thresholds in Drosophila samples

[hannahd1604]

Hi,

I am running AMULET on a few Drosophila samples, but there are fewer doublets than expected (especially when removing the remaining doublets after QC). Do you have any recommendations for adapting the default thresholds for other species?

Best,
Hannah

[ajt986]

Hi Hannah,

I do not have recommendation for a specific species, just that the --expectedoverlap parameter matches the ploidy of the cells that are being investigated. The majority of drosophila cells are diploid, so the default parameter should be correct. Additionally the repeats/ segmental duplicates should be obtained for the species of interest for the filtering step.

It's likely that there is not sufficient read depth to identify more doublets with the default threshold (q-value < 0.01). You can try increasing this threshold to 0.05 or 0.1 to identify more doublets by rerunning the second step of AMULET:

python3 AMULET.py --q 0.05 --rfilter <repeats_segdups_for_drosophila.bed> Overlaps.txt OverlapSummary.txt <newoutput>

Best,
Asa