RPath

RPath identifies causal paths in knowledge graphs (KGs) to prioritize drugs for specific diseases by integrating drug-perturbed and disease-specific transcriptomic signatures.

Key Features:

• Causal Path Identification: Identifies causal paths that connect a drug to a disease within a knowledge graph (KG) by correlating path-derived effects with transcriptional changes from drug-perturbation experiments and anti-correlating them with disease-specific molecular signatures.
• Signature-Based Reasoning: Uses transcriptomic data from drug-perturbation experiments and disease-specific signatures to guide reasoning and prioritize causal mechanisms of action.
• Performance and Validation: Prioritizes clinically investigated drug-disease pairs across multiple datasets and KGs and has demonstrated superior performance compared with other similar methodologies.
• Deconvolution and Target Prediction: Enables deconvolution of predictions to dissect mechanistic bases of specific drug-disease pairings and supports prediction of novel targets.

Scientific Applications:

• Target discovery: Supports identification of candidate therapeutic targets by linking drug effects to disease molecular signatures within KGs.
• Drug repurposing: Prioritizes existing drugs for new indications using transcriptomic anti-correlation combined with causal path evidence.
• Side effect prediction: Facilitates inference of potential side effects by mapping drug-perturbation signatures onto disease-relevant pathways in the KG.
• Systems biology and pharmacology: Provides systems-level interpretation of disease pathophysiology and drug mechanisms for applications in systems biology and pharmacology.

Methodology:

Operates on knowledge graphs to identify causal paths between drugs and diseases, compares those paths to transcriptomic signatures from drug-perturbation experiments and disease-specific molecular profiles, and prioritizes paths that correlate with drug-induced transcriptional changes while anti-correlating with disease signatures.