Dynamic rewiring of biological activity across genotype and lineage revealed by context-dependent functional interactions

Abstract

Coessentiality networks derived from CRISPR screens in cell lines provide a powerful framework for identifying functional modules in the cell and for inferring the role of uncharacterized genes. However, these networks integrate signal across all underlying data, and can mask strong interactions that occur in only a subset of the cell lines analyzed. Here we decipher dynamic functional interactions by identifying significant cellular contexts, primarily by oncogenic mutation, lineage, and tumor type, and discovering coessentiality relationships that depend on these contexts. We recapitulate well-known gene-context interactions such as oncogene-mutation, paralog buffering, and tissue-specific essential genes, show how mutation rewires known signal transduction pathways, including RAS/RAF and IGF1R-PIK3CA, and illustrate the implications for drug targeting. We further demonstrate how context-dependent functional interactions can elucidate lineage-specific gene function, as illustrated by the maturation of proreceptors IGF1R and MET by proteases FURIN and CPD. This approach advances our understanding of context-dependent interactions and how they can be gleaned from these data.

Sparse dictionary learning recovers pleiotropy from human cell fitness screens

“…Despite being highly related genes, NRAS and KRAS fitness effects are in fact anti-correlated (Pearson cor = −0.17) because their activating mutations occur in mutually exclusive cell lines. This presents a paradox that guilt-by-association approaches alone fail to solve, instead requiring supervised learning approaches with cell line mutation data as input ( Kim et al, 2021 ).…”

Optimal construction of a functional interaction network from pooled library CRISPR fitness screens

“…Partial correlation can measure the similarity of two gene knockout profiles after removing the effect of a third gene, set of genes, or another vector. If the functional relationship between two genes varies across the assayed cell lines, the observed correlation can be diluted by the presence of samples in which the relationship is severed [25]. For example, consider common oncogenes KRAS and NRAS, two members of the MAP kinase signaling pathway whose mutually exclusive mutations mean they are essential in different cell lines (Fig.…”

“…4D). CPD protein is known to process pro-IGF1R into its mature form [26] and we recently showed that CPD is also involved in maturation of MET receptor protein specifically in glioma cells [25].…”

Optimal construction of a functional interaction network from pooled library CRISPR fitness screens

“…Partial correlation can measure the similarity of two gene knockout profiles after removing the effect of a third gene, set of genes, or other vector. If the functional relationship between two genes varies across the assayed cell lines, the observed correlation can be diluted by the presence of samples in which the relationship is severed (Kim et al, 2022b). For example, consider common oncogenes KRAS and NRAS , two members of the MAP kinase signaling pathway whose mutually exclusive mutations mean they are essential in different cell lines (Figure 4A).…”