TSCC: Two-Stage Combinatorial Clustering for virtual screening using protein-ligand interactions and physicochemical...

Area of Science:

• Computational chemistry
• Biotechnology
• Drug discovery

Background:

• Databases of 3D compounds and protein complexes are crucial for biotechnology and drug development.
• Screening and retrieving suitable candidates, while managing false positives, remains a significant challenge in post-screening analysis.

Purpose of the Study:

• To develop an improved post-screening analysis method for virtual screening.
• To enhance the accuracy and efficiency of identifying therapeutic and industrial targets from large compound libraries.

Main Methods:

• A two-stage clustering (TSCC) method was employed for virtual screening analysis.
• The first stage involved clustering compounds based on protein-ligand interactions.
• The second stage utilized structure clustering based on physicochemical features to refine candidate selection.

Main Results:

• The TSCC method effectively clustered compounds based on atomic-level protein-ligand interactions and their strengths.
• Structure clustering grouped similar compounds, with the lowest energy representative selected from each cluster.
• This two-stage approach improved the specificity and accuracy of virtual screening analysis.

Conclusions:

• Representing interactions at the atomic level provides a more detailed description of protein-ligand binding.
• The two-stage clustering approach significantly enhances post-screening analysis, improving clustering, mining, and visualization of compound candidates.
• This method leads to better virtual screening enrichment and more effective identification of potential drug leads.