CANDO and the infinite drug discovery frontier

Area of Science:

• Computational biology
• Drug discovery
• Systems pharmacology

Background:

• Drug repurposing offers a faster, cheaper alternative to de novo drug development.
• Predicting drug efficacy and identifying new therapeutic uses for existing compounds remain challenging.
• Computational methods can analyze complex biological data to uncover hidden therapeutic potential.

Purpose of the Study:

• To introduce the Computational Analysis of Novel Drug Opportunities (CANDO) platform for inferring drug behavior through compound-proteome interaction signatures.
• To computationally predict and analyze novel drug repurposing opportunities.
• To validate the platform's predictions through prospective studies.

Main Methods:

• Constructed interaction signatures for 3733 ingestible compounds against 48,278 protein structures.
• Utilized a signature comparison and ranking approach to identify potential drug-indication relationships.
• Benchmarked prediction accuracy for 1439 indications with approved compounds.
• Prospectively validated 49/82 high-value predictions.

Main Results:

• Achieved benchmarking accuracies of 12-25% for indications with approved compounds.
• Successfully validated 49 out of 82 high-value predictions, demonstrating comparable or superior activity to existing drugs.
• Identified novel repurposed therapeutics for seven indications.

Conclusions:

• The CANDO platform effectively predicts novel drug repurposing opportunities by analyzing compound-proteome interactions.
• The approach shows promise for identifying new therapeutic uses for existing compounds, including personalized medicine applications.
• The multiscale modeling framework has broad applicability in medicine and engineering.