Research And Application Of Disease Related Signaling Pathway Enrichment Analysis Methods

Signaling pathways are a series of enzymatic reactions in which signals are transmitted from the extracellular to the intracellular and exert their effects.On the account of molecular biology,the signaling pathway is a biological regulatory network in essence,which includes genes,interactions between genes and biological functions or processes that control the biological behavior of cells,etc.Abnormalities in any part of the biological regulatory network can cause changes in the biological behavior of cells through cascade reactions.Abnormal signaling pathways are closely related to the occurrence and development of many diseases,especially tumors,diabetes mellitus,neurodegenerative diseases,and other complex diseases.Therefore,in order to elaborate the molecular mechanism of important life activities and the occurrence,development rules and mechanisms of diseases at the molecular and cellular levels,it is crucial to do signaling pathway analysis.The commonly used signaling pathway analysis methods are mainly based on limited information such as the expression data of genes or the types of interactions between genes.However,there are still many real biological factors that have not been taken into account in the enrichment process of disease-related signal pathways.Therefore,we developed three more efficient specific signaling pathway analysis methods,to fully parse and use the information of topology contained in signaling pathways,and to better doing disease related signaling pathways enrichment analysis.In addition,all signaling pathway analysis methods can only judge whether signaling pathways is related to diseases based on the p-values,which provides limited help for further researches.Therefore,in order to remedying this shortcoming,we developed a method to assess and classified visualize the disease specificity of disease-related signaling pathways.The research content includes the following several aspects:1 Improved signaling pathway impact analysis with gene weights.From the perspective of genes,few proposed signaling pathway analysis methods have make use of the different importance of genes in signaling pathways,which encompass the constitutive functional nonequivalent roles of genes in real signaling pathways and none of proposed signaling pathway analysis methods tries to associate the importance of genes with related disease.For remedying the shortcomings in the perspective of genes,we improved the signaling pathway impact analysis method(SPIA)with three different gene weights which could reflect the importance of genes from three different aspects and the different importance of a gene to different diseases,called gwSPIA.The validation results showed that gwSPIA could effectively identify disease related signaling pathways.In addition,the validation results also showed that the gene weights used in gwSPIA could reflect the different correlations between genes and different diseases.This indicates that the gene weight can be used to screen gene biomarkers.2 Signaling pathway analysis combined with the strength variations of interactions between genes under different conditions.From the perspective of interactions between genes,existing signaling pathway analysis methods mainly focus on the differential expression of genes in the signaling pathways under two different conditions,ignoring the strength variations of interactions between genes in signaling pathways under two different conditions.We proposed a new signaling pathway analysis method combined with the strength variations of interactions between genes under different conditions,called SPACI,aiming to solve the problem of gwSPIA and existing methods in terms of interactions between genes in signaling pathways.The validation results showed that SPACI method could effectively identify the signaling pathways which were related to the corresponding diseases.The results also showed that the interactions with high strength variations calculated by SPACI method are closely and specifically related to diseases.3 Signaling pathway analysis based on the signal variations received by effector genes under different conditions.From the perspective of the biological functions or processes which control the cell biological behaviors,most signaling pathway analysis methods ignore the correlation between the signal variations the effector genes(directly affect the biological functions or processes of genes in the end of signaling pathway)received from the upstream genes under two different conditions and diseases.Although gwSPIA and SPACI have solved the problems of existing signaling pathway analysis methods in terms of genes and interactions between genes respectively,they have not solved the problems of existing signaling pathway analysis methods in terms of the biological functions or processes which control the cell biological behaviors.Therefore,we proposed a new signal pathway analysis method based on the signal variations received by effector genes under different conditions,called SPFA.The validation results showed that SPFA could effectively identify the signaling pathways related to diseases.The results also showed that SPFA could help us distinguish which biological processes or functions were abnormal in a signaling pathway or which parts of a signaling pathway were abnormal according to the analysis of effector genes to better interpret the signaling pathways and understand the biological behavior of disease cells.4 Assessing the disease specificity of signaling pathways and classifying and visualizing the results of signaling pathway analysis.Exiting signaling pathway analysis methods can only judge whether signaling pathways is related to diseases based on the p-values,which provides limited help for further researches.Therefore,according to the classification information of signaling pathways in the KEGG database,we classified and visualized the results of signaling pathway analysis to provide help for further researches.Then,we developed a method to assess the disease specificity of signaling pathways based on the classification information of signaling pathways.We firstly construct a connection network based on the direct relationship between non-disease signaling pathways and disease signaling pathways in the KEGG database.Then we calculated the disease specificity of signaling pathways according to the degree of non-disease signaling pathways in the network.Verification results showed that the disease specificity calculated in this paper could help distinguish whether the signaling pathway is a disease-specific signaling pathway and provided helpful information for further researches.In this paper,according to the deficiencies of the existing signaling pathway analysis methods in terms of genes,interactions between genes,and effector genes,three signaling pathway analysis methods were proposed to make more comprehensive use of signaling pathways and realize more accurate identification of disease-related signaling pathways.And we developed a signaling pathway disease specificity assessment and visualization method to provide more valuable information for the next study of signaling pathway analysis.