A binary QSAR model for classification of hERG potassium channel blockers

Area of Science:

• Pharmacology and Toxicology
• Computational Chemistry
• Drug Discovery

Background:

• Acquired long QT syndrome, a serious cardiac side effect, is often linked to drug candidates inhibiting the human ether-a-go-go-related-gene (hERG) potassium channel.
• This inhibition can lead to dangerous arrhythmias, posing a significant challenge in clinical drug development.
• Early prediction of hERG channel affinity is vital for mitigating risks in the drug discovery pipeline.

Purpose of the Study:

• To develop and validate quantitative structure-activity relationship (QSAR) models for predicting hERG potassium channel affinity.
• To assess the classification performance of these models in identifying potential hERG blockers.
• To evaluate the utility of 2D descriptors for rapid virtual screening in drug discovery.

Main Methods:

• Generation of binary QSAR models using two distinct sets of molecular descriptors.
• Utilized a set of 32 P_VSA descriptors and another set selected via a feature selection algorithm.
• Models were calibrated with threshold values at IC(50)=1 and 10 microM for classifying hERG blockers.

Main Results:

• The developed QSAR models achieved a classification power of 82-88% for identifying hERG blockers across the full dataset.
• This performance is comparable to existing classification models for hERG affinity.
• The study demonstrated the effectiveness of using 2D descriptors for predicting hERG channel interaction.

Conclusions:

• Binary QSAR models utilizing 2D descriptors offer a versatile and efficient approach for predicting hERG channel affinity.
• These models can be readily integrated into virtual screening protocols for early-stage drug discovery.
• The findings support the use of computational methods for proactive assessment of cardiotoxicity risks associated with drug candidates.