Аннотация:Virtual screening of vendor catalogues followed by experimental assessment of hits is a commonly accepted practice of lead discovery in academic medicinal chemistry. While ZINC database comprises a large chemical space covered by numerous vendors, a majority of these compounds is not available for a large-scale procurement from Russian Federation. Consequently, only a few vendors are easily available for Russian scientists eagering to discover new relevant chemotypes. We have performed several virtual screens for potential antivirals in the catalogues of different Russia-based vendors, and the hits were further evaluated in phenotypic and target-based experimental assays. Hit rates in these assays were strikingly different, from ~30% for anti-flavivirus activity to ~1% for protease inhibition [1-3]. Since the methodology and the retrospective validation of the virtual screens gave a hint for their relevance, we supposed that variation of the hit rates could be inferred from the catalogue composition, because no good hits could be found in a catalogue that does not contain relevant molecules.In the current study, we present the property-based and chemical space-based analysis of the catalogues of the Russian vendors and compare them with our training sets as well as with the other databases, such as ChEMBL and databases of the approved drugs. Based on this analysis, we make conclusions on the applicability of these vendor catalogues for wide scale lead identification.This study was supported by the State Research Funding for FSASI “Chumakov FSC R&D IBP RAS” (Institute of Poliomyelitis) from the Ministry of Science and Higher Education of the Russian Federation,project no. FNZG-2024-0005.