Method and system for detecting partial tandem repeat of KMT2A gene based on next-generation sequencing technology

Abstract

The invention discloses a method and system for detecting partial tandem repeat of a KMT2A gene based on a next-generation sequencing technology, and the method comprises the following steps: S1, after a to-be-detected sample is sequenced through an NGS experimental process, obtaining sequencing data in a fastq format, and carrying out QC and comparison on the sequencing data to obtain comparison data in a bam format; s2, performing depth calculation on the data, performing GC correction on the depth by using loess regression, and removing noise in a to-be-detected sample by using a PCA method and a pre-established normal sample baseline to obtain a final log-ratio; s3, segmenting the log-ratio by using a CBS (Content Based Service) method; and S4, fitting the data by using MCMC-based Gibbs sampling, and calculating whether PTD mutation exists in the KMT2A gene or not. The method is high in detection sensitivity, KMT2A-PTD can be detected in one-time panel data, and the variation state of diagnosis and treatment related genes of a to-be-detected sample is comprehensively evaluated by combining the advantages of high-throughput sequencing.

Description

Technical Field

[0001] The present invention belongs to the technical field of gene engineering, and in particular relates to a method and system for detecting partial tandem duplication of a KMT2A gene based on second-generation sequencing technology. Background Art

[0002] The MLL (mixed lineage leukaemia gene) gene is also called the lysine methyltransferase 2A (KMT2A) gene. It is located in the long arm of chromosome 11, region 2, band 3 (11q23), and encodes histone H3 lysine 4 methyltransferase. It plays an important role in embryonic development and hematopoiesis by regulating the expression of HOX genes. KMT2A gene rearrangement is the cause of some acute lymphoblastic leukemia and acute myeloid leukemia.

[0003] PTD (partial tandem duplication) is a common variant of KMT2A gene rearrangement, which refers to partial tandem duplication of the gene. KMT2A-PTD is characterized by a large number of internal duplications spanning 6-8 exons, and the most common duplication sequences include exons 3-9, exons 3-10, and exons 3-11.

[0004] KMT2A-PTD can be seen in about 10% of patients with acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS) with normal karyotype. The incidence rate is nearly 90% in AML patients with only trisomy 11 abnormalities, and it is a poor prognostic indicator for patients with AML and MDS. In the latest 2024 MDS NCCN guidelines, KMT2A-PTD is included in the IPSS-M (Molecular International Prognostic Scoring System) for MDS. KMT2A-PTD detection is conducive to accurately judging the prognosis of AML and MDS patients. For patients who are initially diagnosed with KMT2A-PTD and test positive, this gene variation can be used as a molecular residual lesion (MRD) monitoring indicator in the future.

[0005] At present, the mainstream method for clinical detection of KMT2A-PTD is qPCR, and the existing patent CN112266962A adopts this method. The qPCR method is simple and rapid, which greatly facilitates the clinical detection of KMT2A (MLL) gene. However, AML and MDS are highly heterogeneous and have diverse molecular changes. The qPCR technology based on single marker detection cannot fully evaluate the molecular variation of patients.

[0006] RNA-based NGS methods, such as the existing patent CN111863133A, have also been reported for detecting KMT2A-PTD. However, RNA is easily degraded and difficult to preserve. Therefore, DNA-based NGS detection of KMT2A-PTD has become a practical need. However, in most gene mutation detection projects based on DNA detection, KMT2A-PTD mutations are not included. The main reason is that the tandem duplication of KMT2A-PTD is a large-fragment tandem duplication involving multiple exons, so there are limitations in NGS detection. Summary of the invention

[0007] The purpose of the present invention is to provide a method and system for detecting partial tandem duplications of the KMT2A gene based on second-generation sequencing technology. The method of the present invention has high detection sensitivity and can detect KMT2A-PTD in a single panel data, combined with the advantages of high-throughput sequencing, to comprehensively evaluate the mutation status of diagnosis and treatment-related genes of the sample to be tested.

[0008] In order to achieve the above purpose, this application adopts the following technical solutions:

[0009] In a first aspect, the present invention provides a method for detecting partial tandem duplication of the KMT2A gene based on the second generation sequencing technology, comprising the following steps:

[0010] S1: After the samples to be tested are sequenced through the NGS experimental process, sequencing data in fastq format are obtained. The sequencing data are QC and aligned to obtain alignment data in bam format;

[0011] S2: Calculate the depth of the data, use loess regression to perform GC correction on the depth, and use the PCA method to remove the noise in the sample to be tested with a pre-established normal sample baseline to obtain the final log-ratio;

[0012] S3: Use the CBS method to segment the log-ratio so that each segment has a relatively consistent copy number state, assuming that the log-ratio conforms to the following distribution:

[0013]

[0014] Among them, r i is the log-ratio within the i segment, p is the tumor purity, C i The absolute copy number for this segment, C j The absolute copy number of segment j for this sample, l j For this sample j, the length of the segment, σ ri is the standard deviation of the log-ratio of the i segment;

[0015] S4: MCMC-based Gibbs sampling was used to fit the data and calculate whether there was a PTD mutation in the KMT2A gene;

[0016] If a segment is formed within the range of exon 1-15 of the KMT2A gene and the number of exons contained in this segment is ≥5, and its absolute copy number is ≥3.5, it is reported as positive;

[0017] Otherwise, the average log-ratio is calculated for the subdivided regions exon2-8, exon2-9, exon2-10, exon3-8, exon3-9 and exon3-10. If the following three points are met at the same time: 1) the average log-ratio ≥ 0.2; 2) the difference in log-ratio with the two previous and subsequent capture bin intervals is > 0.2; 3) at most one bin has a log-ratio < 0, it is reported as positive, otherwise it is reported as negative.

[0018] In the above technical solution, the specific method for calculating the depth of data in S2 is: obtain the total coverage of reads in bam in each target genomic region, and divide it by the width of each target genomic region to obtain the average depth of each region.

[0019] In the above technical scheme, the specific method of using the PCA method in S2 to remove the noise in the sample to be tested with a pre-established normal sample baseline is: pre-select some high-quality normal samples whose processing process is highly matched with the sample to be tested to build a baseline, and obtain the normal sample that is most similar to the sample to be tested in depth deviation from the normal sample baseline through PCA dimensionality reduction and finding the minimum distance, and remove the noise in the sample to be tested and standardize it based on it to obtain the final log-ratio.

[0020] In the above technical solution, the specific method of using the CBS method to segment the log-ratio in S3 is: for each possible segmentation point (the boundary of the copy number change), the variance ratio of the data before and after is calculated, and the segmentation point that makes the variance ratio the largest is selected as the optimal segmentation point. The data is divided into two parts at the selected segmentation point, and the above segmentation process is recursively repeated for each part until the stopping condition is met.

[0021] In the above technical scheme, the specific method of using MCMC-based Gibbs sampling to fit the data in S4 is: find multiple tumor purity and ploidy combinations that fit the log-ratio well through 2D grid search; under each tumor purity and ploidy combination condition, use Gibbs sampling to calculate the posterior probability of the copy number of each segment, determine the initial estimate of the copy number state for each segment through likelihood, and then assign copy numbers to all segments through simulated annealing algorithm, while optimizing the purity estimate.

[0022] In a second aspect, the present invention provides a system for detecting partial tandem duplications of the KMT2A gene based on the second generation sequencing technology, comprising a data comparison unit, a data depth calculation unit, a data processing unit and a data output unit;

[0023] The data comparison unit is used to perform QC and comparison on the sequencing data to obtain comparison data in bam format after the sequencing of the sample to be tested is completed through the NGS experimental process to obtain sequencing data in fastq format;

[0024] The data depth calculation unit is used to perform depth calculation on the data, perform GC correction on the depth using loess regression, and use the PCA method to remove noise in the sample to be tested with a pre-established normal sample baseline to obtain a final log-ratio;

[0025] The data processing unit is used to segment the log-ratio using the CBS method so that each segment has a relatively consistent copy number state, assuming that the log-ratio conforms to the following distribution:

[0026]

[0027] Among them, r i is the log-ratio within the i segment, p is the tumor purity, C i The absolute copy number for this segment, C j The absolute copy number of segment j for this sample, l j For this sample j, the length of the segment, σ ri is the standard deviation of the log-ratio of the i segment;

[0028] The data output unit is used to fit the data using MCMC-based Gibbs sampling, calculate whether the KMT2A gene has a PTD mutation, and report the result;

[0029] If a segment is formed within the range of exon 1-15 of the KMT2A gene and the number of exons contained in this segment is ≥5, and its absolute copy number is ≥3.5, it is reported as positive;

[0030] Otherwise, the average log-ratio is calculated for the subdivided regions exon2-8, exon2-9, exon2-10, exon3-8, exon3-9 and exon3-10. If the following three points are met at the same time: 1) the average log-ratio ≥ 0.2; 2) the difference in log-ratio with the two previous and subsequent capture bin intervals is > 0.2; 3) at most one bin has a log-ratio < 0, it is reported as positive, otherwise it is reported as negative.

[0031] In the above technical solution, in the data depth calculation unit, the specific method for performing depth calculation on the data is: obtaining the total coverage of reads in bam in each target genome region, and dividing it by the width of each target genome region to obtain the average depth of each region.

[0032] In the above technical scheme, in the data depth calculation unit, the specific method of using the PCA method to remove the noise in the sample to be tested with a pre-established normal sample baseline is: pre-select some high-quality normal samples whose processing process is highly matched with the sample to be tested to build a baseline, and obtain the normal sample that is most similar to the sample to be tested in depth deviation from the normal sample baseline through PCA dimensionality reduction and finding the minimum distance, and remove the noise in the sample to be tested based on it and standardize it to obtain the final log-ratio.

[0033] In the above technical solution, in the data processing unit, the specific method of using the CBS method to segment the log-ratio is: for each possible segmentation point (the boundary of the copy number change), the variance ratio of the data before and after is calculated, and the segmentation point that makes the variance ratio the largest is selected as the optimal segmentation point. The data is divided into two parts at the selected segmentation point, and the above segmentation process is recursively repeated for each part until the stopping condition is met.

[0034] In the above technical solution, in the data output unit, the specific method of fitting the data using MCMC-based Gibbs sampling is: finding multiple tumor purity and ploidy combinations that fit the log-ratio well through 2D grid search; under each tumor purity and ploidy combination condition, using Gibbs sampling to calculate the posterior probability of the copy number of each segment, determining the initial estimate of the copy number state for each segment through likelihood, and then assigning copy numbers to all segments through a simulated annealing algorithm, while optimizing the purity estimate.

[0035] The beneficial effects of the present invention are:

[0036] This method is the first to detect KMT2A-PTD at the DNA level using NGS detection technology. Compared with the traditional qPCR-based method, it can not only detect KMT2A-PTD in a single panel data, but also combine the advantages of high-throughput sequencing to comprehensively evaluate the mutation status of the diagnosis and treatment-related genes of the sample to be tested.

[0037] In addition, the dilution data of clinical samples and cell lines showed that the method of the present invention has higher sensitivity and higher consistency compared with the RNA method. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 This is the MLL gene signal of sample 180634787B4 which was negative by NGS;

[0039] Figure 2 This is the MLL gene signal status of sample 180614383B4 which was positive by NGS detection. DETAILED DESCRIPTION

[0040] In order to better illustrate the purpose, technical solutions and advantages of the present invention, the present invention will be further described below in conjunction with specific embodiments. The present invention can be implemented in many different forms and should not be construed as being limited to the embodiments set forth herein. On the contrary, these embodiments are provided so that this disclosure will be thorough and complete, and the concept of the present invention will be fully conveyed to those skilled in the art, and the present invention will only be limited by the claims.

[0041] This method is based on DNA samples and uses a high-throughput sequencing NGS platform to achieve qualitative detection of KMT2A-PTD. The specific method is as follows:

[0042] After the samples to be tested are sequenced through the NGS experimental process, sequencing data in fastq format are obtained. The sequencing data are QC and aligned to obtain alignment data in bam format;

[0043] Obtain the total coverage of reads in each target genomic region in bam, and divide it by the width of each target genomic region to obtain the average depth of each region;

[0044] Depth was corrected for GC using loess regression;

[0045] Preselect some high-quality normal samples whose processing process is highly matched with the sample to be tested to build a baseline. Through PCA dimensionality reduction and finding the minimum distance, obtain the normal sample that is most similar to the sample to be tested in terms of depth deviation from the normal sample baseline. Based on it, remove the noise in the sample to be tested and standardize it to obtain the final log-ratio;

[0046] Use the CBS method to segment the log-ratio so that each segment has a relatively consistent copy number state: for each possible split point (the boundary of the copy number change), calculate the variance ratio of the data before and after, select the split point that maximizes the variance ratio as the best split point, divide the data into two parts at the selected split point, and recursively repeat the above splitting process for each part until the stopping condition is met;

[0047] Assume that the log-ratio follows the following distribution:

[0048]

[0049] Among them, r i is the log-ratio within the i segment, p is the tumor purity, C i The absolute copy number for this segment, Cj The absolute copy number of segment j for this sample, l j For this sample j, the length of the segment, σ ri is the standard deviation of the log-ratio of the i segment;

[0050] Through 2D grid search, multiple tumor purity and ploidy combinations with good log-ratio fit were found; under each tumor purity and ploidy combination, Gibbs sampling was used to calculate the posterior probability of the copy number of each segment, and the initial estimate of the copy number state was determined for each segment through likelihood, and then the copy number was assigned to all segments through the simulated annealing algorithm, while optimizing the purity estimate;

[0051] Calculate whether the KMT2A gene has a PTD mutation;

[0052] If a segment is formed within the range of exon 1-15 of the KMT2A gene and the number of exons contained in this segment is ≥5, and its absolute copy number is ≥3.5, it is reported as positive;

[0053] Otherwise, the average log-ratio is calculated for the subdivided regions exon2-8, exon2-9, exon2-10, exon3-8, exon3-9 and exon3-10. If the following three points are met at the same time: 1) the average log-ratio ≥ 0.2; 2) the difference in log-ratio with the two previous and subsequent capture bin intervals is > 0.2; 3) at most one bin has a log-ratio < 0, it is reported as positive, otherwise it is reported as negative.

[0054] Experimental procedures

[0055] The above steps were used to analyze 32 clinical samples, which were confirmed as KMT2A-PTD positive samples based on the qPCR platform, and 19 samples had qPCR quantitative results. The test results showed that 30 of them were detected as positive, with a sensitivity of 93.75%. The qPCR quantitative results of the 2 samples with negative test results were very low (in Table 1, - represents that the corresponding samples have no qPCR quantitative results).

[0056] Table 1 Comparison of qPCR quantification and NGS detection results

[0057]

[0058]

[0059] Example of negative sample: For sample 180634787B4, when the CBS method was used to divide the log-ratio, chromosome 11 where KMT2A was located was divided into an integral segment. Through Gibbs sampling, the copy number of the segment was calculated to be 2 (normal copy number), and the log-ratio of exon 2-8, 2-9, 2-10, 3-8, 3-9, and 3-10 of KMT2A was not significantly higher than the two capture bin intervals before and after (cut-off>=0.2), so it was determined to be KMT2A-PTD negative. Figure 1 shown.

[0060] Example of positive sample: For sample 180614383B4, when the CBS method was used to divide the log-ratio, exon 2-10 of KMT2A was divided into independent segments. The copy number of this segment was calculated by Gibbs sampling to be 4, which was higher than the threshold of 3.5, so it was determined to be KMT2A-PTD positive. Figure 2 shown.

[0061] It can be seen from this that the method of the present invention has a high sensitivity, and the qPCR result itself is very low and cannot be detected.

[0062] The DNA-based detection method of the present invention was compared with the traditional RNA-based detection method. The specific process was as follows: RNA targeted capture sequencing was performed on 6 samples with qPCR quantitative results and RNA based on the NGS platform, and KMT2A-PTD analysis was performed on the detection data based on the fusion detection technology. The results showed that the RNA samples with low quantification on the 2 qPCR platforms did not capture KMT2A-PTD related signals.

[0063] Among the 6 RNA samples, 4 of them also had DNA samples, and the DNA and RNA test results of these 4 samples were consistent. In Table 2, NA means that the corresponding experiment was not performed.

[0064] Table 2 Comparison of DNA and RNA in 6 samples

[0065] Sample No. MLL-PTD quantitative results DNA test results RNA test results 180634779 24.59 Positive Positive 180634780 11.38 Positive Positive 180634789 2.99 Negative Negative 180643434 1.87 NA Negative 180643435 49.19 NA Positive 180634794 38.32 Positive Positive

[0066] In addition, 9 clinical KMT2A-PTD positive samples were diluted to 20% and 30% tumor purity, a total of 18 diluted samples, and the samples were tested according to the above method. The results showed that 9 samples with 30% purity were qualitatively positive for PTD. Among the 9 samples with 20% purity, 6 samples were qualitatively positive for PTD (6 / 9). Three samples were randomly selected and diluted to 10%, and the test results were all negative.

[0067] Furthermore, a KMT2A-PTD-positive cell line EOL-1 was diluted to 10%, 20%, and 30% cell purity. The results showed that the sample with 10% purity was negative, and the samples with 20% and 30% purity were positive.

[0068] The dilution data of clinical samples and cell lines showed that the method of the present invention has higher sensitivity and higher consistency compared with the RNA method.

[0069] Obviously, the above embodiments are merely examples for the purpose of clear explanation, and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived therefrom are still within the scope of protection of the invention.

Claims

1. A method for detecting partial tandem duplication of KMT2A gene based on second generation sequencing technology, characterized in that: The following steps are involved: S1: After the samples to be tested are sequenced through the NGS experimental process, sequencing data in fastq format are obtained. The sequencing data are QC and aligned to obtain alignment data in bam format; S2: Calculate the depth of the data, use loess regression to perform GC correction on the depth, and use the PCA method to remove the noise in the sample to be tested with a pre-established normal sample baseline to obtain the final log-ratio; S3: Use the CBS method to segment the log-ratio so that each segment has a relatively consistent copy number state, assuming that the log-ratio conforms to the following distribution: Among them, r i is the log-ratio within the i segment, p is the tumor purity, C i The absolute copy number for this segment, C j The absolute copy number of segment j for this sample, l j For this sample j, the length of the segment, σ ri is the standard deviation of the log-ratio of the i segment; S4: MCMC-based Gibbs sampling was used to fit the data and calculate whether there was a PTD mutation in the KMT2A gene; If a segment is formed within the range of exon 1-15 of the KMT2A gene and the number of exons contained in this segment is ≥5, and its absolute copy number is ≥3.5, it is reported as positive; Otherwise, the average log-ratio is calculated for the subdivided regions exon2-8, exon2-9, exon2-10, exon3-8, exon3-9 and exon3-10. If the following three points are met at the same time: 1) the average log-ratio ≥ 0.2; 2) the difference in log-ratio with the two previous and subsequent capture bin intervals is > 0.2; 3) at most one bin has a log-ratio < 0, it is reported as positive, otherwise it is reported as negative.

2. The method according to claim 1, characterized in that: The specific method for calculating the depth of data in S2 is: obtain the total coverage of reads in bam in each target genomic region, and divide it by the width of each target genomic region to obtain the average depth of each region.

3. The method according to claim 1, characterized in that: The specific method of using the PCA method in S2 to remove the noise in the sample to be tested with a pre-established normal sample baseline is as follows: pre-select some high-quality normal samples whose processing process is highly matched with the sample to be tested to build a baseline, and obtain the normal sample that is most similar to the sample to be tested in terms of depth deviation from the normal sample baseline through PCA dimensionality reduction and finding the minimum distance. Based on it, the noise in the sample to be tested is removed and standardized to obtain the final log-ratio.

4. The method according to claim 1, characterized in that: The specific method of using the CBS method to segment the log-ratio in S3 is: for each possible segmentation point, calculate the variance ratio of the data before and after, select the segmentation point that makes the variance ratio the largest as the best segmentation point, divide the data into two parts at the selected segmentation point, and recursively repeat the above segmentation process for each part until the stopping condition is met.

5. The method according to claim 1, characterized in that: The specific method of using MCMC-based Gibbs sampling to fit the data in S4 is: find multiple tumor purity and ploidy combinations that fit the log-ratio well through 2D grid search; use Gibbs sampling to calculate the posterior probability of the copy number of each segment under each tumor purity and ploidy combination, determine the initial estimate of the copy number state for each segment through likelihood, and then assign copy numbers to all segments through simulated annealing algorithm while optimizing the purity estimate.

6. A system for detecting partial tandem duplications of the KMT2A gene based on the second generation sequencing technology, characterized in that: It includes a data comparison unit, a data depth calculation unit, a data processing unit and a data output unit; The data comparison unit is used to perform QC and comparison on the sequencing data to obtain comparison data in bam format after the sequencing of the sample to be tested is completed through the NGS experimental process to obtain sequencing data in fastq format; The data depth calculation unit is used to perform depth calculation on the data, perform GC correction on the depth using loess regression, and use the PCA method to remove noise in the sample to be tested with a pre-established normal sample baseline to obtain a final log-ratio; The data processing unit is used to segment the log-ratio using the CBS method so that each segment has a relatively consistent copy number state, assuming that the log-ratio conforms to the following distribution: Among them, r i is the log-ratio within the i segment, p is the tumor purity, C i The absolute copy number for this segment, C j The absolute copy number of segment j for this sample, l j For this sample j, the length of the segment, σ ri is the standard deviation of the log-ratio of the i segment; The data output unit is used to fit the data using MCMC-based Gibbs sampling, calculate whether the KMT2A gene has a PTD mutation, and report the result; If a segment is formed within the range of exon 1-15 of the KMT2A gene and the number of exons contained in this segment is ≥5, and its absolute copy number is ≥3.5, it is reported as positive; Otherwise, the average log-ratio is calculated for the subdivided regions exon2-8, exon2-9, exon2-10, exon3-8, exon3-9 and exon3-10. If the following three points are met at the same time: 1) the average log-ratio ≥ 0.2; 2) the difference in log-ratio with the two previous and subsequent capture bin intervals is > 0.2; 3) at most one bin has a log-ratio < 0, it is reported as positive, otherwise it is reported as negative.

7. The system according to claim 1, characterized in that: In the data depth calculation unit, the specific method for performing depth calculation on the data is: obtaining the total coverage of reads in bam in each target genome region, and dividing it by the width of each target genome region to obtain the average depth of each region.

8. The system according to claim 1, characterized in that: In the data depth calculation unit, the specific method of using the PCA method to remove the noise in the sample to be tested with a pre-established normal sample baseline is: pre-select some high-quality normal samples whose processing process is highly matched with the sample to be tested to construct a baseline, and obtain the normal sample that is most similar to the sample to be tested in depth deviation from the normal sample baseline through PCA dimensionality reduction and finding the minimum distance, and remove the noise in the sample to be tested and standardize it based on it to obtain the final log-ratio.

9. The system according to claim 1, characterized in that: In the data processing unit, the specific method of using the CBS method to segment the log-ratio is as follows: for each possible segmentation point, the variance ratio of the data before and after is calculated, and the segmentation point that makes the variance ratio the largest is selected as the optimal segmentation point. The data is divided into two parts at the selected segmentation point, and the above segmentation process is recursively repeated for each part until the stopping condition is met.

10. The system according to claim 1, characterized in that: In the data output unit, the specific method of fitting the data using MCMC-based Gibbs sampling is as follows: multiple tumor purity and ploidy combinations that fit the log-ratio well are found through 2D grid search; under the conditions of each tumor purity and ploidy combination, the posterior probability of the copy number of each segment is calculated using Gibbs sampling, and the initial estimate of the copy number state is determined for each segment through likelihood, and then the copy number is assigned to all segments through a simulated annealing algorithm, while optimizing the purity estimate.

Images