Mining brain-wide gene expression data to identify imaging genomic modules via biclustering (Abstract)

Published in Organization for Human Brain Mapping, 2021

Citation: Jingxuan Bao, Mansu Kim, Xiaohui Yao, Trang Le, Patryk Orzechowski, Jingwen Yan, Andrew Saykin, Jason Moore, Li Shen. (2021). "Mining brain-wide gene expression data to identify imaging genomic modules via biclustering." Organization for Human Brain Mapping 2021.

Abstract

Allen Human Brain Atlas (AHBA), a brain-wide genome-wide (BWGW) gene expression data set, is a natural connection between genome and brain. We previously proposed to identify meaningful sub-portions of AHBA, called imaging genomic modules (IGMs), to capture local co-expression patterns across imaging and genomic domains, and then used IGMs to help mine high level imaging genetic associations. Our prior method applied hierarchical clustering twice to partition genes and brain regions of interest (ROIs) separately, and had two limitations: (1) it could only identify grid-like non-overlapping biclusters; and (2) global correlations were used for clustering, which was not suitable for finding local co-expression patterns. In this work, we propose a new method to overcome these limitations.

Abstract

Jingxuan Bao, Mansu Kim, Xiaohui Yao, Trang Le, Patryk Orzechowski, Jingwen Yan, Andrew Saykin, Jason Moore, Li Shen. (2021). "Mining brain-wide gene expression data to identify imaging genomic modules via biclustering." Organization for Human Brain Mapping 2021.