New strategy for assessing the applicability of reactions

Researchers from the University of Münster developed a computer-assisted method to objectively select model substrates for evaluating new chemical reactions. This approach aims to improve the quality of chemical reaction data by considering the complexity and structural properties of real pharmaceutical compounds, thus addressing the subjective bias in substrate selection. By utilizing molecular fingerprints and machine-learning models, they can predict the applicability of reactions across a wide range of substrates without bias, facilitating better data use in both academic and industrial contexts. For more information, please continue reading the article below: