CLC-Pred Synergy: Web Application for Predicting Pairwise Drug Combinations with Synergistic Activity Against NCI60...

Area of Science:

• Computational chemistry
• Pharmacology
• Bioinformatics

Background:

• Pharmacological interventions are key in cancer therapy, but drug resistance and severe side effects limit efficacy.
• Identifying effective drug combinations is challenging due to the vast number of possibilities.
• Structure-activity relationship (SAR) modeling offers a computational approach to predict drug synergy.

Purpose of the Study:

• To develop and validate SAR models for predicting synergistic anticancer drug combinations.
• To create a practical tool for in silico screening of drug pairs.
• To improve the efficacy and reduce the toxicity of cancer therapies through rational drug combinations.

Main Methods:

• Utilized the NCI-ALMANAC database to build training sets of drug pair cytotoxicity profiles.
• Developed SAR models using the PASS DDI platform, encoding chemical structures into PoSMNA descriptors.
• Employed a Bayesian-like algorithm and leave-one-out cross-validation (LOO CO CV) to assess model performance (AUC).

Main Results:

• Generated 104 robust SAR models with predictive accuracy (AUC) exceeding 0.7 (mean AUC = 0.75).
• Successfully predicted synergistic effects for drug combinations across 45 cancer cell lines.
• Implemented the models in the CLC-Pred Synergy web application for practical in silico screening.

Conclusions:

• Developed accurate SAR models for predicting anticancer drug synergy.
• The CLC-Pred Synergy tool facilitates efficient in silico screening of potential drug combinations.
• This approach can accelerate the discovery of more effective and less toxic cancer therapies.