IKZF1 gene exon deletion identification system and method based on high-throughput transcriptome sequencing

Abstract

The present invention discloses an IKZF1 gene exon deletion identification system and method based on high-throughput transcriptome sequencing; the exon deletion identification system is predicted by a random forest model constructed by taking the number of IKZF1 exon connections, differential genes and their expression levels, and IKZF1 exon and intron mutation frequency information of a sample as feature inputs, and can be used to accurately predict and identify any exon deletion with high accuracy and strong specificity, thus providing guidance for patient prognosis and treatment.

Description

Technical Field

[0001] The present invention belongs to the technical field of gene detection, and in particular relates to an IKZF1 gene exon deletion identification system and method based on high-throughput transcriptome sequencing. Background Art

[0002] The IKZF1 gene (IKAROS family IKAROS1) is located on chromosome 7 p12 and has a total of 8 exons. The IKROS protein it encodes plays an important role in the development and function of the hematopoietic system. Local deletion of the IKROS family zinc finger protein 1 (IKZF1) can lead to the loss of the zinc finger DNA binding domain and abnormal function, thereby affecting the normal development and maturation of cells. ALL (acute lymphoblastic leukemia, ALL) is the most common malignant tumor in children, accounting for 26% of childhood cancers. IKZF1 exon deletion is common in ALL, with an incidence of about 15% in children with B-ALL and 30% to 40% in adult patients. IKZF1 gene deletion or sequence mutation is detected in more than 70% of Ph chromosome-positive ALL. IKZF1 exon deletion is closely related to the pathogenesis of high-risk B-ALL, worse treatment prognosis, increased risk of relapse, and whether the IKAROS transcription factor can play a normal role.

[0003] The main types of IKZF1 exon deletions are: 4-7 deletion, 1-8 deletion, 2-7 deletion, 2-3 deletion, 4-8 deletion, 2-8 deletion, 1-3 deletion, 1 deletion, etc. Among all deletion types, 4-7 deletion and 1-8 deletion are the most common, with a proportion of about 30%, followed by 4-8 deletion and 2-7 deletion, and other types of deletions are relatively rare.

[0004] At present, the main methods for detecting IKZF1 exon deletion are chip technology, reverse transcription polymerase chain reaction (RT-PCR) method and multiplex ligation-dependent probe amplification (MLPA). Gene chip technology and reverse transcription polymerase chain reaction (RT-PCR) detection method have the disadvantages of low detection sensitivity, difficulty in quantitative analysis and difficulty in identifying rare IKZF1 mutant transcripts. Multiplex ligation-dependent probe amplification (MLPA) is a qualitative and semi-quantitative analysis technology for DNA sequences to be tested that has been developed in recent years. In the MLPA reaction, two oligonucleotide fragments are hybridized with the target sequence and then the probes are connected with ligase. Since the ligation reaction is highly specific, the ligase can only connect the two probes into a complete nucleic acid single strand when the two probes are completely hybridized with the target sequence. If the target sequence and the probe sequence are not completely complementary, the ligation reaction will not be able to proceed.

[0005] Transcriptome sequencing (RNA-seq) uses high-throughput gene sequencing technology to sequence all gene transcription products in a specimen, and can obtain the gene sequence of all transcripts. RNA-seq data contains rich information such as gene expression, gene fusion, splicing isoforms and gene mutations. It is an effective method to identify genomic damage in patients with diseases and provides key information about patient prognosis and treatment. Summary of the invention

[0006] The purpose of the present invention is to propose an IKZF1 gene exon deletion identification system based on high-throughput transcriptome sequencing, which is used to accurately identify the KZF1 exon deletion type of a sample to be tested and provide guidance for the treatment and prognosis of patients.

[0007] In view of this, the scheme of the present invention is as follows:

[0008] The first aspect of the present invention provides an IKZF1 gene exon deletion identification system based on high-throughput transcriptome sequencing, comprising:

[0009] The characteristic data analysis module is used to analyze the characteristic data of the sample to be tested, including: the characteristic data of the IKZF1 gene in the transcriptome data of the sample to be tested, and the differential gene expression between the sample to be tested and the negative sample;

[0010] An exon deletion recognition module is used to input the characteristic data of the sample to be tested into a deletion recognition model to obtain an IKZF1 gene exon deletion recognition result; the deletion recognition model is trained using a random forest model; the training process is based on the IKZF1 gene characteristic data of each training sample transcriptome data and the differential gene expression amount from the negative sample as input features;

[0011] The IKZF1 gene characteristic data includes the number of exon connections, the mutation frequency of exons, and the mutation frequency of introns; the negative sample is a sample without any deletion type of the IKZF1 gene; the training sample includes positive samples of all deletion types, and each positive sample has one deletion type.

[0012] Furthermore, the types of exon deletions include 4-7 deletion, 1-8 deletion, 2-7 deletion, 2-3 deletion, 4-8 deletion, 2-8 deletion, 1-3 deletion, and 1 deletion.

[0013] Furthermore, the test sample and / or training sample is preprocessed before use, including filtering low-quality reads, removing adapter sequences and redundant bases.

[0014] Furthermore, the mutation frequency of the exon is the exon heterozygous mutation frequency.

[0015] Furthermore, the differential gene expression levels were analyzed using DESeq2 software to find differentially expressed genes and perform differential analysis, the results of which included gene IDs, standardized gene expression values, and log2-converted differential fold values.

[0016] Furthermore, the number of exon connections is calculated using featureCounts software based on a defined gene annotation file to obtain the number of each exon connection in a gene.

[0017] Furthermore, the intron mutation frequency is subjected to an intersection operation on the defined IKZF1 gene intron position file, and a file of the number of heterozygous mutations occurring in the intron is obtained by analysis.

[0018] The second aspect of the present invention is to propose a method for identifying exon deletion of the IKZF1 gene based on high-throughput transcriptome sequencing for non-diagnostic purposes, comprising:

[0019] Analyze the characteristic data of the samples to be tested: including the characteristic data of the IKZF1 gene in the transcriptome data of the samples to be tested, and the differential gene expression between the samples to be tested and the negative samples;

[0020] Inputting the characteristic data of the sample to be tested into the deletion recognition model to obtain the IKZF1 gene exon deletion recognition result; the deletion recognition model is trained by a random forest model; the training process is based on the IKZF1 gene characteristic data of each training sample and the differential gene expression amount with the negative sample as input features;

[0021] The IKZF1 gene characteristic data includes the number of exon connections, the mutation frequency of exons, and the mutation frequency of introns; the negative sample is a sample without any deletion type of the IKZF1 gene; the training sample includes positive samples of all deletion types, and each positive sample has one deletion type.

[0022] The non-diagnostic purposes include predicting samples that cannot be traced back to humans, predicting the type of deletions in the samples themselves, including studying the pathogenesis of high-risk B-ALL, determining whether the IKAROS transcription factor can function normally, and other biological functions.

[0023] The third aspect of the present invention is to provide an electronic device, comprising a processor and a memory, wherein the memory stores a computer program, and when the processor executes the computer program, the method described in the second aspect is implemented.

[0024] A fourth aspect of the present invention is to provide a computer-readable storage medium having a computer program stored thereon, and a processor executes the computer program to implement the method described in the second aspect.

[0025] Compared with the prior art, the beneficial effects of the present invention include but are not limited to:

[0026] The IKZF1 gene exon deletion identification system provided by the present invention is constructed through the IKZF1 exon connection number, differential genes and their expression levels, IKZF1 exon and intron mutation frequency information of the sample and a random forest model, and can be used to accurately predict and identify any exon deletion with high accuracy and strong specificity, providing guidance for patient prognosis and treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0028] Figure 1 The present invention is a flowchart for constructing an IKZF1 exon deletion recognition model.

[0029] Figure 2 This is a volcano plot of the differentially expressed genes analyzed during the model construction process of the present invention.

[0030] Figure 3 This is a heat map of differentially expressed genes analyzed during the model construction process of the embodiment of the present invention. DETAILED DESCRIPTION

[0031] The following provides definitions of some terms used in this specification. Unless otherwise specified, all terms used herein are generally understood by those skilled in the art to which this solution belongs.

[0032] Terminology explanation:

[0033] Random Forest: Random Forest is an ensemble learning method mainly used for classification and regression tasks. Multiple sub-datasets are generated from the original data set using the method of sampling with replacement. Each sub-dataset is used to train a decision tree. At each node, a portion of features is randomly selected for splitting until a specified number of decision trees are generated. Finally, the prediction results of all decision trees are used to determine the final classification result by majority voting. The model builds multiple decision trees and combines their prediction results to improve the accuracy and robustness of the model.

[0034] BAM: It is a commonly used binary format file, mainly used for the storage of large-scale sequencing genome sequence data, especially sequencing data aligned to the reference genome. It consists of two parts: the header and the alignment record, which includes information about the reference sequence, sequencing platform, alignment algorithm, the location, direction, base quality, and other information of sequencing reads. The binary writing method gives it many advantages, such as more compact data storage, saving storage space, and easier reading and writing. The BAM format is supported in a variety of bioinformatics tools and analysis software, and can perform a variety of operations such as data processing, filtering, annotation, etc., which facilitates data exchange and sharing. It has been widely used in genomic research, personalized medicine, population genetics and other fields.

[0035] genomic.rename.gff: A database that records information such as chromosomes and their starting positions, gene names, and regions.

[0036] The inventors discovered several features closely related to IKZF1 exon deletions, and used these features to predict the type of IKZF1 exon deletions through a random forest modeling method. The construction of the IKZF1 gene deletion recognition model based on high-throughput transcriptome sequencing mainly includes three parts: data preprocessing, finding important features, and modeling. Figure 1 shown.

[0037] 1. Data preprocessing

[0038] Fastp was used to perform quality control on the raw sequencing data, including filtering low-quality reads, removing adapter sequences and redundant bases, etc. to improve the accuracy of subsequent analysis. STAR software was used to align the data reads to the specified hg19 reference genome and then SAMBAMBA was used to remove duplicates and sort to obtain a bam file that stores the genome alignment information.

[0039] 2. Important feature extraction

[0040] (1) Number of exon junctions

[0041] Transcriptome sequencing was performed using 145 samples with known IKZF1 exon deletion types and 50 patient samples without IKZF1 exon deletion. By using featureCounts software to analyze the BAM file, the interconnection patterns between the exons in the IKZF1 gene and the specific number of these connections can be obtained in detail. The specific steps include:

[0042] First, prepare a high-quality BAM file, which is obtained by alignment of high-throughput transcriptome sequencing data. Next, use the BAM file as input and perform quantitative analysis using featureCounts software. The software can calculate the number of connections for each exon in the gene based on the predefined gene annotation file (genomic.rename.gff), and find that there are different exon connection patterns and connection numbers in different sample deletion types. Since the total number of reads of the IKZF1 gene in different samples is different, it is important to normalize the number of connections in this step to eliminate the influence of the number of reads on the statistics of the number of exon connections. These data are of great significance for understanding the deletion type and expression regulation of the gene, and are helpful for in-depth exploration of the function of the IKZF1 gene in different biological processes and its potential disease association. In this way, a more comprehensive map of exon connections can be depicted, laying the foundation for the subsequent prediction of IKZF1 exon deletion types.

[0043] (2) Differentially expressed genes

[0044] The most direct purpose of transcriptome sequencing is to try to find genes with significant expression differences between groups and explain the impact of changes in gene expression levels on biological functions. The DESeq2 package is an R language package designed for high-dimensional count data in transcriptome data, focusing on data normalization, visualization, and differential expression analysis. Its main function is to use the empirical Bayes method to estimate the prior values ​​of logarithmic fold change and deviation, and calculate the posterior values ​​of these statistics. It is currently one of the most frequently used R packages for identifying differential molecules. DESeq2 software was used to find the differentially expressed genes corresponding to each deletion type for samples of each deletion type and samples of ALL patients without IKZF1 exon deletion (no IKZF1 gene exon deletion). The differential analysis results included gene ID, standardized gene expression value, log2-transformed difference fold (log2FoldChange) value, significant p value, and corrected p value (default FDR correction) and other main information. The screening was performed according to the conditions of |log2FoldChange|≥1 and padj<0.01, and up-regulated and down-regulated genes were distinguished by "up" and "down". After DESeq2 identified the significantly differentially expressed genes, they can be represented by volcano plots through the R package (ggplot2). Volcano plots are one of the common statistical charts (such as Figure 2 ).

[0045] After analysis, it was found that for each deletion type, the differential gene expression levels were different from those of samples without IKZF1 deletion.

[0046] (3) Mutation frequency of IKZF1 exons

[0047] First, the bcftools mpileup command in the bcftools software was used to convert the BAM files of the samples into standardized variant annotation files (vcf format). Next, the bedtools software was used to perform an intersection operation on these vcf files and the predefined IKZF1 gene exon position files to obtain the mutation frequency and specific position information in each exon. When defining the mutation type, we defined the variants with an allele frequency between 20% and 80% as heterozygous mutants, the sites with an allele frequency below 20% as homozygous wild types, and the sites with a gene frequency greater than 80% as homozygous mutants. With this definition, the number of heterozygous mutation types occurring in the IKZF1 exon of each sample can be accurately counted. Since heterozygous deletion is the most common type and the frequency of pure and missing types is less than 1%, the heterozygous deletion type is mainly discussed. These results provide important data support for subsequent molecular mechanism studies.

[0048] (4) Intron mutation frequency

[0049] The statistical method of intron mutation frequency is similar to that of exon mutation frequency, except that the bedtools software needs to perform an intersection operation with the pre-defined IKZF1 gene intron position file. The final analysis obtains the number file of heterozygous mutations occurring in introns.

[0050] 3. Random Forest Model Construction

[0051] (5) The feature vectors in the sample are organized into a specific model input format as the input of the decision tree, and the feature vectors include: the number of exon connections, the expression level of differentially expressed genes, the mutation frequency of IKZF1 exons, and the mutation frequency of introns. The sample is divided into a training set and a test set in a ratio of 7:3. The RandomizedSearchCV model tuning tool is used in the model to help find the best model parameter combination, and the best parameter combination and its corresponding cross-validation score are output. Finally, the best model is used to evaluate on the test set and the accuracy of the test set is output.

[0052] Example

[0053] A total of 145 samples with deletions (each sample had only one deletion, and each deletion type had multiple samples) and 50 samples without deletions were collected for high-throughput transcriptome sequencing data, and the following operations were performed:

[0054] (1) Sequencing data quality control (taking test samples as an example)

[0055] All data were analyzed using fastp and quality control to obtain the following table:

[0056] Sample Raw_reads Raw_bases Clean_reads Q20(％) Q30(％) Clean_bases Clean_rate(%) test_read1 53662631 8103057281 52670725 98.3 94.92 7625487814 94.11 test_read2 53662631 8103057281 52670725 97.24 91.72 7377764375 91.05

[0057] (2) Comparison of STAR with reference genome

[0058] The sequencing data is aligned to the human reference genome hg19 as follows:

[0059] Sample Mapping % rRNA_Mapping% test 90.28 4.18

[0060] (3) Feature extraction

[0061] Feature 1: Exon connection number. Use featureCounts to count the number of different IKZF1 exon connection patterns in the sample. The results are shown in the following table (part of the samples):

[0062]

[0063] Feature 2: Differentially expressed genes. Use DESeq2 to find differentially expressed genes between samples with specific exon deletion types and samples without exon deletion. Taking the IKZF1 exon 4-7 deletion type as an example, up-regulated and down-regulated genes were obtained. The heat map of the differential gene expression in the two groups of samples is shown below: Figure 3 shown.

[0064] Feature 3: IKZF1 exon mutation frequency. The heterozygous mutations of exons are counted as follows (part of the samples):

[0065] Sample No. Missing Type Number of heterozygous mutations Exon Number of heterozygous mutations Exon Sample_control_3 No missing 0 - 0 - Sample_control_22 No missing 0 - 0 - Sample_control_37 No missing 0 - 0 - Sample48 4-7del 4 exon8 0 - Sample49 4-7del 8 exon8 1 exon1 Sample34 4-7del 11 exon8 0 -

[0066] Feature 4: IKZF1 intron mutation frequency. The statistics of heterozygous intron mutations are as follows (partial samples):

[0067]

[0068] (4) Model construction

[0069] The two groups of samples collected above, 145 with IKZF1 deletion and 50 without deletion, were subjected to data quality control, genome alignment and feature extraction to obtain data containing the characteristic information of the two groups of samples, in which the characteristic information was used as the characteristic value, the IKZF1 exon deletion type of the sample was used as the target value, and the sorted characteristic data was used as the input of the random forest model. A random forest classifier containing 1000 decision trees was created by the Random Forest algorithm, and the random seed was set to 0 to ensure the reproducibility of the results. The sklearn.metrics module was used to evaluate the quality of the model prediction results, and the evaluation indicators included precision, recall rate, F1-score, classification report, etc. The metrics.accuracy_score function in the module was used to compare the labels predicted by the model with the true labels and give the overall accuracy of the classifier on a given test data set. Finally, the classifier obtained by tuning the RandomizedSearchCV model had an accuracy of 98.21% and was saved.

[0070] (5) Model testing

[0071] 20 samples of IKZF1 deletion type to be tested were collected and high-throughput transcriptome sequencing was performed on these 20 samples. Then the model was used to identify and predict the exact type according to the following steps:

[0072] 1. Perform quality control on the obtained data using fastp, and align the hg19 human reference genome with STAR to obtain a bam file.

[0073] 2. Use featureCounts software to analyze the BAM file to obtain a file containing 13 IKZF1 exon patterns and connection number information.

[0074] 3. Use featureCounts to analyze the BAM file, and then after normalization, the expression levels (read counts) of all genes can be calculated. To be consistent with the characteristic values ​​of the model, for each sample to be tested, all differentially expressed genes and their expression levels with samples without exon deletions need to be screened out.

[0075] 4. Use bcftools software to obtain the mutation frequencies of IKZF1 exons and IKZF1 introns respectively.

[0076] 5. The number of IKZF1 exon connections, differentially expressed genes and their expression levels, and IKZF1 exon and intron mutation frequency information of all the above samples were sorted into a table corresponding to the model input, and the saved models were loaded respectively to predict the deletion type of the sample IKZF1 through model.predict.

[0077] After model prediction, the prediction results of the 20 samples to be tested were consistent with the IKZF1 deletion type obtained by the MLPA method. The model typing was accurate and reliable. The specific results are as follows:

[0078]

[0079]

[0080] Although the present invention is disclosed as above, the protection scope of the present invention is not limited thereto. Those skilled in the art may make various changes and modifications without departing from the spirit and scope of the present invention, and these changes and modifications will fall within the protection scope of the present invention.

Claims

1. An IKZF1 gene exon deletion identification system based on high-throughput transcriptome sequencing, characterized in that: include: The characteristic data analysis module is used to analyze the characteristic data of the sample to be tested, including: the characteristic data of the IKZF1 gene in the transcriptome data of the sample to be tested, and the differential gene expression between the sample to be tested and the negative sample; An exon deletion recognition module is used to input the characteristic data of the sample to be tested into a deletion recognition model to obtain an IKZF1 gene exon deletion recognition result; the deletion recognition model is trained using a random forest model; the training process of the deletion recognition model is based on the IKZF1 gene characteristic data of each training sample transcriptome data, and the differential gene expression amount with the negative sample as input features; The IKZF1 gene characteristic data includes the number of exon connections, the mutation frequency of exons, and the mutation frequency of introns; the negative sample is a sample without any deletion type of the IKZF1 gene; the training sample includes positive samples of all deletion types, and each positive sample has one deletion type.

2. The IKZF1 deletion recognition system according to claim 1, characterized in that The types of exon deletions include 4-7 deletion, 1-8 deletion, 2-7 deletion, 2-3 deletion, 4-8 deletion, 2-8 deletion, 1-3 deletion, and 1 deletion.

3. The IKZF1 deletion identification system according to claim 1, characterized in that The test sample and / or training sample is preprocessed before use, including filtering low-quality reads, removing adapter sequences and redundant bases.

4. The IKZF1 deletion recognition system according to claim 1, characterized in that The exon mutation frequency is the exon heterozygous mutation frequency; the exon heterozygous mutation is a variation in which the allele frequency is between 20% and 80%.

5. The IKZF1 deletion recognition system according to claim 1, characterized in that The differential gene expression was determined using DESeq2 software to identify differentially expressed genes and perform differential analysis. The results included gene IDs, standardized gene expression values, and log2-converted differential fold values.

6. The IKZF1 deletion identification system according to claim 1, characterized in that The exon connection number is calculated using featureCounts software based on a defined gene annotation file to calculate the number of each exon connection in a gene.

7. The IKZF1 deletion recognition system according to claim 1, characterized in that The intron mutation frequency performs an intersection operation on the defined IKZF1 gene intron position file, and obtains a file of the number of heterozygous mutations occurring in the intron by analysis; the intron heterozygous mutation is a variation with an allele frequency between 20% and 80%.

8. A method for identifying exon deletions in the IKZF1 gene based on high-throughput transcriptome sequencing, for non-diagnostic purposes, characterized in that: include: Analyze the characteristic data of the samples to be tested: including the characteristic data of the IKZF1 gene in the transcriptome data of the samples to be tested, and the differential gene expression between the samples to be tested and the negative samples; Inputting the characteristic data of the sample to be tested into the deletion recognition model to obtain the IKZF1 gene exon deletion recognition result; the deletion recognition model is trained by a random forest model; the training process of the deletion recognition model is based on the IKZF1 gene characteristic data of each training sample and the differential gene expression amount with the negative sample as input features; The IKZF1 gene characteristic data includes the number of exon connections, the mutation frequency of exons, and the mutation frequency of introns; the negative sample is a sample without any deletion type of the IKZF1 gene; the training sample includes positive samples of all deletion types, and each positive sample has one deletion type.

9. An electronic device, characterized in that The method comprises a processor and a memory, wherein the memory stores a computer program, and when the processor executes the computer program, the method according to claim 8 is implemented.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and the processor executes the computer program to implement the method according to claim 8.

Images