Members of MI-6’s research and development team are scheduled to give oral presentations at the 49th Symposium on Chemoinformatics, scheduled to take place on Tuesday 10–Wednesday 11 November 2026.
Presentation details
Day 1
AI-Guided Search for Low-Impact Reaction Solvents
- Presenter: Elisa Margarita MENDOZA ZAMARRIPA
- Co-authors: José Mauro VASQUEZ CASTILLO, Chia Hsiu CHEN, and Art Wei Yao Ang
- Presentation number: 1O15
- Session: Oral presentation session 3
- Session date and time: Tuesday 10 November 2026, 14:55–16:25 JST (scheduled)
Abstract
Solvents are essential to chemical manufacturing, but they account for a large share of its waste and environmental impact. This presentation introduces a workflow that uses generative AI, machine learning, and cheminformatics to propose low-impact solvent candidates and rank them by reaction performance and environmental impact.
In a Diels–Alder case study, the workflow identified a candidate predicted to perform nearly as well as acetic acid, a solvent known to give a fast rate for this reaction. The candidate’s environmental impact is estimated to be about 70% lower than that of acetic acid.
Day 2
Cost-to-Go Reranking of Reaction Candidates in Double-Ended Synthesis Planning
- Presenter: Hardik Tankaria
- Co-author: Yuan Weilin
- Presentation number: 2O11
- Session: Oral presentation session 7
- Session date and time: Wednesday 11 November 2026, 14:40–15:55 JST (scheduled)
Abstract
Computer-aided synthesis planning seeks a route from the starting materials a laboratory already holds to a target molecule. This presentation introduces double-ended synthesis planning, an approach that searches for synthesis routes from both the starting materials and the target molecule.
Conventional approaches generate and then discard many reactions a route needs, because each is judged on its own plausibility rather than on whether it leads towards the available starting materials.
The presentation introduces a method that reorders candidates using the planner’s own cost-to-go estimate, requiring no retraining. The planner then finds routes to more target molecules with less search effort, improving its ability to find synthesis routes while reducing running costs.
Event details
- Dates: Tuesday 10–Wednesday 11 November 2026
- Event: 49th Symposium on Chemoinformatics
- Venue: Tenpyo Hall, Nara Prefectural Convention Center
- Organiser: Division of Chemoinformatics, The Chemical Society of Japan
- Official website:49th Symposium on Chemoinformatics
Research and development at MI-6
MI-6 combines materials science with AI, data science, and robotics to advance and accelerate materials research and development. We develop and demonstrate new R&D methods by combining digital technologies, such as machine learning and computational science, with technologies for working with physical materials, such as autonomous synthesis and evaluation. Our aim is to address a wide range of R&D challenges, from materials discovery to process optimisation
