Computational Strategy Revealing the Structural Determinant of Ligand Selectivity towards Highly Similar Protein Targets

Background: Poor selectivity of drug candidates may lead to toxicity and side effects accounting for as high as 60% failure rate, thus, the selectivity is consistently significant and challenging for drug discovery. Objective: To find highly specific small molecules towards very similar protein targets, multiple strategies are always employed, including (1) To make use of the diverse shape of binding pocket to avoid steric bump; (2) To increase binding affinities for favorite residues; (3) To achieve selectivity through allosteric regulation of target; (4) To stabalize the inactive conformation of protein target and (5) To occupy dual binding pockets of single target. Conclusion: In this review, we summarize computational strategies along with examples of their successful applications in designing selective ligands, with the aim to provide insights into everdiversifying drug development practice and inspire medicinal chemists to utilize computational strategies to avoid potential side effects due to low selectivity of ligands.