Figure 4
Schematic of analysis pipeline.
(Art credit: Kate Mahan.) We developed an innovative pipeline for extracting new insights from existing publicly available resources, including five gene expression datasets, three SNP datasets and multiple databases, in order to generate hypotheses for potential treatment strategies in subsets of AD patients. (1) We identified genes differentially expressed in AD in multiple gene expression data sets, leading to a list of 24 genes downregulated in AD. (2) We searched all of GEO to identify expression patterns for these genes in additional (not necessarily AD) datasets, revealing gender differences for some genes. (3) For patient subsets of interest (defined by gender and APOE4 status), we performed GWAS in each of three separate SNP data sets (AD vs healthy control) in regions up and downstream of 24 genes, allowing prioritization of genes having nearby SNPs. (4) The genes thus identified provide insights that can help facilitate development of new therapeutics. For example, searching the CMAP database identified existing drugs warranting further investigation for specific subsets of AD patients. In addition, further elucidating the biology of these genes can allow identification of entirely new therapeutic targets.
