Page Comparison

Comparing the coverage between “Whole QCA TD datasets” Vs “Gen 1 + Gen 2 TD datasets”, without any overlap with the Lim-Mobley Benchmark molecules.

On top of Gen 1 + Gen 2 datasets, the following datasets seems to provide good coverage for TIG* parameters

Fragment stability benchmark
OpenFF-benchmark-ligand-fragments-v1.0
OpenFF Substituted Phenyl Set 1

Questions to answer:

What would be a good number of points to have for each interpolated torsion parameter ?
1. Do the slope changes we see with addition of data really matter ?
Is it okay if there is gap between the transitions from single bond to double bond, like in TIG3 plot below ?
Is this the right way to check for coverage?

Code Block

Fragment Stability Benchmark
OpenFF Fragmenter Validation 1.0
OpenFF Full TorsionDrive Benchmark 1
OpenFF Gen 2 Torsion Set 1 Roche 2
OpenFF Gen 2 Torsion Set 2 Coverage 2
OpenFF Gen 2 Torsion Set 3 Pfizer Discrepancy 2
OpenFF Gen 2 Torsion Set 4 eMolecules Discrepancy 2
OpenFF Gen 2 Torsion Set 5 Bayer 2
OpenFF Gen 2 Torsion Set 6 Supplemental 2
OpenFF Group1 Torsions
OpenFF Group1 Torsions 2
OpenFF Group1 Torsions 3
OpenFF Primary Benchmark 1 Torsion Set
OpenFF Primary Benchmark 2 Torsion Set
OpenFF Primary TorsionDrive Benchmark 1
OpenFF Rowley Biaryl v1.0
OpenFF Substituted Phenyl Set 1
OpenFF-benchmark-ligand-fragments-v1.0
Pfizer Discrepancy Torsion Dataset 1
SMIRNOFF Coverage Torsion Set 1
TorsionDrive Paper

Code Block

OpenFF Gen 2 Torsion Set 1 Roche 2
OpenFF Gen 2 Torsion Set 2 Coverage 2
OpenFF Gen 2 Torsion Set 3 Pfizer Discrepancy 2
OpenFF Gen 2 Torsion Set 4 eMolecules Discrepancy 2
OpenFF Gen 2 Torsion Set 5 Bayer 2
OpenFF Gen 2 Torsion Set 6 Supplemental 2
OpenFF Group1 Torsions
OpenFF Group1 Torsions 2
OpenFF Group1 Torsions 3

TIG1c