Machine Learning in Computer-Aided Drug Design Volume II

This Research topic is a second volume of the successful Research Topic “Machine Learning in Computer-Aided Drug Design”. Please see the first Volume here.Machine learning (ML) has become an indispensable tool in modern science, transforming research in structural biology, molecular dynamics, computational chemistry, and drug discovery. Advances in deep learning (DL), including transformer-based architectures and denoising diffusion probabilistic models (DDPMs), have enabled major breakthroughs such as highly accurate protein structure prediction and generative molecular design.Despite these advances, important challenges remain. While ML models can achieve remarkable predictive performance, interpreting how they arrive at their predictions is often difficult. Moreover, model performance depends strongly on the quality and diversity of the training data, making the integration of ML with domain expertise and experimental validation essential for developing robust, interpretable, and biologically meaningful models. The greatest impact of ML in drug discovery will come from approaches that combine computational prediction with mechanistic understanding.In computer-aided drug design (CADD), ML has demonstrated considerable value in predicting ligand binding sites and binding modes, accelerating molecular dynamics (MD) simulations, improving quantum mechanical/molecular mechanical (QM/MM) calculations, optimizing lead compounds, and supporting virtual screening. Beyond improving efficiency and reducing development costs, ML also offers opportunities to generate new biological hypotheses and deepen our understanding of molecular recognition, cellular signaling, and biomolecular dynamics.For this second volume, we invite contributions that apply state-of-the-art ML methodologies while providing rigorous biological, chemical, or physical interpretation of the results. Topics of interest include, but are not limited to:• ML approaches for accelerating molecular dynamics simulations;• Implicit solvent models and force field development;• Discovery and mechanistic interpretation of cryptic binding pockets;• Prediction and characterization of allosteric regulation;• Applications to challenging or traditionally undruggable targets;• Toxicity prediction and safety assessment;• Integration of structural, biochemical, and clinical datasets;• ML applications to chromatin organization, genome architecture, chromosome conformation capture (3C)-based techniques, fluorescence in situ hybridization (FISH), chromatin polymer models, and epigenetic modifications.We particularly encourage submissions that go beyond benchmarking or comparing existing ML algorithms. Studies that combine methodological innovation with mechanistic interpretation, theoretical insight, or experimental validation are especially welcome. Such contributions have the potential to advance computer-aided drug design, molecular biosciences, and the development of more interpretable and trustworthy machine learning methods. Topic editors: Ognjen Perisic, Dr.SUGANDHA SAXENA, Stefan Ivanov Source: https://www.frontiersin.org/research-topics/83457/machine-learning-in-computer-aided-drug-design-volume-ii