Description
Hi there,
I tried version 2.0.0 on Terra (github.com/HumanCellAtlas/skylab:optimus_v2.0.0_terra) with a human cell 10xv3 chemistry dataset.
I used fastq1 and fastq2 and 10xv3 whitelist and reference genome and annotation as input. And at the same time, run this dataset with CellRanger and zUMIs. The final output for the sparse matrix was collected.
However, I found between these workflows, Optimus seems to underestimate a bunch of cells clearly both with their number of detected genes and number of counts.
Do you have any idea why this happen? Is it because of the CB detection or aligner (but they all use STAR)?
I noticed you have a version particular designed for 10xv3 datasets. What's the difference between them? Could this difference be the reason for this?
Cheers.
Activity