Clustering a large number of compounds. 1. Establishing the method on an initial sample

Area of Science:

• Drug discovery and development
• Computational chemistry
• Cheminformatics

Background:

• The National Cancer Institute's Division of Cancer Treatment has updated its drug-screening program.
• A repository of approximately 230,000 compounds is available for screening under the new protocol.

Purpose of the Study:

• To develop and establish a clustering method for extracting a representative sample from a large compound repository.
• To apply this method to an initial sample of 4980 compounds for drug screening optimization.

Main Methods:

• A clustering algorithm was implemented to group compounds based on molecular fragments.
• The method was established using an initial dataset of 4980 compounds.
• Complex molecular fragments were utilized to differentiate a large number of compounds.

Main Results:

• The study successfully established a clustering method for a large-scale compound library.
• The chosen molecular fragments proved effective in distinguishing numerous compounds within the sample.

Conclusions:

• The developed clustering method provides a viable approach for selecting representative compound samples for drug screening.
• This methodology can aid in optimizing the efficiency and scope of cancer drug discovery programs.