第82回インシリコ・メガバンク研究会を下記のとおり行いますのでご案内いたします。今回は大阪大学免疫学フロンティア研究センター・Alexis Vandenbon 先生を講師としてお迎えし、「Immuno-Navigator, a batch-corrected coexpression database, reveals cell type-specific gene networks in the immune system」というタイトルで講演していただきます。
・日時：平成29年2月10日(金) 5:00‐6:30 pm
・演題：Immuno-Navigator, a batch-corrected coexpression database, reveals cell type-specific gene networks in the immune system
・講師：Alexis Vandenbon （大阪大学免疫学フロンティア研究センター）
・概要：Large amounts of experimental data available in public databases contain an enormous potential for elucidating gene regulatory interactions. However, in practice it is difficult to extract new knowledge or hypotheses from such data. One obstacle is the study-specific biases or batch effects that are present in the original data. Although gene coexpression is often used for inferring biological networks, how batch effects influence such networks is not well studied. Here, we prepared a large collection of gene expression data for 24 cell types of the mouse immune system. We found widespread batch effects in this data, and showed that they strongly affect gene coexpression estimates. Removal of batch effects considerably improved the consistency between inferred correlations and prior knowledge. Using the processed data, we constructed Immuno-Navigator, a batch-corrected gene coexpression database. Using our database, we generated hypotheses about candidate regulators in specific immune cells. In one application we successfully predicted known regulators of importance in naturally occurring Treg cells from their expression correlation with a set of Treg-specific genes. For one high-scoring gene, integrin β8 (Itgb8), we experimentally confirmed an association between Itgb8 expression and Treg-specific epigenetic remodelling. We believe that Immuno-Navigator (sysimm.ifrec.osaka-u.ac.jp/immuno-navigator/) will be of great use for generating hypotheses for specific cell types in studies of the immune system.