functional annotation and identification of candidate disease genes by computational analysis of normal tissue gene expression data候选致病基因的功能注释和识别正常组织基因表达数据的计算分析.pdfVIP

Functional Annotation and Identification of Candidate Disease Genes by Computational Analysis of Normal Tissue Gene Expression Data 1 2 3 2 2 2 Laura Miozzi , Rosario Michael Piro , Fabio Rosa , Ugo Ala , Lorenzo Silengo , Ferdinando Di Cunto , Paolo Provero2* 1 Institute of Plant Virology, CNR, Turin, Italy, 2 Molecular Biotechnology Center and Department of Genetics, Biology and Biochemistry, University of Turin, Torino, Italy, 3 ISI Foundation, Turin, Italy Abstract Background: High-throughput gene expression data can predict gene function through the ‘‘guilt by association’’ principle: coexpressed genes are likely to be functionally associated. Methodology/Principal Findings: We analyzed publicly available expression data on normal human tissues. The analysis is based on the integration of data obtained with two experimental platforms (microarrays and SAGE) and of various measures of dissimilarity between expression profiles. The building blocks of the procedure are the Ranked Coexpression Groups (RCG), small sets of tightly coexpressed genes which are analyzed in terms of functional annotation. Functionally characterized RCGs are selected by means of the majority rule and used to predict new functional annotations. Functionally characterized RCGs are enriched in groups of genes associated to similar phenotypes. We exploit this fact to find new candidate disease genes for many OMIM phenotypes of unknown molecular origin. Conclusions/Significance: We predict new functional annotations for many human genes, showing that the integration of different data sets and coexpression measures significantly improves the scope of the results. Combining gene expression