Reagent-based and product-based computational approaches in library design

• Published in: Current opinion in chemical biology Vol. 7; no. 3; pp. 326 - 330
• Main Author: Jamois, Eric A
• Format: Journal Article
• Language: English
• Published: England Elsevier Ltd 01.06.2003
• Subjects: Computational Biology - trends; Drug Design; Indicators and Reagents; Peptide Library; Structure-Activity Relationship
• ISSN: 1367-5931; 1879-0402
• DOI: 10.1016/S1367-5931(03)00053-X

The design of combinatorial libraries involves the consideration of all synthesizable compounds (the virtual library), followed by the selection of a suitably sized subset for actual synthesis and experimentation. Several approaches to this task can be envisaged, involving either reagent-based or product-based considerations. Reagent-based design considers the properties of the building blocks rather than those of the final products. Although popular with chemists, this approach overlooks the extent of chemical transformations involved in generating products. In effect, several important properties cannot be derived from building blocks alone and require access to product structures. Several studies have demonstrated the superiority of product-based designs in yielding diverse and representative subsets. Although more computationally intensive, the latter approach provides a basis for more sophisticated designs where reagent-based and product based considerations can be combined for a best-of-breed approach.