We are presenting an analysis of forty databases ranging in size from a few thousand to nearly 100 million molecules, comprising over 200 million structures in total, for their tautomeric conflicts based on 119 tautomeric transform rules expressed as SMIRKS. A tautomeric conflict is defined as an occurrence of two or more structures within a data set that were identified by this rule set as being tautomers of each other. Out of these rules, 79 yielded at least one conflict. The databases analyzed spanned a wide variety of types including experimentally based structure collections, drug collections, large aggregating databases such as PubChem, and commercial catalogs. Almost all databases analyzed showed intra-database tautomeric conflicts. The conflict rates were typically in the range of a few tenths of a percent of the database, which for the largest collections yielded more than 100,000 cases per database. While not all such conflicts are data quality issues in large chemical information resources, many may point to insufficient tautomeric normalization of structures.

Tautomeric conflicts in small to large small-molecule databases

ACS Spring 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.