Preparing anti-SARS-CoV-2 agent EIDD-2801 by a practical and scalable approach, and quick evaluation machine learning.

EIDD-2801 is an orally bioavailable prodrug, which will be applied for emergency use authorization from the U.S. Food and Drug Administration for the treatment of COVID-19. To investigate the optimal parameters, EIDD-2801 was optimized a four-step synthesis with high purity of 99.9%. The hydroxylamination procedure was telescoped in a one-pot and the final step was precisely controlled on reagents, temperature and reaction time. Compared to the original route, the yield of the new route was enhanced from 17% to 58% without column chromatography. The optimized synthesis has been successfully determinated on a decagram scale: the first step at 200 g and the final step at 20 g. Besides, the relationship between yield and temperature, time, and reagents in the deprotection step was investigated Shapley value explanation and machine learning approach-decision tree method. The results revealed that reagents have the greatest impact on yield estimation, followed by the temperature.