Sequence-based virtual screening using transformers

Ligand-Transformer is a deep learning framework that predicts protein–ligand binding and conformational changes using only sequence and 2D-structure inputs, enabling efficient virtual screening and uncovering mechanisms of drug action.