High-throughput transcriptome sequencing-based IKZF1 gene exon deletion recognition system and method

Abstract

Provided are a high-throughput transcriptome sequencing-based IKZF1 gene exon deletion recognition system and method. The exon deletion recognition system is obtained by inputting the number of IKZF1 exon junctions, differential genes and expression quantities thereof, and IKZF1 exon and intron mutation frequency information of samples as features into a constructed random forest model and performing prediction, and can be used for accurately predicting and recognizing any exon deletion with high accuracy and high specificity, providing guidance for patient prognosis and treatment.

Description

A System and Method for IKZF1 Gene Exon Deletion Recognition Based on High-Throughput Transcriptome Sequencing Technical Field

[0001] This invention belongs to the field of gene detection technology, specifically relating to an IKZF1 gene exon deletion identification system and method based on high-throughput transcriptome sequencing. Background Technology

[0002] The IKZF1 gene (IKAROS1 of the IKAROS family) is located on chromosome 7 p12 and has 8 exons. The IKROS protein it encodes plays a crucial role in the development and function of the hematopoietic system. Local deletion of IKROS family zinc finger protein 1 (IKZF1) leads to the loss and dysfunction of the zinc finger DNA-binding domain, thereby affecting normal cell development and maturation. ALL (acute lymphoblastic leukemia) is the most common malignant tumor in children, accounting for 26% of childhood cancers. IKZF1 exon deletions are common in ALL, occurring in approximately 15% of children with B-cell ALL and 30%–40% of adults. In Ph-chromosome-positive ALL, IKZF1 gene deletions or sequence mutations are detected in over 70% of cases. IKZF1 exon deletions are closely related to the pathogenesis of high-risk B-cell ALL, poorer treatment prognosis, increased risk of relapse, and the ability of IKAROS transcription factors to function properly.

[0003] The main types of exon deletions in IKZF1 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. Among all deletion types, 4-7 deletion and 1-8 deletion are the most common, each accounting for about 30%, followed by 4-8 deletion and 2-7 deletion. Other deletion types are relatively less common.

[0004] Currently, the main methods for detecting IKZF1 exon deletions include microarray technology, reverse transcription polymerase chain reaction (RT-PCR), and multiplex ligation-dependent probe amplification (MLPA). Microarray technology and RT-PCR have drawbacks such as low detection sensitivity, difficulty in quantitative analysis, and difficulty in identifying rare IKZF1 mutant transcripts. MLPA is a qualitative and semi-quantitative analysis technique developed in recent years for detecting DNA sequences. In MLPA, two oligonucleotide fragments hybridize with the target sequence, and then a ligase ligates the probes. Due to the high specificity of the ligation reaction, the ligase can only connect the two probes into a single nucleic acid strand if the two probes are completely hybridized with the target sequence. If the target and probe sequences are not perfectly complementary, the ligation reaction will fail.

[0005] RNA sequencing (RNA-seq) is a high-throughput gene sequencing technology that sequences all gene transcripts in a sample, yielding the gene sequences of all transcripts. RNA-seq data contains rich information on gene expression, gene fusions, splice isoforms, and gene mutations, making it an effective method for identifying genomic damage in patients and providing crucial information on patient prognosis and treatment. Summary of the Invention

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

[0007] In view of the above, the solution of the present invention is as follows:

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

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

[0010] The exon deletion identification module is used to input the feature data of the sample to be tested into the deletion identification model to obtain the IKZF1 gene exon deletion identification result; the deletion identification model is trained using a random forest model; the training process is based on the IKZF1 gene feature data of the transcriptome data of each training sample, and the differential gene expression level with negative samples as input features;

[0011] The IKZF1 gene feature data includes the number of exon connections, the mutation frequency of exons, and the mutation frequency of introns; the negative samples are samples of the IKZF1 gene without any deletion type; the training samples contain positive samples with 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 samples and / or training samples are 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 differentially expressed genes were identified using DESeq2 software, and differential analysis was performed. The results included gene IDs, standardized gene expression values, and log2-converted fold change values.

[0016] Furthermore, the number of exon connections is calculated using the featureCounts software based on a predefined gene annotation file, which calculates the number of connections for each exon in the gene.

[0017] Furthermore, the mutation frequency of the introns is obtained by performing an intersection operation on the defined IKZF1 gene intron location files to analyze and obtain a file showing the number of heterozygous mutations occurring in the introns.

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

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

[0020] The feature data of the sample to be tested is input into the deletion identification model to obtain the IKZF1 gene exon deletion identification result; the deletion identification model is trained using a random forest model; the training process is based on the IKZF1 gene feature data of each training sample and the differential gene expression level compared with the negative sample as input features;

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

[0022] The aforementioned non-diagnostic purposes include predicting samples that cannot be traced back to a person, predicting the type of missing data in the sample itself, including for studying the pathogenesis of high-risk B-ALL, determining whether IKAROS transcription factors can function normally, and other biological functions.

[0023] A third aspect of the invention is to provide an electronic device comprising a processor and a memory, the memory storing a computer program, wherein the processor, when executing the computer program, implements the method as described in the second aspect.

[0024] A fourth aspect of the present invention is to provide a computer-readable storage medium on which a computer program is stored, wherein a processor executes the computer program to implement the method as 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 this invention is constructed using the IKZF1 exon linkage number, differentially expressed genes and their expression levels, and IKZF1 exon and intron mutation frequency information of the sample and constructed through a random forest model. It can be used to accurately predict and identify any exon deletion with high accuracy and strong specificity, providing guidance for patient prognosis and treatment. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 is a flowchart of the construction process of the IKZF1 exon deletion identification model described in this invention.

[0029] Figure 2 is a volcano diagram of differentially expressed gene analysis during the model construction process of this invention.

[0030] Figure 3 is a heatmap of differentially expressed gene analysis during the model construction process of this embodiment of the invention. Detailed Implementation

[0031] The following provides definitions for some of the terms used in this specification. Unless otherwise stated, all terms used herein have the meaning commonly understood by one of ordinary skill in the art to which this solution pertains.

[0032] Terminology Explanation:

[0033] Random Forest is an ensemble learning method primarily used for classification and regression tasks. It generates multiple subsets of the original dataset using sampling with replacement. Each subset is used to train a decision tree. At each node, a subset of features is randomly selected for splitting, continuing until a specified number of decision trees are generated. Finally, the predictions from all decision trees are used to determine the final classification result through majority voting. The model constructs multiple decision trees and combines their predictions to improve the model's accuracy and robustness.

[0034] BAM is a commonly used binary file format primarily used for storing large-scale sequencing genomic sequence data, especially sequencing data aligned to a reference genome. It consists of a header and an alignment record, which includes information such as the reference sequence, sequencing platform, alignment algorithm, and the position, orientation, and base quality of the sequencing reads. The binary format offers many advantages, such as more compact data storage, space saving, and easier reading and writing. The BAM format is supported by various bioinformatics tools and analysis software, enabling data processing, filtering, annotation, and other operations, facilitating data exchange and sharing. It has been widely used in genomics research, personalized medicine, and population genetics.

[0035] genomic.rename.gff: A database containing information such as chromosomes and their origins, gene names, and regions.

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

[0037] 1. Data Preprocessing

[0038] FastP was used for quality control of the raw sequencing data, including filtering low-quality reads, removing adapter sequences and redundant bases, to improve the accuracy of subsequent analyses. STAR software was used to align the data reads to a specified hg19 reference genome, followed by SAMBAMBA reordering to obtain a BAM file containing genome alignment information.

[0039] 2. Key Feature Extraction

[0040] (1) Number of exon connections

[0041] Transcriptome sequencing was performed on 145 samples with known IKZF1 exon deletion types and 50 patient samples without IKZF1 exon deletions. By analyzing the BAM files using featureCounts software, detailed information on the interconnections between exons in the IKZF1 gene and the specific number of these connections could be obtained. Specific operational steps included:

[0042] First, a high-quality BAM file was prepared, obtained from high-throughput transcriptome sequencing data after alignment processing. Next, the BAM file was used as input for quantitative analysis using featureCounts software. This software can calculate the number of links for each exon in a gene based on a predefined gene annotation file (genomic.rename.gff), revealing different exon linking patterns and numbers in different sample deletion types. Since the total number of IKZF1 gene reads aligned to different samples varies, this step requires normalization of the linking count to eliminate the influence of read count on exon linking statistics. These data are crucial for understanding gene deletion types and expression regulation, and help to further explore the function of the IKZF1 gene in different biological processes and its potential disease associations. This method allows for a more comprehensive mapping of exon linking patterns, laying the foundation for subsequent prediction of IKZF1 exon deletion types.

[0043] (2) Differentially expressed genes

[0044] The most direct purpose of transcriptome sequencing is to identify genes with significant differential expression between groups and to explain the impact of changes in gene expression levels on biological functions. The DESeq2 package is an R package designed for high-dimensional counting data in transcriptome datasets, focusing on data normalization, visualization, and differential expression analysis. Its main function is to use empirical Bayesian methods to estimate prior values ​​for logarithmic fold changes and deviations, and to calculate posterior values ​​for these statistics. It is one of the most frequently used R packages for identifying differentially expressed molecules. For each deletion type of sample and ALL patients without IKZF1 exon deletions (no IKZF1 gene exon deletions were present), the DESeq2 software was used to identify the differentially expressed genes corresponding to each deletion type. The differential analysis results included key information such as gene ID, standardized gene expression value, log2 fold change (log2FoldChange) value, significance p-value, and corrected p-value (default FDR correction). Filtering was performed based on |log2FoldChange|≥1 and padj<0.01, and up- and down-regulated genes were distinguished by "up" and "down" respectively. After DESeq2 identified significantly differentially expressed genes, they could be represented using a volcano plot using the R package (ggplot2), a common statistical chart (as shown in Figure 2).

[0045] Analysis revealed that for each deletion type, the differential gene expression levels were different from those in samples without IKZF1 deletion.

[0046] (3) Mutation frequency of IKZF1 exons

[0047] First, the BAM files of the samples were converted into standardized variant annotation files (VCF format) using the `bcftools mpileup` command in the BCFtools software. Next, these VCF files were intersected with predefined IKZF1 gene exon location files using BCFtools software to obtain the mutation frequency and specific location information in each exon. When defining mutation types, variants with allele frequencies between 20% and 80% were defined as heterozygous mutants, allele frequencies below 20% were defined as homozygous wild-type, and allele frequencies above 80% were defined as homozygous mutants. This definition allows for accurate counting of the number of heterozygous mutation types occurring in the IKZF1 exons of each sample. Since heterozygous deletions are the most common type, with homozygous deletions occurring at a frequency of less than 1%, the focus was primarily on heterozygous deletions. These results provide important data support for subsequent molecular mechanism studies.

[0048] (4) Mutation frequency of introns

[0049] The statistical methods for intron mutation frequencies are similar to those for exon mutation frequencies. The difference is that the bedtools software requires an intersection operation with a predefined IKZF1 gene intron location file. The final analysis yields a file containing the number of heterozygous mutations occurring within the introns.

[0050] 3. Random Forest Model Construction

[0051] (5) The feature vectors in the samples are formatted into a specific model input format and used as input to the decision tree. These feature vectors include: exon connectivity, differentially expressed gene expression levels, mutation frequencies of IKZF1 exons, and mutation frequencies of endons. The model is split into training and testing sets in a 7:3 ratio. The RandomizedSearchCV model tuning tool is used to find the optimal combination of model parameters, outputting the optimal parameter combination and its corresponding cross-validation score. Finally, the optimal model is evaluated 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 type of deletion, and multiple samples had each type of deletion) and 50 samples without deletions were collected and processed as follows:

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

[0055] The following table was obtained after quality control analysis of all data using FASTP:

[0056]

[0057] (2) STAR alignment with reference genome

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

[0059]

[0060] (3) Feature extraction

[0061] Feature 1: Exon Connection Count. The number of different IKZF1 exon connections in the samples was counted using featureCounts. The results are shown in the table below (partial sample):

[0062]

[0063] Feature 2: Differentially expressed genes. DESeq2 was used to identify differentially expressed genes in samples with specific exon deletion types and those without exon deletions. Taking the IKZF1 exon 4-7 deletion type as an example, upregulated and downregulated genes were obtained, and the heatmap of differentially expressed genes in the two groups of samples is shown in Figure 3.

[0064] Feature 3: IKZF1 exon mutation frequency. The heterozygous mutations in the exons are statistically analyzed as follows (partial sample):

[0065]

[0066] Feature 4: IKZF1 intron mutation frequency. The heterozygous mutations in introns are statistically analyzed as follows (partial sample):

[0067]

[0068] (4) Model building

[0069] The collected 145 IKZF1 exon deletion cases and 50 non-deletion cases were subjected to data quality control, genome alignment, and feature extraction to obtain data containing feature information of both groups of samples. The feature information was used as feature values, and the IKZF1 exon deletion type of the samples was used as the target value. The processed feature data was used as input to a random forest model. A random forest classifier with 1000 decision trees was created using the Random Forest algorithm, with a random seed of 0 to ensure reproducibility. The sklearn.metrics module was used to evaluate the quality of the model's prediction results, including precision, recall, F1-score, and classification report. The metrics.accuracy_score function in the module compares the model's predicted labels with the true labels and gives the classifier's overall accuracy on a given test dataset. Finally, the classifier, fine-tuned using RandomizedSearchCV, achieved an accuracy of 98.21%, and this classifier was saved.

[0070] (5) Model Testing

[0071] Twenty samples of IKZF1 deletion types were collected and subjected to high-throughput transcriptome sequencing. The specific deletion type was then identified and predicted using a model following these steps:

[0072] 1. The obtained data was quality controlled using FASTP, and the data was aligned with the hg19 human reference genome using STAR to obtain the BAM file.

[0073] 2. The featureCounts software was used to analyze the BAM file to obtain a file containing information on the 13 IKZF1 exon modes and connection numbers.

[0074] 3. Use featureCounts to analyze the BAM file and then normalize it to calculate the expression level (read count) of all genes. In order to keep the model's feature values ​​consistent, for each test sample, it is necessary to screen out all differentially expressed genes and their expression levels compared with samples without exon deletions.

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

[0076] 5. Compile the IKZF1 exon connection number, differentially expressed genes and their expression levels, and IKZF1 exon and intron mutation frequency information for all the above samples into a table corresponding to the model input. Load the saved model and use model.predict to predict the IKZF1 deletion type of the sample.

[0077] Based on the model prediction, the predicted results of the 20 test samples were consistent with the IKZF1 deletion type obtained by the MLPA method. The model classification was accurate and reliable. The specific results are as follows:

[0078]

[0079]

[0080] While the present invention has been disclosed above, its scope of protection is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and all such changes and modifications will fall within the scope of protection of the present invention.

Claims

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

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

4. The IKZF1 missing identification system according to claim 1, characterized in that, The mutation frequency of the exons is the exon heterozygous mutation frequency.

5. The IKZF1 missing identification system according to claim 1, characterized in that, The differentially expressed genes were identified using DESeq2 software, and differential analysis was performed. The results included gene ID, standardized gene expression value, and log2 transformation fold change.

6. The IKZF1 missing identification system according to claim 1, characterized in that, The number of exon connections was calculated using the featureCounts software based on a predefined gene annotation file, which counted the number of connections for each exon in the gene.

7. The construction method according to claim 1, characterized in that, The mutation frequency of the introns is obtained by performing an intersection operation on the defined IKZF1 gene intron location files to analyze the number of heterozygous mutations occurring in the introns.

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 test samples: including the IKZF1 gene characteristic data in the transcriptome data of the test samples, and the differential gene expression levels between the test samples and negative samples; The feature data of the sample to be tested is input into the deletion identification model to obtain the IKZF1 gene exon deletion identification result; the deletion identification model is trained using a random forest model; the training process is based on the IKZF1 gene feature data of each training sample and the differential gene expression level compared with the negative sample as input features; The IKZF1 gene feature data includes the number of exon connections, the mutation frequency of exons, and the mutation frequency of introns; the negative samples are samples of the IKZF1 gene without any deletion type; the training samples contain positive samples with all deletion types, and each positive sample has one deletion type.

9. An electronic device, characterized in that, It includes a processor and a memory, the memory storing a computer program, and the processor executing the computer program to implement the method as described in claim 8.

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 as described in claim 8.

Images