Anne Mai Wassermann1, Jürgen Bajorath. 1. Department of Life Science Informatics, B-IT, LIMES Program Unit Chemical Biology and Medicinal Chemistry, Rheinische Friedrich-Wilhelms-Universität, Dahlmannstr. 2, D-53113 Bonn, Germany.
Abstract
BACKGROUND: Bioisosteric replacements are commonly understood to be replacements of groups of atoms in bioactive compounds that retain their specific activity and retain, or further improve, compound potency. Such chemical modifications are of high interest in medicinal chemistry and are often considered in compound exploration and optimization. RESULTS: We have applied the matched molecular pair formalism to carry out a large-scale data mining study to identify bioisosteres in publicly available active compounds with similar potency. Our data mining effort has identified a set of 96 nonredundant bioisosteric replacements, approximately half of which were previously unobserved. However, a number of replacements commonly considered to be bioisosteric did not meet our extended bioisostere selection criteria, which included high frequency of occurrence, limited potency alterations and activity across different target families. Furthermore, many commonly known bioisosteric replacements were found to be dependent on the structural context in which they occurred. CONCLUSION: The systematic analysis of public domain compound data presented herein provides an alternative route to the identification of bioisosteric replacements and further extends the spectrum of currently known bioisosteres. We provide a compendium of bioisosteric replacements that are well supported by currently available compound data.
BACKGROUND: Bioisosteric replacements are commonly understood to be replacements of groups of atoms in bioactive compounds that retain their specific activity and retain, or further improve, compound potency. Such chemical modifications are of high interest in medicinal chemistry and are often considered in compound exploration and optimization. RESULTS: We have applied the matched molecular pair formalism to carry out a large-scale data mining study to identify bioisosteres in publicly available active compounds with similar potency. Our data mining effort has identified a set of 96 nonredundant bioisosteric replacements, approximately half of which were previously unobserved. However, a number of replacements commonly considered to be bioisosteric did not meet our extended bioisostere selection criteria, which included high frequency of occurrence, limited potency alterations and activity across different target families. Furthermore, many commonly known bioisosteric replacements were found to be dependent on the structural context in which they occurred. CONCLUSION: The systematic analysis of public domain compound data presented herein provides an alternative route to the identification of bioisosteric replacements and further extends the spectrum of currently known bioisosteres. We provide a compendium of bioisosteric replacements that are well supported by currently available compound data.
Authors: Anna Gaulton; Louisa J Bellis; A Patricia Bento; Jon Chambers; Mark Davies; Anne Hersey; Yvonne Light; Shaun McGlinchey; David Michalovich; Bissan Al-Lazikani; John P Overington Journal: Nucleic Acids Res Date: 2011-09-23 Impact factor: 16.971
Authors: Darren V S Green; Stephen Pickett; Chris Luscombe; Stefan Senger; David Marcus; Jamel Meslamani; David Brett; Adam Powell; Jonathan Masson Journal: J Comput Aided Mol Des Date: 2019-10-21 Impact factor: 3.686
Authors: Anthony R Bradley; Ian D Wall; Darren V S Green; Charlotte M Deane; Brian D Marsden Journal: J Chem Inf Model Date: 2014-10-09 Impact factor: 4.956