Add text describing pipeline processing to uns

[dosumis]

Description from Iris. Applies to Droplet. Double check that this is sufficiently generic and that Iris is happy to include this text on all datasets

UMI counts (quantifications) from libraries of the same experiment are aggregated into a single quantification table.  Taking into account the batches, doublets are removed with scrublet .  The data is then cell-filtered (low-quality UMIs removed), and then gene-filtered (genes with less than 3 UMIs mapped are removed).  Finally, the data is normalised.  This is the X matrix in the final anndata object.