Identifying compound-target associations by combining bioactivity profile similarity search and public databases...

Area of Science:

• Computational chemistry
• Drug discovery and development
• Bioinformatics

Background:

• Molecular target identification is crucial for advancing drug discovery.
• Existing methods may not efficiently associate novel compounds with their targets.
• Understanding compound-target interactions is key to drug development.

Purpose of the Study:

• To introduce a computational approach, bioactivity profile similarity search (BASS), for associating molecular targets with small molecules.
• To validate the efficacy of BASS using a comprehensive bioactivity profile database.
• To explore BASS's potential in identifying targets of uncharacterized compounds and enabling scaffold hopping.

Main Methods:

• Developed the bioactivity profile similarity search (BASS) computational approach.
• Constructed a bioactivity profile database using 4296 compounds from the NCI-60 screen.
• Defined neighbor compounds based on bioactivity profile similarity (threshold > 0.75).
• Linked potential targets to neighbor compounds using known target annotations of query compounds.

Main Results:

• Successfully predicted compound-target associations, with approximately 45% verified retrospectively.
• Demonstrated BASS's capability to identify structurally diverse compounds with similar bioactivities.
• Highlighted the method's potential for identifying targets of uncharacterized molecules.

Conclusions:

• BASS is a viable computational tool for molecular target identification in drug discovery.
• The approach offers insights into compound promiscuity and polypharmacology.
• BASS facilitates scaffold hopping, aiding in the search for novel molecules against specific targets.