Dear Phipson lab team,
Thank you for developing this useful tool.
I had 3 questions:
- for best practice if classifySex() and findMfDoublet() needs to be run on raw expression counts (out of cellranger) or if first samples need to be filtered based on QCs (mito.pct, total read counts, etc.) if that affect the classification performance.
- Is it better to run it independently for each 10X library or if all samples can be merged together into a single matrix to be inputed in cellXY? I seem to have obtained different results using single vs merged approaches.
- Do you recommend running classifySex after filtering doublets (from findMfDoublet) or it shouldn't significantly affect the M/F classification if some doublets are present?
Thanks for the help!
Dear Phipson lab team,
Thank you for developing this useful tool.
I had 3 questions:
Thanks for the help!