Dear community,
I have a question that bothers me for a long time. Is there are MD-specific repositories, that automatically generate metadata and might be considered as good for FAIR sharing? (For now Zenodo and NOMAD are considered, but first is general purpose and second seems to be concentrating on QM data).
If there are no such repositories, what would be a minimal set of data, that would be considered as FAIR data set to share? Which files from raw data should be shared and which should be additionally generated?
Hi,
There is no MD-specific repo, some uses zenodo, other uses local alternative, other github (but it works only to input data due to size limit of the repo), etc. But a lot of discussions around the topic, but not a final standard.
Here some thought (that may help to start a discussion here)
- provide all input files (to allow reproducibility)
- provide compact version of output files
- provide scripts that allow reproducibility of the results
- provide a README file where you describe the content of your repo, including information on code (and version) and force field is used, reference to the papers or simular where the simulation results are published,
- eventually license for use.
\Alessandra