Developing Novel Ligands with Binding Affinity Using Machine Learning
Yang Ha · Structural Dynamics · 2025
With the advance of machine learning, the process of designing ligand molecules with binding affinities has become increasingly efficient. In this study, we have developed an autoencoder to generate large amount of drug/ligand candidates from a seed molecule (SMILES string). Those candidates have similar structure or property with the given molecule. Further evaluation methods are then applied to further nail down the promising ones. This can greatly accelerate the process of drug development and assist scientists to study protein ligand interactions.