Predicting the Activities of Drug Excipients on Biological Targets using One-Shot Learning

Abstract

Excipients are major components of drugs and are used to improve drug attributes such as stability and appearance. Excipients approved by the U.S. Food and Drug Administration (FDA) are regarded as safe for humans in allowed concentrations, but their potential interactions with drug targets have not been investigated systematically, which might influence a drug’s efficacy. Deep learning models have been used for the identification of ligands that could bind to the drug targets. However, due to the limited available data, it is challenging to reliably estimate the likelihood of a ligand–protein interaction. One-shot learning techniques provide a potential approach to address this low data problem as these techniques require only one or a few examples to classify the new data. In this study, we apply one-shot learning models to data sets that include ligands binding to G-protein-coupled receptors (GPCRs) and kinases. The predicted results suggest that one-shot learning could be used for predicting ligand–protein interactions, and the models attain better performance when protein targets contain conserved binding pockets. The trained models are also used to predict interactions between excipients and drug targets, which provides a potential efficient strategy to explore the activities of drug excipients. We find that a large number of drug excipients could interact with biological targets and influence their function. The results demonstrate how one-shot learning can be used to make accurate predictions for excipient–protein interactions, and these methods could be used for selecting excipients with limited drug–protein interactions.

The Experimentalist’s Guide to Machine Learning for Small Molecule Design

“…Recently, the continued improvements in computer hardware, , in combination with advancements in novel computational algorithms, have provided an unprecedented opportunity to accelerate the development process. Such levels of computing power have been frequently utilized in computational biology and have been applied to a variety of research areas, ranging from protein engineering, − drug design and optimization, − large-scale modeling of biological systems, − and clinical diagnostics. − In this review, the applications of machine learning (ML) algorithms on small molecule design will be discussed.…”

Leveraging machine learning models for peptide–protein interaction prediction

“…Moreover, ML/DL techniques exhibit a degree of generalizability. Some advanced techniques like transfer learning or one-shot learning models, which have been applied in protein engineering and protein-ligand interaction prediction, [128][129][130][131] could facilitate the models trained on certain peptide-protein binding datasets to generalize to other peptide-protein complexes.…”