Link or store "known" computed results?

openforcefield / protein-ligand-benchmark

Protein-Ligand Benchmark Dataset for Free Energy Calculations

MIT License

150 stars 15 forks source link

Link or store "known" computed results? #40

Open IAlibay opened 2 years ago

IAlibay commented 2 years ago

Especially if this ends up being a per-release benchmark run, it would be good to have a simple manner for folks to work out what historical computed free energies were for a given edge based on a given OpenFF release.

Even if it's not stored here specifically, having a link to where it is stored would be great.

dotsdl commented 2 years ago

We are working to gather historical benchmark runs performed by @dfhahn at Janssen, in particular:

predicted deltaG
uncertainty in deltaG
software environment info
forcefield
version of protein-ligand-benchmark

We should decide on other required keys we need for downstream data.

dotsdl commented 2 years ago

@dfhahn has shared with us historical data.

The data can be found here: https://github.com/dfhahn/protein-ligand-benchmark-analysis

The zenodo corresponding Zenodo record: https://zenodo.org/record/6538976#.YocTYB7MJA4

The data can be found in 00_data (raw) or in benchmarkpl/20*/10_results/.

dotsdl commented 2 years ago

I will perform a research cycle in the existing data and report back here on how it is structured, what is present, possible issues or omissions, etc.