Integrating Multi-Omic Data to Identify Mechanisms of Drug Sensitivity and Biomarker Hypotheses

Study Context

A biotechnology company developing a targeted oncology therapy sought to understand the biological determinants of response to their compound across tumor models. Initial experimental work demonstrated heterogeneous sensitivity across cancer cell lines — some showed strong transcriptional and phenotypic responses to pathway inhibition, while others exhibited little or no response. The team used Plex to integrate experimental datasets with global biological knowledge to generate mechanistic hypotheses.

Data Inputs

• Cell line drug sensitivity data — tumor models classified as sensitive or insensitive based on phenotypic response from in vitro screening
• Transcriptomic data — RNA-seq experiments identifying genes up/downregulated following compound treatment, plus pathway enrichment signatures
• Genetic perturbation and dependency data — gene essentiality, loss-of-function perturbation signatures, copy number alterations, and CRISPR screen results

Plex Analytical Framework

Plex integrated these heterogeneous datasets within a cross-indexed knowledge graph linking genes, proteins, pathways, regulatory networks, genomic alterations, disease annotations, cell line models, perturbation datasets, and published literature. Researchers input experimental data — including gene lists, molecular signatures, and cell line classifications — and retrieved biologically connected results ranked by relevance across datasets.

Oncogenic Signaling Pathway Interactions

Transcriptional signatures associated with therapy response were enriched for genes linked to well-characterized oncogenic signaling pathways regulating tumor proliferation, transcriptional control, and cellular stress responses. These connections were identified through cross-dataset relationships between RNA-seq signatures, perturbation datasets, and pathway annotations. Importantly, these pathways had been described individually in prior literature but were automatically reconstructed by Plex through the integrated knowledge graph.

Genomic Alterations Associated with Drug Sensitivity

Comparative analysis of sensitive versus insensitive cell lines identified a cluster of genes with increased expression or copy number within a specific chromosomal region frequently amplified across multiple cancer types. Nearly all genes with increased expression in sensitive tumor models mapped to this locus, suggesting that genomic amplification in this region may influence sensitivity to the therapy. These genomic features represent potential biomarkers for future translational studies.

Chromatin Regulators Modulating Transcriptional Response

RNA-seq analysis revealed that chromatin remodeling proteins and transcriptional co-factors are connected to the targeted pathway. These regulators influence transcriptional programs controlling cell identity, differentiation, and tumor progression — suggesting that pathway inhibition alters broader transcriptional states rather than a single signaling cascade. This aligns with emerging evidence that transcriptional regulatory networks play a central role in oncogenic cell state transitions.

Implications for Drug Development

• Patient selection strategy — genomic and transcriptional features associated with sensitivity serve as candidate biomarkers for clinical studies
• Mechanistic understanding — the therapy influences transcriptional regulatory networks rather than acting solely through a single pathway
• Combination therapy opportunities — chromatin regulators and transcriptional co-factors represent potential combination partners