Pred-hERG: A Novel web-Accessible Computational Tool for Predicting Cardiac Toxicity.

Area of Science:

• Cardiovascular Pharmacology
• Computational Chemistry
• Drug Discovery

Background:

• Blockage of the hERG K(+) channel is a significant cause of lethal cardiac arrhythmia.
• The hERG channel's ligand promiscuity makes it a critical antitarget in early drug development.
• Identifying potential hERG blockers early is crucial to mitigate cardiac risks.

Purpose of the Study:

• To develop an innovative and freely accessible web server for the early identification of hERG channel blockers and non-blockers.
• To provide a tool for screening chemical libraries and assessing potential cardiac risks associated with drug candidates.

Main Methods:

• Compiled the largest publicly available curated hERG dataset, comprising 5,984 compounds.
• Developed robust and externally predictive binary and multiclass computational models for hERG channel interaction.
• Integrated these models into a user-friendly web service.

Main Results:

• Achieved a high predictive performance with a binary classification model (CCR≈0.8) and a multiclass model (accuracy≈0.7).
• The developed web server, Pred-hERG, offers predictions from both binary and multiclass models.
• Provides probability maps of atomic contributions to aid in understanding prediction rationale.

Conclusions:

• The Pred-hERG web server is a valuable, freely accessible resource for early-stage drug development.
• Facilitates the identification of potential hERG blockers, thereby reducing the risk of drug-induced cardiac arrhythmia.
• The platform is designed for continuous updates and improvements as new data becomes available.