Network search algorithms and scoring functionsfor advanced-level computerized synthesis planning

Abstract

In 2020, a“hybrid”expert-AI computer program called Chematica (a.k.a. Synthia)was shown to autonomously plan multistep syntheses of complex natural prod-ucts, which remain outside the reach of purely data-driven AI programs. Theability to plan at this level of chemical sophistication has been attributed mainlyto the superior quality of Chematica's reactions rules. However, rules alone arenot sufficient for advanced synthetic planning which also requires appropriatelycrafted algorithms with which to intelligently navigate the enormous networksof synthetic possibilities, score the synthetic positions encountered, and rankthe pathways identified. Chematica's algorithms are distinct fromprêt-à-porteralgorithmic solutions and are product of multiple rounds of improvements,against target structures of increasing complexity. Since descriptions of theseimprovements have been scattered among several of our prior publications, theaim of the current Review is to narrate the development process in a morecomprehensive manner.