Leveraging AlphaFold2 structural space exploration for generating drug target structures in structure-based virtual screening

TL;DR

This paper introduces a method to improve drug discovery by modifying AlphaFold2 protein structures to better suit virtual screening.

Contribution

A novel approach using MSA mutations and optimization strategies to generate drug-target structures for virtual screening.

Findings

Genetic algorithms enhance virtual screening accuracy when active compounds are abundant.

Random search is more effective with limited active compound data.

Modified AlphaFold2 structures outperform PDB-derived structures for poor screening targets.

Abstract

Computational virtual screening (VS) plays a vital role in early-stage drug discovery by enabling the efficient selection of candidate compounds and reducing associated costs. However, the absence of experimentally determined three-dimensional protein structures often limits the applicability of structure-based VS. Advances in protein structure prediction, notably AlphaFold2, have begun to address this gap. Yet, studies indicate that direct use of AlphaFold2-predicted structures often leads to suboptimal VS performance—likely because these structures fail to capture ligand-induced conformational changes (apo-to-holo transitions). To overcome this, we propose an approach that explores and modifies the structural space of AlphaFold2 predictions to generate conformations more amenable to VS. Our method deliberately alters the multiple sequence alignment (MSA) by introducing alanine mutations at key residues in the ligand-binding site, thereby inducing significant conformational shifts. The exploration process is guided by iterative ligand docking simulations, with mutation strategies optimized either by a genetic algorithm or via random search. Our evaluation shows that when sufficient active compounds are available, the genetic algorithm significantly enhances VS accuracy. In contrast, with limited active compound data, a random search strategy proves more effective. Moreover, our approach is particularly promising for targets that yield poor screening results when using experimentally determined structures from the PDB. Overall, these findings underscore the practical utility of modified AlphaFold2-derived structures in VS and expand the potential of computationally predicted protein models in drug discovery.

•Modified protein alignments yield drug-friendly AlphaFold2 structures.•Genetic search optimizes predicted structures for improved virtual screening.•Enhanced screening performance seen for targets with poor PDB data.•Method works for new targets and compares favorably with AlphaFold3.

Modified protein alignments yield drug-friendly AlphaFold2 structures.

Genetic search optimizes predicted structures for improved virtual screening.

Enhanced screening performance seen for targets with poor PDB data.

Method works for new targets and compares favorably with AlphaFold3.

Figures