paths get mixed up when running the pipeline in parallel on multiple datasets

[rgutzen]

In principle executing the pipeline multiple time is straight-forward as you just need to change the PROFILE parameter, e.g.:

snakemake --config PROFILE=LENS
snakemake --config PROFILE=IDIBAPS

However, since the linking of stages works by copying and editing the stage config files as temporary files temp_config.yaml in the pipeline directory, the second snakemake run will overwrite this temporary file with other parameter settings.

A reasonable fix would be to store the working version of the config file ('temp_config.yaml') not in the pipeline directory but in the corresponding output directory (which is profile specific). Furthermore, the config doesn't need to be temporary anymore and could actually be helpful to have stored together with the results, because the reports (which also contain the configs) are more prone to failing (e.g. currently on the HBP clusters).

[rgutzen]

the edited config files for the pipeline execution are now stored as working_config.yaml in the output directories.