Systematic, computational discovery of multicomponent reactions and one-pot sequences.

Discovery of new types of reactions is essential to organic chemistry because it expands the scope of molecular scaffolds that can be made and can help trace new and more economical syntheses of existing structures1-4. Within the spectrum of all possible reaction classes, those that build complex scaffolds from multiple simple components in one step (i.e., multicomponent reactions, MCRs5-11; Figure 1a) and/or proceed sequentially in one pot12-14 are appealing as they may capitalize on the use of unstable intermediates, minimize separation and purification operations, and increase the overall step- and atom-economy15 as well as “greenness”16,17 of synthesis. However, the number of known MCR classes remains limited to several hundred (Figure 1b and interactive map at https://mcrmap.allchemy.net), and the number of publications reporting new ones appears to decrease in recent years (Figure 1c), perhaps because the most popular reactivity patterns (e.g., isocyanide, β-dicarbonyl, or imine-based MCRs) and their straightforward extensions and combinations18 have been studied in nearly exhaustive detail. The discovery process continues to rely, per Ivar Ugi’s famous quote, on thoughtful “reflection […] or combinatorial techniques,”5 and even when supported by high-end quantum mechanical methods, requires prior knowledge of “fitting” substrates (e.g., recent ref. 19 vs. closely analogous MCRs from earlier ref. 20,21). The “reflection” regards not only the choice of substrates but, above all, careful consideration of possible cross-reactivities between numerous intermediates, reagents, and side-/by-products present in a multicomponent mixture – in effect, translating into a laborious analysis of complex networks of mechanistic steps. Here, we show that computers equipped with accurate and broad knowledge of mechanistic transforms can perform such network analyses rapidly and in a high-throughput manner, and can guide systematic discovery, ranking, and yield estimation of unprecedented types of MCRs, one-pot sequences and even organocatalytic reactions, several of which we validate by experiment. These results evidence that synthesis-planning algorithms are no longer limited to skillful manipulation of the existing knowledge-base of “full reactions”22-28 but can assist in its creative expansion.