[FEAT] Adapt script analyse_citeseq_after_dsb.R

- remove any functionality about thresholds
- make sure that the dsb normalized ADT counts that can be 0 are set to
  0 to not break the clustering
- save the SCE object including BREMSC clusters, and all UMAP
  coordinates in the colData
- save the colData of the final SCE object as tsv