HierVLS hierarchical docking protocol for virtual ligand screening of large-molecule databases

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• High-throughput screening (HTS) requires efficient methods to analyze vast combinatorial chemistry libraries.
• Virtual ligand screening (VLS) is crucial for identifying potential drug candidates against specific protein targets.

Purpose of the Study:

• To develop HierVLS, a rapid hierarchical docking approach for VLS.
• To enhance the efficiency of identifying active compounds from large molecular libraries.

Main Methods:

• HierVLS employs a multi-level docking strategy, starting with a coarse-grained search and progressing to finer, more accurate energy calculations.
• The method optimizes ligand-protein-solvent interactions in a hierarchical manner.
• Each ligand is processed in approximately 4 minutes, significantly faster than previous methods.

Main Results:

• HierVLS accurately predicted the binding conformation for all 37 tested cocrystal structures.
• Calculated binding energies showed good correlation with experimental binding constants.
• In a library of over 10,000 molecules, HierVLS identified 26 of 37 known binders within the top 2% ranked by affinity.
• Screening a 55,000-compound library against protein-tyrosine phosphatase 1B (PTP1B) identified known binders, including the top-ranked compound.

Conclusions:

• HierVLS is an effective and efficient tool for virtual ligand screening.
• The method accelerates the identification of lead compounds from large combinatorial databases.
• Future work will address challenges with metal-containing sites and water-mediated interactions.