Layered Molecular Editor for Automatic Construction of Chemical Structure Data Sets and Its Application.

Herein, the layered molecular editor (LME) method is developed to automatically construct chemical structures by modifying various moieties of molecules in chemical reactions, such as substituent groups. With an open application programming interface (API), LME could be widely applied in the screening of chemical reactions and the modification effect of substituent groups, integrating with other computational chemistry programs. By means of LME, data sets of Diels-Alder (DA) and Claisen rearrangement reactions with different substituent groups have been created from templates of substrates, and the relationships between their chemical activities and corresponding chemical structures are investigated by automatic geometric structure optimization and transition state searching using GFN2-xTB, iEIP, and machine learning methods. This shows the scaling ability and reliability of the developed LME method.