Hole filling and library optimization: application to commercially available fragment libraries

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Drug Discovery

Background:

• Compound libraries are crucial for drug discovery.
• Vast chemical space exists beyond in-house libraries.
• Efficiently expanding libraries into unexplored chemical regions is vital.

Purpose of the Study:

• To develop an automated method for filling holes in existing compound libraries.
• To apply this method to augment libraries with external compounds, specifically fragments.
• To ensure efficient exploration of chemical space without unnecessary library expansion.

Main Methods:

• An automated hole-filling (HF) method, Canvas HF, was developed.
• It utilizes 2D chemical fingerprint distances to identify and select compounds for underrepresented chemical space regions.
• An extension allows simultaneous consideration of chemical properties to bias compound selection.

Main Results:

• Canvas HF effectively fills chemical space 'holes' without selecting only peripheral compounds.
• The method demonstrates robustness across different databases and retrieval numbers.
• Performance was favorably compared to sphere exclusion and random selection methods.

Conclusions:

• Canvas HF provides an efficient and effective strategy for augmenting compound libraries from external sources.
• The parameter-free nature of the method enhances its broad applicability in hole-filling tasks.
• This approach aids in optimizing compound library diversity and coverage for drug discovery.