Analysis method of Tissue-specific gene set weight

Abstract

Gene set testing is one of the indispensable analysis methods for analyzing high-dimensional genome data. Although the expression and function of many genes are tissue-specific, gene set testing is performed in a way that does not distinguish between tissue types. In order to cope with this challenge, this article uses the tissue-specific gene activity information in the human protein atlas to calculate the tissue-specific gene weights, and uses the filtered gene set in the molecular signature database to generate the tissue-specific gene set weights. In order to prove the validity of these weights, this article use the weights to perform gene set tests on the gene expression data of the three diseases, and use the p-values generated by the tests to weight. Through the weighted FDR analysis, this research can conclude the proposed tissue-specific gene set weights significantly improve the statistical power of gene set testing, and more accurately identified the biological gene set association information of high-dimensional genome data.