The Analysis Of Gene Alternative Splicing Mechanism Based On Data Mining

The process of gene expression in eukaryotic organisms mainly includes transcription,splicing and translation,and gene splicing is the pivotal step of such process.Alternative splicing means that multiple proteins can be encoded by only one gene.Gene alternative splicing is a ubiquitous phenomenon in eukaryotic cells.We already know that gene splicing is a highly regulated process involving various regulatory signals like splicing sites,splicing regulatory elements and branch sites.The splicing regulatory elements can be classified as four classes:Exon Splicing Enhancer(ESE),Exon Splicing Suppressor(ESS),Intron Splicing Enhancer(ISE),and Intron Splicing Suppressor(ISS),depending on their function and location.These signals give a combinatorial regulation on the production of mature RNA and translation of proteins.Modeling and analyzing of these regulatory signals are of great significance for the comprehensive understanding of the function of regulatory signals and gene alternative splicing process.Numerous studies indicate that the abnormal alternative splicing is an important pathogenesis for complex and inherited diseases.Genomic variants can induce abnormal alternative splicing by changing the gene splicing regulatory signals,which is an important cause of disease occurrence.To further understand the genetic mechanism of alternative splicing regulation and the pathogenic mechanisms of mutation in the splicing level,this dissertation aims to model and analyze splicing regulatory elements and splicing sites by sequential pattern mining,and further to quantify the effects of genome mutation on the two types of signals.The main innovations of this dissertation are as following:(1)A thorough analysis of the functional properties of RNA splicing regulatory elements and its genetic mutation on splicing regulation.First of all,according to the sequence characteristics of splicing regulatory elements,we selected and verified the quantified recognition ability of sequential pattern mining methods for the splicing functions of splicing regulatory elements,and then used this method to predict and identify potential new splicing regulatory elements.Then,it was discovered through quantitative analysis that the different splicing elements showed similar functions.The frequency characteristics of multiple regulatory elements are detected in the following statistics.Finally,the splicing effect of exon mutation and single nucleotide polymorphism(SNPs)was analyzed.The results of the experiment showed that:a.The mutations that cause exon skipping are more likely to destroy exon splicing enhancer and generate exon splicing inhibition;b.Both exon splicing enhancer and exon splicing inhibition are avoided by SNPs compared with simulated mutations that do not under evolutionary selection.(2)Apply sequential pattern mining to studying the splicing effect of the cancer somatic mutations on splicing site regions.Firstly,extracted all cancer somatic mutations in splicing site regions.Then,compared the splicing score changes of single nucleotide polymorphisms,random mutations and cancer somatic mutations in splice site regions,and we found the cancer somatic mutations have significantly weaken the signal strength of splicing sites.Finally,using TCGA RNA-Seq data to validate the predictions of our method,and found that the RNA-Seq result of ZNF7 were in consistence with the prediction results of the method.And this study provides an approach to explorer cancer pathogenesis.