Interpretable Deep-Learning p<i>K</i>_a Prediction for Small Molecule Drugs via Atomic Sensitivity Analysis.

ABSTRACT: Machine learning (ML) models now play a crucial role in predicting properties essential to drug development, such as a drug's logscale acid-dissociation constant (pK_a). Despite recent architectural advances, these models often generalize poorly to novel compounds due to a scarcity of ground-truth data. Further, these models lack interpretability. To this end, with deliberate molecular embeddings, atomic-resolution information is accessible in chemical structures by observing the model response to atomic perturbations of an input molecule. Here, we present BCL-XpKa, a deep neural network (DNN)-based multitask classifier for pK_a prediction that encodes local atomic environments through Mol2D descriptors. BCL-XpKa outputs a discrete distribution for each mo

SUBMITTER: DeCorte J 

PROVIDER: S-EPMC11733947 | biostudies-literature | 2025 Jan

REPOSITORIES: biostudies-literature