The availability of compounds is crucial through the discovery process for every industry (e.g. materials or drug discovery). Although various approaches have been developed recently, it seems that most approaches have limitations as the chemical space grows. In order to offer an alternative approach for molecule prioritization for researchers to retrosynthetic tools, in the presented Compound Price prediction Network (CoPriNet), authors focused on developing a graph neural network-based compound price prediction method. For the development of this novel approach, the Mcule team provided the curated database of purchasable compounds with required availability, price and other data. The novel approach could help speed up virtual screening in discovery processes such as drug or material discovery.
Open Access Article available at: https://doi.org/10.1039/D2DD00071G