determined by two global parameters μ and ω, the scale parameter β remains unchanged.

[0093] The observed peak height γ and shape parameter α i Substituting the scale parameter β into formula (1) directly obtains the weight of a single allele, and then multiplying the weights of each allele in a genotype to obtain the evidence weight of the genotype.

[0094] Substitute the genotype evidence weight into the subsequent hypothesis formula (10) to obtain Hp and Hd under maximum likelihood estimation. If the result is not optimal, μ, m i , ω will be updated respectively according to the feedback results, thus iterating.

[0095] For donor i, update the mixing ratio m i After that, its expected peak height can be updated by the following formula:

[0096] μ i =α i ×β (9)

[0097] Each update μ i After that, combined with the updated mixing ratio m i The observed peak height γ and the expected peak height μ under i The standard deviation between them updates the coefficient of variation ω, and then further obtains the updated α of donor i according to equations (8) and (7): i and β, and continue iterating until the optimal result is obtained.

[0098] S3. Calculate the probability of evidence for genotype:

[0099] After obtaining the genotype evidence weight, multiply the genotype evidence weight and the genotype frequency to obtain the corresponding genotype evidence probability.

[0100] To use IdentiFiler Plus TMTaking the PCR amplification kit as an example, a 5-color reagent amplification system includes 16 loci, and the frequency library needs to include the frequency of each allele on each locus in the population. The method of this embodiment uses the Iden population frequency library available internationally, which has x loci. As shown in Table 2.

[0101] Table 2 Locus information

[0102]

[0103] The file information in the frequency library is shown in Table 3:

[0104] Table 3 Frequency library information

[0105]

[0106] S4. Calculate the likelihood ratio under the two hypotheses:

[0107] The likelihood ratio calculation is based on two assumptions, the prosecution assumption and the defense assumption. The prosecution accuses someone, claiming that someone with a certain genotype of DNA is in the mixed DNA sample, and the defense argues that the genotype is a random individual in the population. The prosecution hopes to control the parameters to maximize the possibility of their assumptions under their assumptions, while the defense hopes to control the parameters to maximize the possibility of DNA evidence under their assumptions. It can be simply understood as: a third party sees a fact, but the prosecution and the defense, in their respective positions, use the results of their own simulation operations to explain this fact, and try to make the results more consistent with their own statements. Then, on the basis of both parties doing their best, the ratio of the two is fixed, which is the absolute evidentiary force of the evidence. This is the likelihood ratio, and its calculation is shown in the following formula:

[0108]

[0109] Among them, P(E|Hp) is the strength of evidence under the prosecution hypothesis, and P(E|Hd) is the strength of evidence under the defense hypothesis. P(E|G,Hp) is the weight of evidence of genotype under the prosecution hypothesis, and P(G|Hp) is the probability of evidence of genotype under the prosecution hypothesis. P(E|G,Hd) is the weight of evidence of genotype under the defense hypothesis, and P(G|Hd) is the probability of evidence of genotype under the defense hypothesis. Σ summation is to consider the product of the probability of evidence and the weight of evidence of each genotype obtained by permutation and combination under two hypotheses, and then sum all the products after obtaining them.

[0110] Example 2

[0111] This embodiment provides an application example of the method described in Embodiment 1.

[0112] 1. Get mixed DNA samples from multiple people, amplify them by PCR, and put them on a genetic analyzer to get an fsa file. Put the fsa file on GeneMapper to get a preliminary mixed STR map. The blue, green, and yellow channels are as follows: Figure 1 As shown, the red and orange channels are Figure 2 shown.

[0113] 2. Export the alleles and allele peak height information at each locus to a table. The basic information is shown in Table 4.

[0114] Table 4 Allele information of samples

[0115]

[0116]

[0117]

[0118] 3. The reference typing used in this example is the typing of standard product 9947, as shown in Table 5:

[0119] Table 5 Genotype of standard 9947

[0120]

[0121] 4. The allele frequency file used in this embodiment is the population frequency library of the relevant identifier published by the National Institute of Standards and Technology of the United States.

[0122] 5. Set the initial parameters of the gamma distribution model. There are 4 initial parameters of the model, namely, mixing ratio, peak height variation coefficient, expected peak height and maximum number of iterations, as shown in Table 6.

[0123] Table 6 Initial parameters of the gamma distribution model

[0124] Mixing ratio Peak height variation coefficient Peak Expectation Maximum number of iterations 0.33333 0.2～1 500-8000 50

[0125] 6. The calculation results under the prosecution assumption and the defense assumption are obtained, as shown in Table 7.

[0126] Table 7 Calculation results

[0127]

[0128]

[0129] It should be noted that the donor mixing ratio is the optimal mixing ratio of the donor under the prosecution and defense propositions, the expected peak height change value is the expected peak height of the allele, and the change value is the peak height variation coefficient.

[0130] For those skilled in the art, various corresponding changes and modifications can be made according to the above technical solutions and concepts, and all of these changes and modifications should be included in the protection scope of the claims of the present invention.

Claims

1. A method for analyzing mixed DNA using a gamma model, characterized in that: The steps include: S1, data preprocessing, obtaining loci, alleles and fluorescence intensity peak height information of alleles in mixed DNA samples; S2. Weight of evidence analysis: S2.

1. Determine the donor genotype: First, according to the number of alleles analyzed in step S1, the locus with the largest number of alleles is found, and the minimum number of donors is evaluated according to the number of alleles of the locus; after the number of donors n is determined, the genotype of the donor is obtained according to the permutation and combination of alleles; S2.2, Peak Height Modeling: The gamma distribution model is shown as follows: Among them, γ is the observed peak height, α and β are the shape parameter and scale parameter to be determined respectively; the shape parameter α of the gamma distribution model is determined by the mixing ratio and the peak height variation coefficient, and the scale parameter β of the gamma distribution model is determined by the peak height variation coefficient and the peak height expectation value; after determining α and β, the observed peak height γ of a peak is directly substituted into the gamma distribution model to obtain the weight of the corresponding allele; S2.3, calculation parameters: After initializing the mixing ratio, peak height variation coefficient, and expected peak height of the gamma distribution model, α and β are derived according to the peak height variation coefficient ω; Mean μ and variance σ of the gamma distribution 2 The calculation of is shown in formula (2) and formula (3): μ=α·β (2) s 2 =a·b 2 (3) The coefficient of variation of peak height ω is defined as the ratio of the standard deviation to the mean of the peak height; where the standard deviation is the standard deviation of the observed peak height and the expected peak height, and the mean of the peak height is the expected peak height, then: Substituting in the mean and variance: So the relationship between the peak height variation coefficient and shape parameter is as follows: Because of formula (2), the relationship between the peak height variation coefficient and the scale parameter is as follows: β=μ·ω 2 (7) In summary, after assigning the initial values ​​of the peak height variation coefficient ω and the peak height mean μ, the scale parameter β of the gamma distribution model can be obtained according to formula (7); For the gamma distribution model of mixed DNA samples, the mixing ratio of each donor needs to be considered. Each donor has its own shape parameter, while the scale parameter is a global parameter that is the same for all donors. For a specific allele, the shape parameters of each donor are calculated by the mixing ratio and known parameters; according to formula (6), we have: Among them, α i is the shape parameter of donor i, m i is the mixing ratio of donor i; Since β is determined by two global parameters, μ and ω, the scale parameter β remains unchanged; The observed peak height γ and shape parameter α i Substituting the scale parameter β into formula (1) directly obtains the weight of a single allele, and then multiplying the weights of each allele in a genotype to obtain the evidence weight of the genotype; S3. Calculate the probability of evidence for genotype: After obtaining the evidence weight of the genotype, the evidence weight of the genotype and the genotype frequency are multiplied to obtain the evidence probability of the corresponding genotype. The genotype frequency is obtained through the frequency library; S4. Calculate the likelihood ratio under the two hypotheses: The likelihood ratio calculation is based on two assumptions, the prosecution assumption and the defense assumption. The prosecution accuses someone, claiming that someone with a certain genotype has DNA in a mixed DNA sample, and the defense argues that the genotype is a random individual in the population. The prosecution hopes to control the parameters to maximize the possibility of their assumptions under their assumptions, while the defense hopes to control the parameters to maximize the possibility of DNA evidence under their assumptions. Then, on the basis of both parties doing their best, the ratio between the two is fixed, which is the absolute evidentiary force of the evidence. This is the likelihood ratio, and its calculation is shown in the following formula: Among them, P(E|Hp) is the strength of evidence under the prosecution hypothesis, and P(E|Hd) is the strength of evidence under the defense hypothesis; P(E|G,Hp) is the weight of evidence of genotype under the prosecution hypothesis, and P(G|Hp) is the probability of evidence of genotype under the prosecution hypothesis; P(E|G,Hd) is the weight of evidence of genotype under the defense hypothesis, and P(G|Hd) is the probability of evidence of genotype under the defense hypothesis; Σ summation is to consider the product of the probability of evidence of each genotype obtained by permutation and combination and the weight of evidence under two assumptions, and then sum up all the products after obtaining them.

2. The method according to claim 1, characterized in that The specific process of step S1 is: After obtaining the mixed DNA sample, PCR amplification is performed first, and the amplified products are separated by capillary electrophoresis. In capillary electrophoresis, fluorescently labeled probes are used to identify DNA fragments of different lengths. The genetic analyzer uses laser excitation to capture and record the fluorescence signal of each DNA fragment. The fluorescence signal is converted into data, which contains the genotype information of each sample on different markers and the migration time and fluorescence intensity of each DNA fragment. The peak of fluorescence intensity is then analyzed to identify the alleles at each site.

3. The method according to claim 1, characterized in that In step S4, if Hp and Hd calculated by the maximum likelihood estimation according to equation (10) are not optimal, μ and m in step S2.3 are i ,ω will be updated respectively according to the feedback results; For donor i, update the mixing ratio m i After that, its expected peak height is updated by: m i =a i ×b (9) Each update μ i After that, combined with the updated mixing ratio m i The observed peak height γ and expected peak height μ under i The standard deviation between them updates the coefficient of variation ω, and then further obtains the updated α of donor i according to equations (8) and (7): i and β, and continue the iterative calculation of steps S3 and S4 until the optimal result is obtained in step S4.

4. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 3 is implemented.

5. A computer device, characterized in that: The method comprises a processor and a memory, wherein the memory is used to store a computer program; when the processor is used to execute the computer program, the method described in any one of claims 1 to 3 is implemented.

Images