Open-Access Activity Prediction Tools for Natural Products. Case Study: hERG Blockers

Area of Science:

• Pharmacology
• Computational Chemistry
• Drug Discovery

Background:

• hERG channel interference can cause fatal cardiac arrhythmia.
• Drug development is frequently halted due to hERG-associated toxicity.
• Early identification of potential hERG activity is crucial for drug safety.

Purpose of the Study:

• To review and compare computational tools for predicting hERG channel activity.
• To assess the predictive power of various established computational methods.
• To identify needs for improved prediction tools, particularly for natural products.

Main Methods:

• Review of computational tools based on 3D QSAR, similarity, and machine learning.
• Screening of a dataset containing known hERG active and inactive synthetic and natural compounds.
• Comparative analysis of prediction tools including Similarity Ensemble Approach, SuperPred, SwissTargetPrediction, HitPick, admetSAR, PASSonline, Pred-hERG, and VirtualToxLab™.

Main Results:

• Quantification and comparison of the predictive power of multiple computational tools.
• Demonstration of the utility of these tools for evaluating drug candidates.
• Identification of limitations in current tools, especially concerning natural products.

Conclusions:

• Computational methods offer valuable approaches for early-stage hERG activity prediction.
• A comparative analysis aids researchers in selecting appropriate tools for drug development.
• There is a significant unmet need for specialized prediction tools for natural products in hERG toxicity assessment.