The aim of the Synthetically Accessible Virtual Inventory (SAVI) is to computationally generate a very large number of easily and cheaply synthesizable structures for use in drug discovery. While synthesizability, reliability and cost are the main driving criteria behind the generation of the SAVI library, we aim to include structures that possess a range of properties seen as desirable in current cutting-edge drug development. Through contributions from international collaborators, SAVI combines richly annotated chemical transforms from the LHASA project knowledgebase provided by maintainers, with the chemoinformatics toolkit CACTVS as well as custom code and hardware infrastructure for processing a large amount of chemical data. In addition to the procedure of the reversal of the original transforms for the forward-synthetic purposes of SAVI, the algorithm includes pre-computation of a number of properties relevant to drug discovery. Ultimately, the project aims at compiling a library containing one billion easily synthetically accessible screening samples, all annotated with one or more computer-proposed synthetic routes; as well as offering the user the ability to filter and search this set by a variety of structural and other criteria via GUI. As an update of the project we present the current progress, challenges and future direction.

Synthetically accessible virtual inventory (SAVI)

251st National Meeting (2016)

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