Abstract
A fundamental goal of biological research is to determine the interactions and functional relationships between genes and their coded proteins that drive biological responses. Understanding the response of the global transcriptome in the context of pathogenesis and drug-related adversities can reveal gene-response relationships that contribute to biogical insights and more accurate and reliable mechanism-based safety assessments. Although transcriptomic data provide a framework to systematically determine gene activity, their high dimensionality and complexity can make interpretation and analysis challenging. Gene co-expression analysis addresses these difficulties in analyzing transcriptomics data by first constructing networks of genes that are co-expressed across treatments, reducing complexity, and then inferring biological relevance and gene-pathology associations for each network. Variation in gene expression in bulk tissue helps define co-expression relationships, but the cell type heterogeneity, inherent to bulk tissue, can also complicate biological interpretations. Consequently, interpretation of some tissue gene co-expression patterns may be subject to the confounding influence of variations in cellular composition obscuring intra-cell-type-specific co-expression network responses. In this review, we highlight methods designed to capture cell type-specific co-expression patterns and discuss their potential utility for understanding mechanisms of toxicity and pathogenesis.