To plan the syntheses of small organic molecules, chemists use retrosynthesis, a problem-solving technique in which target molecules are recursively transformed into increasingly simpler precursors. Computer-aided retrosynthesis would be a valuable tool but at present it is slow and provides results of unsatisfactory quality. Here we use Monte Carlo tree search and symbolic artificial intelligence (AI) to discover retrosynthetic routes. We combined Monte Carlo tree search with an expansion policy network that guides the search, and a filter network to pre-select the most promising retrosynthetic steps. These deep neural networks were trained on essentially all reactions ever published in organic chemistry. Our system solves for almost twice as many molecules, thirty times faster than the traditional computer-aided search method, which is based on extracted rules and hand-designed heuristics. In a double-blind AB test, chemists on average considered our computer-generated routes to be equivalent to reported literature routes.

Planning chemical syntheses with deep neural networks and symbolic AI

The field of total synthesis has a rich history and a vibrant future. Landmark advances and revolutionary strides in the logic of synthesis have put the practicing chemist in the enviable position of being able to create nearly any molecule with enough time and effort. The stage is now set for organic chemists to aim for "ideality" in the way molecules are synthesized. This perspective presents a simple and informative definition of "ideality" and demonstrates its use during the self-evaluation of several syntheses from our laboratory.

Aiming for the Ideal Synthesis

The Green ChemisTREE: 20 years after taking root with the 12 principles

Driven by remarkable advances in the understanding of structure and reaction mechanisms, organic synthesis will be increasingly directed to producing bioinspired and newly designed molecules.

Synthesis at the molecular frontier

The article discusses the mechanisms and the applications in organic synthesis, materials, supramolecular, and polymer synthesis of most organic reactions mediated by single electron transfer. Each reaction or class of reactions will be discussed by starting with the original discovery publication, followed by a summary of all or most review articles published in the field, and a discussion of the mechanism(s) and of the most important methodologic and synthetic developments since the most recent review was published. The mechanisms of most organic reactions are considered to proceed by two-electron transfer pathways, even though both biology and radical chemistry rely extensively on one-electron transfer processes. Radicals generated by homolytic cleavage at high temperature were traditionally employed in the industrial production of polymers and to a lesser extent in the synthesis of organic molecules.

Single electron transfer in radical ion and radical-mediated organic, materials and polymer synthesis.

I. Foreword by E. J. Corey II Introduction by K. C. Nicolaou III. Preface IV. Explanation of the Use of Colors in the Schemes and Text V. List of Abbreviations VI. List of Named Organic Reactions (this main section includes 250 two-page entries) VII. Named Organic Reactions in Alphabetical Order VIII. Appendix: Listing of the Named Reactions by Synthetic Type and by their Utility IX. References X. Index

Strategic Applications of Named Reactions in Organic Synthesis: Background and Detailed Mechanisms

Molecules and Medicine

The discovery that cortistatins A and J show noteworthy antiangiogenic activity prompted an investigation of the possibility that simpler and much more easily made compounds based on a steroid core might have useful bioactivity. These studies have led to the development of several potent, water-soluble compounds that may be suitable for local application to treat ocular wet macular degeneration, an important cause of blindness, as well as for treatment of various other angiogenesis-dependent diseases. One of these substances was tested in a mouse retinal angiogenesis model and found to inhibit angiogenesis at a locally administered dose of 500 pmol. Comparison of cell migration data for this and two other synthetic compounds with published data on cortistatin A indicate that they inhibit vascular endothelial growth factor-induced cell migration of human umbilical vein endothelial cells more strongly than cortistatin A.

Barbara Czako

Papers

Strategic Applications of Named Reactions in Organic Synthesis: Background and Detailed Mechanisms

Molecules and Medicine

Discovery of potent and practical antiangiogenic agents inspired by cortistatin A.

A short, scalable synthesis of the carbocyclic core of the anti-angiogenic cortistatins from (+)-estrone by B-ring expansion.

Discovery of Potent and Practical Antiangiogenic Agents Inspired by Cortistatin A.