Additional file 3: of PubChem atom environments

PubChem atom types. This file contains all atom types generated in this study for PubChem Substance and PubChem Compound in SMARTS format. Each SMARTS string is followed by a semicolon-separated incidence vector that specifies in how many records in which repository the particular atom type was found. Order of repositories is PubChem Substance followed by PubChem Compound. File is sorted by lexicographical order of SMARTS strings.