Fig. 2: High-throughput generation of experimental training data for reaction forward prediction models. | Nature Communications

Fig. 2: High-throughput generation of experimental training data for reaction forward prediction models.

From: Expediting hit-to-lead progression in drug discovery through reaction prediction and multi-dimensional optimization

Fig. 2

A Generalized Minisci reaction scheme highlighting the large chemical and reaction space. B Miniaturized reaction screening workflow utilized to generate the data set. C Matrix of possible fragment and carboxylic acid combinations including conducted screenings and their yield range. D Analysis of the binary reaction outcome. E Analysis of the yield across all reactions. F Example of the used screening plate highlighting the results from one particular fragment and carboxylic acid combination. Colors of the pie charts: Light blue: Mono-alkylation product, Grey: Starting material, Dark blue: Di-alkylation product, Dark grey: Side products. The percentage values represent the amount of observed mono-alkylated product determined by liquid chromatography-mass spectrometry (LCMS).

Back to article page