Drug repurposing:

Insights from Multi-Omics and Machine Learning

Drug repurposing offers a high-impact opportunity to identify new therapeutic uses for existing drugs—especially in complex, treatment-resistant cancers. Genome Enhancer is a machine learning–powered platform that applies upstream regulatory modeling to multi-omics data, integrating RNA-seq and DNA methylation to unravel hidden mechanisms behind disease phenotypes.

In a case study on cisplatin-resistant ovarian cancer, Genome Enhancer was used to analyze transcriptomic and epigenomic profiles of treated and untreated cancer cells. The goal was: to identify druggable targets and repurposing-ready drugs that could help overcome resistance.

The analysis revealed several transcription factors (EP300, NFYA, RELA, RXRA, HSF1) driving gene expression changes in the resistant phenotype. Through regulatory network reconstruction, key master regulators upstream of these TFs were identified—including PDGFRA, VRK1, and the CDK1–Cyclin B1 complex.

These master regulators represent pivotal control points in the disease-specific signaling network—and offer promising entry points for therapeutic intervention.

How Master Regulators Are Identified

  1. RNA-seq, ChIP-seq, ATAC-seq, DNA-methylation and other multi-omics data Input:
    The workflow begins with identification of differentially expressed and epigenetically regulated genes from patient-derived omics data.
  2. TF Binding Site Prediction:
    Promoter and enhancer regions of these genes are analyzed using the TRANSFAC® database and MATCH™/CMA algorithms to detect overrepresented transcription factor binding sites (TFBSs) and their combinations.
  3. Pathway Tracing:
    Using TRANSPATH®, the tool reconstructs upstream signaling pathways that could explain the activation of these TFs.
  4. Master Regulator Detection:
    Specialized graph algorithms scan the pathway network to pinpoint nodes at the top of these cascades—called master regulators. These are molecules whose activity explains the global shift in transcriptional programs seen in the pathology.

How Drugs Are Prioritized

  1. Druggability Assessment:
    Each identified master regulator is checked against the HumanPSD™ database for known drug–target relationships and literature evidence.
  2. Biological Activity Prediction (PASS):
    For regulators with no direct drug hits, the system uses PASS (Prediction of Activity Spectra for Substances) to identify chemical compounds likely to interact with those targets based on structural similarity and known activity profiles.
  3. Compound Filtering:
     The final drug list includes:
    • Approved FDA drugs with known efficacy in related pathways
    • Investigational compounds with high predicted interaction scores
    • Prioritization based on network-level modulation and existing pharmacokinetics/safety data

In this study, Paclitaxel and Etoposide were highlighted as promising candidates for repurposing in cisplatin-resistant ovarian cancer based on their predicted ability to modulate the disease-specific regulatory network.

What You’ll Gain

  • A full systems-level view of molecular dysregulation in cisplatin resistance
  • A prioritized list of master regulators and actionable targets
  • Repurposing candidates from FDA-approved and investigational drug libraries
  • A replicable methodology for >3,900 diseases using Genome Enhancer