(532d) Enhanced Process Characterization through Data-Rich Experimentation and Process Modeling: Generating an Intermediate Towards Islatravir

Tight development timelines and increasing expectations around process knowledge at time of filing have driven investments in data-rich experimentation tools aimed to maximize knowledge capture from each experiment. Commonly reaction characterization is accomplished through design of experiment methodologies coupled with statistical analyses and response surface modeling. Typically, however, these statistical analyses are restricted to single timepoints. Recently, automated reactor and sampling technologies have enabled scientists to capture temporal data throughout each experiment. This greater data density has enabled the generation of data-driven models through techniques such as dynamic surface response methodology (DRSM) and the leveraging of these models to improve process robustness and efficiency [[1]].

Herein, we present the use of reaction and sampling automation to generate an extensive reaction dataset during the completion of a design of experiments as part of process characterization of one of the penultimate intermediates in the synthesis of islatravir, a late-stage drug substance candidate. Data-driven models generated from this dataset were used to inform reaction stability and drive process decisions.

[[1]] J. Jurica, J. McMullen. 2021. Org. Proc. Res. Dev., 25, 2, 282-291