Method for detecting dominant clones resulting from exogenous dna insertion mutations

Abstract

The application discloses a detection method based on dominant clones caused by exogenous DNA insertion mutation, and the method comprises the following steps: 1, preprocessing of virus IS information data by positive and negative strand correction and redundant integration; 2, after the preprocessing is completed, IS caused by non-specific amplification is identified to remove false positive IS detection results caused by PCR non-specific amplification; 3, according to a site clustering algorithm, IS in adjacent positions is combined and clustered; 4, according to the clustering results, a clone plane is analyzed, and dominant clones are described; and the application utilizes a PCR primer similarity ranking algorithm to identify and remove non-specific amplification caused by possible PCR primer similarity, so that the false positive of IS identification is significantly reduced in actual application, and the accuracy is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of bioinformatics, and particularly relates to a detection method of dominant clones caused by exogenous DNA insertion mutation. BACKGROUND

[0002] Gene therapy refers to a treatment method of using molecular biology means to introduce exogenous DNA into the genome of a gene-deficient cell to restore the normal function of the cell. However, safety has always been one of the important problems that have plagued the development of the therapy in the development process of gene therapy. Integration vectors are a class of vector DNA sequences commonly used in gene therapy, which are used to load exogenous DNA fragments and integrate into the host genome in an insertion manner. Among them, lentiviral vectors (LVs), adeno-associated virus vectors (AAV) and other vectors have become ideal tool vectors for gene therapy due to their high gene introduction efficiency or stable expression ability in target cell genes.

[0003] When the integration vector is integrated into the host genome, it can cause instability and rearrangement of the host cell genome, and can possibly lead to disorder of gene expression of the host cell, escape from immune recognition of the host and long-term self-sustaining survival, and thus eventually lead to the occurrence of cancer. At present, in preclinical studies and some clinical trials using viral vectors, carcinogenic phenomena caused by integration of vectors into the genome have been observed. In an AAV viral gene therapy experiment on dogs, researchers found that some therapeutic gene fragments carried by AAV viruses were integrated near the genes controlling growth on the chromosomes of the dogs, and some liver cells of the dogs divided faster than other cells and formed cell clusters, which had the possibility of inducing cancer. Therefore, after the gene therapy product is put on the market, detection of the carcinogenic risk caused by vector insertion is one of the important links of gene therapy. However, in the process of continuous cell therapy, the occurrence time of this potential tumorigenicity is often significantly lagging and difficult to predict. At present, there is no good method to quantitatively evaluate the tumorigenicity and tumorigenicity of insertion mutation. Most of the current IS detection methods are based on the method of amplifying the ends of the insertion fragment by PCR to identify IS. In actual application, when the target fragment is low, non-specific amplification phenomenon will occur, resulting in false positive results of IS identification. SUMMARY

[0004] To solve the prior art in the process of persistent cell therapy, the occurrence time of such potential tumorigenicity often has obvious hysteresis and is difficult to predict, and there is no better method to quantitatively evaluate the tumorigenicity and tumorigenicity of insertion mutation; the current IS detection method is mostly based on the method of amplifying the end of the insertion fragment to identify IS, and in actual application, when the target fragment is low, non-specific amplification phenomenon will occur, resulting in the defect of false positive results of IS identification, and the application provides a detection method based on the dominant clone caused by exogenous DNA insertion mutation.

[0005] In order to solve the above technical problems, the application provides the following technical solutions:

[0006] The application provides a detection method based on the dominant clone caused by exogenous DNA insertion mutation, which comprises the following steps:

[0007] Step 1, pre-processing of positive and negative strand correction and redundancy integration of virus IS information data;

[0008] Step 2, after pre-processing, IS caused by non-specific amplification is identified to remove false positive IS detection results caused by PCR non-specific amplification;

[0009] Step 3, according to the site clustering algorithm, IS in the adjacent position is combined and clustered;

[0010] Step 4, according to the clustering result, the clone plane is analyzed, and the dominant clone is described.

[0011] As a preferred technical solution of the application, the virus IS information data in step 1 is a set of virus IS, wherein each IS indicates the integration position on the host genome.

[0012] As a preferred technical solution of the application, the specific operation of correcting the positive and negative strands of the virus IS information data in step 1 is that if the gene where the IS is located is a negative strand and the IS read is a negative strand, the IS is converted into a positive strand.

[0013] As a preferred technical solution of the application, the method for redundancy integration of the virus IS information data is that if the site and the positive and negative strand information of the IS are completely consistent, it is combined into one record; the site and the chromosome information corresponding to the IS are encoded to simplify it into a number as the basis of the site clustering algorithm.

[0014] As a preferred technical solution of the application, the method for identifying IS caused by non-specific amplification in step 3 comprises three steps:

[0015] Step A, extract all PCR primer sequences, calculate the reverse complementary sequence according to the primer sequence, become the primer sequence library S1;According to the position of each IS, extract several base pairs before and after it to form the sequence library S2 to be matched;On the target genome, a sufficient number of random sites are randomly selected to form the background sequence library S3;

[0016] Step B, matching scoring, for each sequence in S2 and S3, match to each primer sequence and its reverse complementary sequence in the primer sequence library respectively;In the matching results of the sequence and its reverse complementary sequence, select the highest matching score in all scoring as the similarity score of the IS and the primer;After matching, each IS in S2 and S3 obtains a similarity score vector, and the vector length is equal to the number of primers;For each primer, the similarity score corresponding to the primer in S2 and S3 is sorted, and the sorted score is converted into a percentile, such as the score ranked in the front 1%, all converted into 1, and so on;After conversion, each IS corresponds to a percentile score vector, which represents the ranking of the similarity of the sequence and the primer sequence in all sequences;

[0017] Step C, judgment of non-specific amplification;If there is no non-specific amplification, the sequences in S2 and S3 are completely random, and the distribution of the corresponding similarity ranking score is theoretically uniform;And in the case of non-specific amplification, the sorting distribution of the score in S2 will be biased to the front part of the ranking;By calculating the ranking vector of each IS in step 2, the probability of each IS in step 2 being non-specific amplification is obtained;The calculation method is to multiply the percentile score vector of each IS, if the probability is less than a certain threshold T, the IS is marked as non-specific amplification;The IS marked as specific amplification will be removed in the subsequent analysis.

[0018] As a preferred technical scheme of the present application, the step 3 is combined according to the position clustering algorithm, and the method of clustering is that the IS is grouped according to the chromosome, and the IS in each chromosome;The IS distance matrix is constructed according to the distance between each IS genomic site, and then the hierarchical clustering algorithm is used to cluster the distance matrix;After hierarchical clustering, the clusters of hierarchical clustering results are extracted according to a certain distance threshold, and the IS cluster after recognition is called UIS, and the representative site of UIS is determined by the read support number of all sites in the IS cluster.

[0019] As a preferred technical solution of the present application, the method of analyzing the clonal plane according to the clustering result and describing the dominant clone in step 4 is to calculate and evaluate the dominant clone of the sample from two dimensions, i.e. the diversity of the cell type of the independent integration site of the cell and the uniformity of the proportion of the independent integration site of the cell type, i.e. the clonal diversity and the clonal uniformity, by using the proportion of the read support number of the top 10 IS of the sample.

[0020] The present application has the following beneficial effects:

[0021] The detection method of the dominant clone caused by the exogenous DNA insertion mutation in the present application uses the PCR primer similarity ranking algorithm to identify and remove the non-specific amplification caused by the possible PCR primer similarity, which significantly reduces the false positive of IS identification and improves the accuracy in actual application. The present application realizes the detection of the dominant clone by evaluating the IS diversity and IS uniformity of the sample, thereby more reliably evaluating the tumor caused by the insertion mutation. The present application quantitatively evaluates the state of the cell disordered replication caused by the IS insertion by calculating the dominant clone plane, predicts the tumorigenicity of the insertion mutation, and increases the practicability of the IS detection in the clinical level. BRIEF DESCRIPTION OF DRAWINGS

[0022] The accompanying drawings are included to provide a further understanding of the present application, and constitute a part of the specification, which together with the embodiments of the present application, serve to explain the present application, and do not constitute a limitation of the present application.

[0023] In the drawings:

[0024] Figure 1 is a flowchart of the detection method of the dominant clone caused by the exogenous DNA insertion mutation in the present application;

[0025] Figure 2 is a sample diagram of the selection of the dominant clone caused by the exogenous DNA insertion mutation in embodiment 2 of the present application;

[0026] Figure 3 is a diversity and uniformity calculation result diagram of 9 samples of the detection method of the dominant clone caused by the exogenous DNA insertion mutation in embodiment 2 of the present application;

[0027] Figure 4 is a clonality prediction result diagram of the clonal plane of the detection method of the dominant clone caused by the exogenous DNA insertion mutation in embodiment 2 of the present application;

[0028] Figure 5 is a comparison diagram of the similarity of the primer sequence and the random site ranking distribution of the detection method of the dominant clone caused by the exogenous DNA insertion mutation in embodiment 2 of the present application;

[0029] Figure 6 FIG. 2 is a diagram of an IS sequence distribution curve of filtered samples of a method for detecting dominant clones caused by exogenous DNA insertion mutation in embodiment 2 of the present application;

[0030] Figure 7 FIG. 3 is a diagram of a clone plane calculation result of a method for detecting dominant clones caused by exogenous DNA insertion mutation in embodiment 2 of the present application;

[0031] Figure 8 FIG. 4 is a diagram of a top 10 IS proportion of 9 samples of a method for detecting dominant clones caused by exogenous DNA insertion mutation in embodiment 2 of the present application. DETAILED DESCRIPTION

[0032] The preferred embodiments of the present application are described below in conjunction with the accompanying drawings, and it should be understood that the preferred embodiments described here are only used to illustrate and explain the present application, and are not used to limit the present application.

[0033] Embodiment: As shown in the following, the method for detecting dominant clones caused by exogenous DNA insertion mutation in the present application comprises the following steps: Figure 1

[0034] Step 1, pre-processing of positive and negative strand correction and redundancy integration of virus IS information data;

[0035] Step 2, after pre-processing, identifying IS caused by non-specific amplification to remove false positive IS detection results caused by PCR non-specific amplification;

[0036] Step 3, merging and clustering IS in adjacent positions according to site clustering algorithm;

[0037] Step 4, analyzing the clone plane according to the clustering result and describing the dominant clones.

[0038] In the step 1, the virus IS information data is a set of virus IS, wherein each IS indicates the integration position on the host genome.

[0039] In the step 1, the specific operation of positive and negative strand correction of virus IS information data is that if the gene where the IS is located is a negative strand and the IS read is a negative strand, the IS is converted to a positive strand.

[0040] In the step 1, the method for redundancy integration of virus IS information data is that if the site and positive and negative strand information of the IS are completely consistent, it is merged into 1 record; the site and chromosome information corresponding to the IS are encoded to simplify it into 1 number as the basis of site clustering algorithm. ​

[0041] The method for identifying IS caused by nonspecific amplification in step 3 comprises three steps:

[0042] Step A: Extract all PCR primer sequences and calculate the reverse complementary sequence based on the primer sequence to form the primer sequence library S1; extract several base pairs before and after each IS according to its position to form the sequence library to be matched S2; on the target genome, randomly select a sufficient number of random sites to form the background sequence library S3;

[0043] Step B, matching and scoring, for each sequence in S2 and S3, match each primer sequence in the primer sequence library and its reverse complementary sequence respectively; in the matching results of the sequence and its reverse complementary sequence, select the highest matching score among all scores as the score of the similarity between the IS and the primer; after the matching is completed, each IS in S2 and S3 obtains a similarity score vector, and the vector length is equal to the number of primers; for each primer, the similarity scores corresponding to the primer in S2 and S3 are ranked, and the ranked scores are converted to percentiles, such as scores ranked in the top 1%, all converted to 1, and so on; after the conversion is completed, each IS corresponds to a percentile score vector, which indicates the ranking of the similarity between the sequence and the primer sequence among all sequences;

[0044] Step C: Determination of nonspecific amplification. If nonspecific amplification does not occur, the sequences in S2 and S3 are completely random, and the corresponding similarity ranking scores are theoretically evenly distributed. However, in the event of nonspecific amplification, the ranking distribution of the scores in S2 will be biased toward the top-ranked sites. By calculating the ranking vector of each IS in step 2, the probability that each IS in step 2 is nonspecifically amplifiable is obtained. The calculation method is to multiply the percentile score vectors of each IS. If the probability is less than a certain threshold T, the IS is marked as nonspecifically amplified. IS marked as specifically amplified will be removed from subsequent analyses.

[0045] Among them, in the step 3, the ISs at adjacent positions are merged according to the site clustering algorithm, and the clustering method is to group the ISs according to chromosomes, and the ISs in each chromosome; construct an IS distance matrix according to the distance between each IS genomic site, and then cluster the distance matrix using a hierarchical clustering algorithm; after the hierarchical clustering is completed, the IS cluster after extraction and identification of the hierarchical clustering result clusters according to a certain distance threshold is called UIS, and the representative site of UIS is determined by the site with the highest read support number among all the sites that make up the IS cluster.

[0046] Among them, the method of analyzing the cloning plane according to the clustering results in step 4 and describing the dominant clones is to use the proportion of the read segment support number of the top 10 IS in the sample to calculate and evaluate the dominant clones of the sample from two dimensions, namely the diversity of cell types at independent integration sites of the cells and the uniformity of the proportion of cell types at independent integration sites, that is, clonal diversity and clonal uniformity.

[0047] The diversity index is calculated by Simpson's Diversity Index.

[0048] D=1-Σ(ni / N)2

[0049] Where D is the Simpson index, n is the number of reads supporting each detected UIS, and N is the total number of UIS.

[0050] The calculation method of the uniformity index is derived from the Shannon index. The specific calculation method is as follows:

[0051] H=-Σ(pi*log2(pi))

[0052] Where H is the Shannon index, p is the proportion of each IS in the total IS, log2 represents the binary logarithm, ln represents the base of the binary logarithm, and S represents the total number of IS.

[0053] In a coordinate axis, diversity and uniformity are used as the x-axis and y-axis measures respectively, and the sample can be projected as a point on the coordinate axis. This method uses more than 210 samples in the local database to build a machine learning discriminant model based on support vector machine. The diversity index and uniformity index of the sample are used as independent variables, and clonality is used as the dependent variable. The support vector machine is used for training to find a dividing hyperplane in the sample space to separate samples with dominant clones from clone samples without dominant clones (such as Figure 4 After obtaining the dividing hyperplane, this method uses it to assess the dominant clonality of a new sample. This plane determines whether cells have developed dominant clones. The specific judgment rule is: if the sample is below the cloning plane, it is a dominant clone; otherwise, it is not. If the cloning plane results indicate that cells have not developed dominant clones, the sample is polyclonal; if the cloning plane results indicate that cells have developed dominant clones, the sample is oligoclonal.

[0054] The present application utilizes a PCR primer similarity ranking algorithm to identify and remove non-specific amplification caused by possible PCR primer similarity, significantly reducing false positives in IS identification and improving accuracy. The present application evaluates the IS diversity and IS uniformity of the sample to detect dominant clones, thereby providing a more reliable assessment of tumors caused by insertion mutations. This method quantitatively evaluates the state of cell disordered replication caused by IS insertion by calculating the dominant clone plane, predicts the tumorigenicity of insertion mutations, and increases the clinical utility of IS detection.

[0055] In Example 2, as shown in Figures 2 to 8 9 samples were selected, 4 with dominant clones and 5 without dominant clones, and then the top 10 IS of the 9 samples was detected using the method (the top 10 IS of the 9 samples, of which samples 2, 4, 5, and 6 had obvious dominant clones); the detection of the example samples showed the effect of false positive IS caused by non-specific amplification on the true results; and after processing by the method, the false positives were significantly reduced, indicating that the method has good performance in removing false positive IS caused by non-specific amplification. First, the method extracts all PCR primer sequences used in the experiment, calculates the reverse complement sequence of the primer sequence to form a primer sequence library S1. Then, the method extracts the 100 bp flanking sequences of each IS in 2 samples to form a sequence library S2 to be matched, which contains IS flanking sequences from two samples. At the same time, 1000 random sites are randomly selected on the target genome to form a random site background sequence library S3.

[0056] After obtaining the S1, S2, and S3 sequence libraries, the method matches each sequence in the sample sequence library (S2) and the random site sequence library (S3) to the 3 primer sequences and their respective reverse complements in S1. In the similarity scores a and b of each primer sequence and its reverse complement, M = max(a, b) is selected as the similarity score of the sequence and the primer, and finally the similarity scores M1, M2, and M3 of each primer are used as the final score vector, as shown in Table 2. After scoring, each sequence in S2 and S3 obtains a set of similarity score vectors. The similarity score vectors in S2 and S3 are sorted according to different primers, and the sorted scores are converted to percentiles, such as converting the top 1% score to 1 and the top 1%-2% score to 2, and so on. After conversion, each sequence in S2 and S3 corresponds to a ranking vector with a length of 3, indicating the ranking of the best similarity score of the sequence and the 3 primer sequences among all sequences.

[0057] Density distribution of IS scores in samples compared with scores of random sequences. As can be seen from the density distribution, the ranking distribution of the two samples is obviously biased towards the top, indicating that the sequences near the IS in the two samples have higher similarity to the PCR primers. As Figure 5 Comparison of the similarity of the primer sequence with the ranking distribution of random sites in two samples with obvious non-specific amplification leading to false positives (sample 1, sample 2.

[0058] One obvious feature of non-specific amplification is that it occurs multiple times in different samples, while IS randomly integrated in the human genome cannot appear in two samples. We found multiple ISs detected in both samples. After comparing the scores of such ISs with other ISs, we found that the density distribution of the scores of the ISs detected in both samples is more seriously biased towards the top, as Figure 6 As can be seen, in samples 1 and 2 with common detected ISs, the similarity to the primers is the highest, while after filtering by the method, the ranking distribution of the similarity of the ISs to the primers tends to be that of random sites. This shows that the method has a good effect on reducing false positives in IS identification and improving accuracy, which also indirectly shows that the correlation between the top score and non-specific amplification.

[0059] Figure 7 As shown, the results of the clonal plane calculation. The x-axis represents the diversity of the sample, and the y-axis represents the evenness of the sample. As can be seen, samples with dominant clonal phenomena are below the clonal plane, while samples without dominant clonal phenomena are above the clonal plane, indicating that the clonal plane defined by the method can accurately distinguish samples carrying dominant clones.

[0060] Finally, the method multiplies the ranking scores in S2, converts them into the probability that the IS is not a non-specific amplification sequence with the three primers, and then marks ISs with a probability higher than T as specific amplification sequences and removes them. After removing the ISs marked as non-specific amplification, the skewness of the overall score curve distribution decreases significantly, basically showing a uniform distribution, indicating that the method has a good recognition effect on non-specific amplification of ISs.

[0061] After obtaining the diversity and evenness indicators of each sample, the method establishes a coordinate axis with diversity and evenness as the x-axis and y-axis, respectively, and projects the diversity and evenness calculation results of each sample as a point on the x-axis and y-axis coordinates of the coordinate axis, respectively. Then draw the clonal plane determined by machine learning on the coordinate axis and observe the relative position of the sample to the clonal plane (as Figure 7 If the sample is below the clonal plane, it is a dominant clone, otherwise it is not a dominant clone. In Figure 7As can be seen, the samples 2, 4, 5 and 6 with dominant clonal phenomenon are below the clonal plane and are indicated by dots; and the samples 1, 3, 7, 8 and 9 without dominant clonal phenomenon are above the clonal plane and are indicated by triangles. The judgment results are shown in Table 5. The discrimination results show that the method can accurately judge the clonality of all samples.

[0062] In summary of the above examples, the method uses the PCR primer similarity ranking algorithm to identify and remove the non-specific amplification caused by possible PCR primer similarity, significantly reduces the false positive of IS identification in practical application, and improves the accuracy of IS identification. In addition, the method realizes the detection of dominant clones by evaluating the IS diversity and IS uniformity of the sample, thereby more reliably evaluating the tumor caused by insertion mutation, and increasing the practicality of IS detection in the clinical level.

[0063] Finally, it should be noted that: the above only for the preferred embodiments of the present application, and not for limiting the present application, although the foregoing embodiments of the present application are described in detail, for those skilled in the art, it still can be modified, or part of the technical features of the equivalent replacement. Any modification, equivalent replacement, improvement, etc. within the spirit and principles of the present application, should be included in the protection scope of the present application.

Claims

1. A method for detecting dominant clones based on mutations caused by exogenous DNA insertion, characterized in that: The following steps are involved: Step 1: Pre-process the virus IS information data by correcting positive and negative chains and integrating redundancy; Step 2: After pretreatment, IS caused by nonspecific amplification is identified to remove false positive IS test results caused by nonspecific PCR amplification; Step 3: Merge and cluster the ISs at adjacent locations according to the site clustering algorithm; Step 4: Analyze the clone plane based on the clustering results and describe the dominant clones; The method for identifying IS caused by nonspecific amplification in step 2 includes three steps: Step A: Extract all PCR primer sequences and calculate the reverse complementary sequence based on the primer sequence to form the primer sequence library S1; extract several base pairs before and after each IS according to its position to form the sequence library to be matched S2; on the target genome, randomly select a sufficient number of random sites to form the background sequence library S3; Step B, matching and scoring: For each sequence in S2 and S3, each primer sequence in the primer sequence library and its reverse complementary sequence are matched; in the matching results of the sequence and its reverse complementary sequence, the highest matching score among all scores is selected as the similarity score between the IS and the primer; after the matching is completed, each IS in S2 and S3 obtains a similarity score vector, and the vector length is equal to the number of primers; for each primer, the similarity scores corresponding to the primer in S2 and S3 are ranked, and the ranked scores are converted to percentiles; After the conversion is completed, each IS corresponds to a percentile score vector, which indicates the ranking of the similarity between the sequence and the primer sequence among all sequences; Step C: Determination of nonspecific amplification. If nonspecific amplification does not occur, the sequences in S2 and S3 are completely random, and the corresponding similarity ranking scores are theoretically evenly distributed. However, in the event of nonspecific amplification, the ranking distribution of the scores in S2 will be biased towards the top-ranked sites. By calculating the ranking vector of each IS in step B, the probability of each IS in step B being nonspecifically amplifiable is obtained. The calculation method is to multiply the percentile score vectors of each IS. If the probability is less than a certain threshold T, the IS is marked as nonspecific amplification; the IS marked as nonspecific amplification will be removed from subsequent analysis.

2. The method for detecting dominant clones based on exogenous DNA insertion mutation according to claim 1, characterized in that: The virus IS information data in step 1 is a collection of virus ISs, wherein each IS indicates the integration position on the host genome.

3. The method for detecting dominant clones based on exogenous DNA insertion mutation according to claim 1, characterized in that: The specific operation of performing positive and negative strand correction on the virus IS information data in step 1 is that if the gene where the IS is located is the negative strand and the IS read segment is the negative strand, the IS is converted to the positive strand.

4. The method for detecting dominant clones based on exogenous DNA insertion mutation according to claim 3, characterized in that: The method for redundantly integrating virus IS information data is as follows: if the IS site and positive and negative chain information are completely consistent, they are merged into one record; the site and chromosome information corresponding to the IS are encoded and simplified into one number as the basis of the site clustering algorithm.

5. The method for detecting dominant clones based on exogenous DNA insertion mutation according to claim 1, characterized in that: In the step 3, the ISs at adjacent positions are merged according to the site clustering algorithm, and the clustering method is to group the ISs according to chromosomes, and the ISs in each chromosome; construct an IS distance matrix according to the distance between each IS genomic site, and then cluster the distance matrix using a hierarchical clustering algorithm; after the hierarchical clustering is completed, the IS clusters obtained by extracting and identifying the hierarchical clustering results according to a certain distance threshold are called UIS, and the representative site of the UIS is determined by the site with the highest read support number among all the sites constituting the IS cluster.

6. The method for detecting dominant clones based on exogenous DNA insertion mutation according to claim 1, characterized in that: In step 4, the method for analyzing the clone plane according to the clustering results and describing the dominant clones is to use the proportion of the read segment support of the top 10 IS in the sample to calculate and evaluate the dominant clones of the sample from two dimensions, namely the diversity of cell types at independent integration sites of the cells and the uniformity of the proportion of cell types at independent integration sites, that is, clone diversity and clone uniformity.

Images