QMrxn20 Database ================ All geometries are stored as individual XYZ files in this folder structure: transition-states/rxn/label.xyz reactant-conformers/label/number.xyz reactant-complex-constrained-conformers/rxn/label/number.xyz reactant-complex-unconstrained-conformers/rxn/label/number.xyz product-conformers/rxn/label/number.xyz Folders ------- transition-states Transition states for a given reaction. reactant-conformers Reactants without Y, i.e. in the limit of infinite separation. reactant-complex-constrained-conformers Reactants with Y, after constrained geometry optimization. reactant-complex-unconstrained-conformers Reactants with Y, after unconstrained geometry optimization. product-conformers Products for both reactions, i.e. in the limit of infinite separation. Files ----- All energies available are given as a single list energies.txt with the columns Name Possible entries Example ----------------------------------------------------------------- label A_B_C_D_E_A reaction e2 | sn2 e2 geometry ts | r | rcc | rcu | pc rcu number 02 filename reactant-unconstrained-conformers/e2/A_B_C_D_E_A/02.xyz energy Hartree -237.194529 method mp2 | hf | lccsd mp2 and the following abbreviations ts Transition state r Reactant conformers rcc Reactant complex constrained conformers rcu Reactant complex unconstrained conformers pc Product conformers mp2 MP2/6-311G(d) hf HF/6-311G(d)//MP2/6-311G(d) lccsd DF-LCCSD/cc-pVTZ//MP2/6-311G(d) Based on this list and machine learning predictions, activation energies have been calculated in the file barriers.txt with the columns Name Possible entries Example ----------------------------------------------------------------- label A_B_C_D_E_A reaction e2 | sn2 e2 method mp2 | hf | lccsd mp2 reactant rcc | rcu rcu activation kcal/mol 42 *_ts same as energies.txt for transition state *_r same as energies.txt for reactant Note that rows without an entry for filename_r have their activation energy value inferred via ML as described in the paper.