ChemAIRS R&D never stops

ChemAIRS® retrosynthesis system accelerates drug discovery

ChemAIRS®, a leading AI-aided synthetic route design system, can provide multiple synthetic routes for unknown molecules in a short period of time and provide chemists with different synthetic ideas, thus to greatly improve route design efficiency and shorten the discovery cycle of drugs, particularly when encountered with complex molecules or molecules with new scaffolds, of which even experienced chemists need some time to review the relevant literature to come up with a reasonable route. If the experiment fails afterwards, the route needs to be redesigned. While ChemAIRS® has strong advantages of big data and computational calculation power over human brains, it supports and provides human chemists with practical synthetic ideas with diversity, thus enhancing the efficiency and speeding up the drug discovery cycle with a difference.

Single-step reaction to test ChemAIRS® recognition capability

With the continuous development of ChemAIRS®, the software functionality and route prediction capability have been significantly improved. The main improvements in route prediction include the filtering of infeasible single-step reactions and the enhancement of the rationality of covalent bond cut-off.

ChemAIRS® recognition capability of single-step reaction feasibility was tested by using 100 single-step feasible reactions and 100 single-step infeasible reactions provided by chemists. We ran 3 versions of ChemAIRS®, the one released in March, June and September this year, to see its improvements. The result showed that the passing ratios of 100 single-step feasible reactions in March, June and September versions were 53%, 63% and 66% respectively. When tested with 100 single-step infeasible reactions, the results showed that the filtering ratios were 41%, 49% and 71% for the March, June and September versions respectively. As shown in the results, the recognition capability of single-step reaction feasibility has been improved significantly with ChemAIRS®, leading to better feasibility in route design.

Four examples of single-step infeasible reactions are shown, all of which were effectively identified and successfully filtered in the September version of ChemAIRS®. Reaction a and b have selectivity problems, while reaction c has a chiral conformation problem and reaction d has a nucleophilic substitution site problem. These are the reactions that chemists would avoid in route design.

ChemAIRS® route prediction of unknown molecules

Take the unknown molecule which contains a fused ring and a spiro ring structure shown as an example. Since the molecular skeleton is not very common, it is difficult to give a synthetic route without the assistance of relevant databases. Therefore, chemists need to try a variety of molecular cut-off ways with the reference of a lot of literature when designing this molecule. While using ChemAIRS®, multiple synthetic routes can be quickly (usually within 5 min) provided. The following is the top-ranked route shown in the system, from which each of the step is well designed and supported by literature. Besides, the starting materials are commercially available. In this route, the step 3a+3b→5b and 5a+5b→6b are effective in building fused ring and spiro ring. For details of the specific literature referenced: MedChemComm; vol. 5; 7; (2014); p. 973–983 and Bioorganic and Medicinal Chemistry Letters; vol. 18; 19; (2008); p. 5259–5262.

With this unknown molecule, an experienced chemist requires over an hour to come up with a successful route, by searching a large amount of literature. Compared to an experienced human chemist, ChemAIRS® demonstrates similar capability to design synthetic routes.

Wanna try ChemAIRS yourself? please go to website https://www.chemical.ai/ and register with a promotion code CF41MT to redeem your one-month free trial.

About Chemical.AI

Founded in 2018, Chemical.AI is one of the technology leading companies leveraging Artificial Intelligence (AI) and big data to transform research and development to shape the future of chemistry. ChemAIRS® boosts success rates in synthesis route design and prediction with diverse synthesis strategies. In 2021, Chemical.AI also established an AIoT automation lab in Shanghai to bring dry lab and wet lab together to provide more professional and efficient chemical synthesis services to the industry. For more information, please go to https://chemical.ai/

For inquires, please contact bd@chemical.ai