We are happy to share another publication co-authored by Chemspace with you!
The goal of this work was to a novel graph neural network architecture for chemical yield prediction. This can be super useful for synthesis planning to efficiently score synthesis routes, saving time and reagents. The ML models in this work use structural information about participants of the transformation as well as molecular and reaction-level descriptors to separate zero- and non-zero yielding reactions.
You can find the link to the full article here!
We also want to thank all the collaborators for such productive work!
Dzvenymyra Yarish, Sofiya Garkot, Oleksandr O. Grygorenko, Dmytro S. Radchenko, Yurii S. Moroz, Oleksandr Gurbych
The approach described in the paper is designed to help obtain compounds with a high synthesis rate. Meanwhile, we at Chemspace have already delivered a product with all the necessary features - Freedom Space! This product combines well-established reactions on a selected set of Building Blocks to create a collection of 201M compounds for your projects, that can be delivered very fast.
