BCL::MolAlign: Three-Dimensional Small Molecule Alignment for Pharmacophore Mapping.

Area of Science:

• Computational chemistry
• Cheminformatics
• Drug discovery

Background:

• Flexible molecular alignment is crucial for computer-aided drug discovery (CADD).
• Existing software for high-quality flexible alignment is limited.
• Property-based alignment offers a promising approach.

Purpose of the Study:

• Introduce BCL::MolAlign, a novel, freely available molecular alignment program.
• Evaluate BCL::MolAlign's performance in recovering native ligand poses.
• Assess the predictive power of BCL::MolAlign's alignment score for ligand activity.

Main Methods:

• BCL::MolAlign utilizes pregenerated conformers and on-the-fly bond rotation to handle ligand flexibility.
• Monte Carlo Metropolis sampling is employed to explore relative orientations of atom pairs.
• Performance is benchmarked against MOE, ROCS, FLEXS, maximum common substructure (MCS), and RosettaLigand on diverse datasets.

Main Results:

• BCL::MolAlign outperformed MOE, ROCS, and FLEXS in recovering native ligand poses across six datasets.
• The BCL::MolAlign alignment score showed higher predictivity of ligand activity than MCS similarity across 10 datasets.
• BCL::MolAlign recovered a larger fraction of native binding poses compared to MCS-based alignment and RosettaLigand on a benchmark set.

Conclusions:

• BCL::MolAlign is an effective and freely accessible tool for property-based flexible molecular alignment.
• The program demonstrates superior performance in pose recovery and activity prediction compared to existing methods.
• BCL::MolAlign represents a significant advancement for computer-aided drug discovery applications.