mRAISE: an alternative algorithmic approach to ligand-based virtual screening.

Area of Science:

• Computational Chemistry and Cheminformatics
• Drug Discovery and Medicinal Chemistry

Background:

• Ligand-based virtual screening (LBVS) is crucial for identifying novel lead molecules in drug discovery.
• Existing LBVS methods face challenges in result reliability, biologically relevant molecular alignments, and screening efficiency.
• There is a need for robust LBVS tools that can be practically applied in daily drug discovery workflows.

Purpose of the Study:

• To introduce and evaluate a novel ligand-based virtual screening method, mRAISE.
• To assess mRAISE's performance in terms of ranking accuracy and alignment quality compared to state-of-the-art methods.
• To provide a new dataset for evaluating the quality of molecular alignments generated by LBVS methods.

Main Methods:

• mRAISE utilizes structural alignments combined with a descriptor-based bitmap search engine (RAISE) for efficient screening.
• On-the-fly generated alignments are scored using an independent shape-based scoring function for compound ranking.
• Performance was evaluated using the Directory of Useful Decoys dataset and a newly introduced dataset for alignment quality assessment.

Main Results:

• mRAISE achieved an average Area Under the ROC Curve of 0.76 and an average enrichment factor of 20.2 at 1% on the Directory of Useful Decoys dataset.
• The method consistently ranks among top-performing LBVS tools.
• For alignment quality, mRAISE achieved a root-mean-square deviation < 2.0 Å for 80.8% of alignment pairs within the top ten ranked conformations.

Conclusions:

• mRAISE offers a reliable and efficient solution for ligand-based virtual screening in drug discovery.
• The method provides high-quality molecular alignments suitable for medicinal chemistry investigation.
• The introduced dataset and mRAISE software are valuable resources for the research community.