6 0 4 | N A T U R E | V O L 5 5 5 | 2 9 m A R c h 2 0 1 8 ARTicLE doi:10.1038/nature25978 Planning chemical syntheses with deep neural networks and symbolic AI marwin h. S. Segler1,2, mike Preuss3 & mark P. Waller4 Retrosynthetic analysis is the canonical technique used to plan the synthesis of small organic molecules1,2. In retrosynthesis, a search tree is built by ‘working backwards’, analysing molecules recursively and transforming them into simpler precursors until one obtains a set of known or commercially available building-block molecules (Fig. 1)3,4. Given that transformations are formally reversed chemical reactions, the plan can be then carried out in the laboratory in the forward direc- tion to synthesize the target compound3,4. Transformations are derived from successfully conducted series of similar reactions with analogous starting materials, and are often named after their discoverers (‘named reactions’)5. At each retrosynthetic step, a
