ALiBERO: evolving a team of complementary pocket conformations rather than a single leader

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Accurate ligand binding in docking and virtual screening (VS) requires receptors to adopt specific conformational changes.
• Experimental structures and homology models often fail to capture these crucial receptor dynamics, limiting VS performance.
• Previous work introduced Ligand-guided Backbone Ensemble Receptor Optimization (LiBERO) to improve pocket selection for VS.

Purpose of the Study:

• To introduce ALiBERO, an advanced computational tool for optimizing receptor pockets for enhanced virtual screening.
• To expand upon the LiBERO method by enabling the selection of multiple receptor pockets, not just single ones.
• To automate the iterative sampling-selection procedure for pocket optimization.

Main Methods:

• ALiBERO employs a dual pocket selection strategy combining exhaustive combinatorial search with incremental addition of pockets.
• The method iteratively selects pockets that maximize the discrimination between known active compounds and decoys.
• The human estrogen receptor alpha was used as a reference case for algorithm design and implementation.

Main Results:

• Optimized pockets generated by ALiBERO demonstrated significantly improved VS performance.
• The enhanced VS performance was validated using large, independent test sets of active and inactive ligands.
• ALiBERO effectively improves the ability to identify biologically relevant ligands.

Conclusions:

• ALiBERO represents a significant advancement in computational methods for drug discovery.
• The tool enhances virtual screening accuracy by optimizing receptor pocket conformations.
• ALiBERO facilitates more effective identification of potential drug candidates through improved ligand-receptor interaction prediction.