In drug-discovery, a large amount of experimental data is generated to characterize potential drugs, ranging across chemistry and biological disciplines.  As a project progresses, this data can be used to build statistical models that can help predict data for a molecule design before it is even synthesized.   We have developed quantitative structure-activity relationship (QSAR) statistical models and applied them in compound progression of small molecule therapeutic targets. Both software integration and ease in the evaluation of model performance have guided decisions in synthesis and assay progression. A probabilistic approach has shown to be easy to interpret and to apply in compound prioritization. Case studies are shown where intelligent use of model predictions has impacted therapeutic projects saving resources and increasing the overall probability of success.

Case studies in the use of machine learning models applied to small molecule therapeutic targets

2019 Fall National Meeting

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