Improve management of anndata during finemapping

[edg1983]

Current behaviour and possible enhanchments

• generate an anndata per locus: this can be made more efficient by generating an anndata for a chunk of loci of N size where N is set by a parameter. This can increase flexibility and reduce the amount of files generated during work.
• publish a single anndata for each pipeline run: For re-usability, it would be nice to save one anndata per trait or per study (here, study means a study like UKB)

[linear]

HT-29 Save anndata by study or trait after finemap

Now we save a single anndata for each pipeline run.

For re-usability, it would be nice to save one anndata per trait or per study (here, study means a study like UKB)