fmcsR: mismatch tolerant maximum common substructure searching in R

Area of Science:

• Computational chemistry
• Cheminformatics
• Drug discovery

Background:

• Accurate structural similarity measurement is crucial for drug discovery and chemical genomics.
• Existing maximum common substructure (MCS) methods are accurate but limited by rigid matching policies.
• A new mismatch-tolerant search method for flexible MCSs (FMCSs) is introduced to overcome these limitations.

Purpose of the Study:

• To develop and evaluate a novel FMCS algorithm for improved small molecule similarity analysis.
• To enhance the accuracy and efficiency of virtual screening processes.
• To provide a user-friendly R package (fmcsR) for FMCS computation.

Main Methods:

• Implementation of the FMCS algorithm with time-consuming steps in C++ and an R interface (fmcsR).
• Inclusion of utilities for pairwise compound comparisons, similarity searching, clustering, and visualization.
• Evaluation of FMCS performance against existing MCS tools and other virtual screening methods.

Main Results:

• The fmcsR package demonstrates superior time performance compared to existing MCS tools.
• Enabling mismatch tolerance in FMCSs results in larger common substructures and improved enrichment of active compounds in similarity searches.
• FMCS outperforms seven other virtual screening methods in both overall and early enrichment performance.

Conclusions:

• The FMCS approach significantly enhances the effectiveness of similarity searching for drug discovery.
• The fmcsR package offers an efficient and flexible tool for analyzing small molecule structural similarities.
• Flexible matching in MCS algorithms is a valuable strategy for improving virtual screening and identifying potential drug candidates.