Teaching Computer-Assisted Retrosynthesis Reaction Prediction with Open-Source Software

Over the past decade, numerous computer-assisted retrosynthesis reaction prediction methods have been reported in the chemical literature, several of which are also shared as open-source code. Reports of integrating these modern open-source retrosynthetic prediction techniques into the chemical education classroom, however, is largely absent. This article describes our efforts to teach a computer-assisted retrosynthesis workshop series using open-source reaction prediction software and methods. In addition to three workshops on selected open-source reaction prediction methods, the series included workshops on Python programming and introductory cheminformatics using the RDKit cheminformatics toolkit. A pre and post survey was distributed to better understand participant familiarity and workshop outcomes. Participants indicated that they had minimal experience with cheminformatics, and open-source retrosynthesis software, however after the workshop series they indicated increased familiarity, interest, and desire to pursue additional training related to open-source retrosynthesis software and methods. We conclude this article with our experiences and ideas for future improvements. The workshop materials and code are available with an open source license, and educators are encouraged to adapt and improve the lessons (https://github.com/UA-Libraries-Research-Data-Services/retrosynthesis).