Performance when combining scores

[mglev1n]

When using pgsc_calc for more than a few scores, the pipeline seems to take a very long time. Despite using the --parallel argument and having access to HPC resources, the step that combines scores into a single scoring file seems to run serially (reading score1, performing QC, appending to scoring file; reading score2, performing QC, appending, ...). Is there a more efficient approach for creating a combined scoring file (either using the pgsc_calc pipeline, or preparing the combined scoring file prior to running the pipeline)?

[smlmbrt]

Thanks @mglev1n, we agree and have opened some issues to explore how to make this step faster. It will coincide with allowing wide scoring files (e.g. multiple weights for a single set of variants) as well.