TheJacksonLab/OMG_PhysicalProperties: OMG-Property-Database

Monomer-level properties for 12M synthetically accessible polymers in the Open Macromolecular Genome derived via quantum chemistry.

[1] Duplicated data in QM calculated OMG monomers (Feb 14, 2025)

Please apply "drop_duplicates" for "methyl_terminated_product" before constructing a dataset for your purpose. 

There are 10 duplicated methyl terminated monomers among 32,529 QM-calculated OMG methyl terminated geometries (pareto greedy) and 5 duplicates among 15,147 (test). There are 15 overlapping methyl terminated monomers between the train and test dataset. Regardless of duplicates in "methyl_terminated_product", these duplicates do not have a significant impact on the active learning test performance.

There are no duplicates in "product" (ended with asterisks to indicate the repeating points), but some duplicates in "methyl terminated monomers" (asterisks replaced by methyl groups). This can happen when added methyl groups make different repeating units have the same methyl terminated structure.