Sequences dropped by split libraries

Check drop of sequences at the split library stage and if that has any correlation with env or something else.

The easiest that comes to mind to do this is to change params in split libraries to only discard sequences based on barcode and then compare the numbers per sample ...

Start by reproducing the split library stage and compare sequence counts per sample when filtering is limited to barcode-based discards. Determine whether sequence loss correlates with the environment or another factor, and document the observed counts and correlation.

Written by the indexing model from the issue text.