Integrative Cancer Pharmacogenomics to Infer Large-Scale Drug Taxonomy

Area of Science:

• Computational biology
• Pharmacogenomics
• Drug discovery

Background:

• Accurate identification of drug targets and mechanism of action (MoA) is crucial for optimizing anticancer drug efficacy.
• Current MoA prediction methods rely on prior information, limiting their applicability to novel, uncharacterized small molecules.
• There is a need for unbiased approaches to elucidate drug relationships and classify new compounds using basic input data.

Purpose of the Study:

• To develop an integrative computational pharmacogenomic approach, Drug Network Fusion (DNF), for inferring scalable drug taxonomies.
• To enable classification of new experimental compounds with minimal prior information by integrating drug structural information, high-throughput drug perturbation, and drug sensitivity profiles.
• To provide a flexible tool for identifying drug-drug relationships and potential new targets or MoA for cancer drugs.

Main Methods:

• Developed Drug Network Fusion (DNF), an integrative computational pharmacogenomic framework.
• Integrated drug structural information, high-throughput drug perturbation data, and drug sensitivity profiles.
• Applied DNF to infer scalable drug taxonomies and classify new experimental compounds.

Main Results:

• DNF successfully inferred pertinent and novel drug-drug relationships.
• The DNF taxonomy enabled classification of new experimental compounds with minimal prior information.
• The scalability of DNF facilitated the identification of key drug relationships across different drug categories.

Conclusions:

• DNF is a valuable resource for the cancer research community, providing new hypotheses on compound MoA.
• DNF aids in identifying novel drug targets and offers insights for drug repurposing.
• The approach supports potential clinical applications in precision medicine by elucidating drug relationships.