cell ranger v5 include-introns

[hk20013106]

Hi velocyto team,

I have a quick question regarding the pre-velocyto process using cell ranger. The new Cell Ranger 5.0 has a new "include-introns" option for counting intronic reads. https://support.10xgenomics.com/single-cell-gene-expression/software/pipelines/latest/algorithms/overview

Am I right that I should not use this option for veoloctyo analyses?

Thank you very much!

[yaolutian]

same question here thanks

[jon-xu]

Same question here. We tried running velocyto on both cellranger outputs (default setting and include-introns setting). Here are what we observed:

1. cellranger with include-introns has more (about 5%) valid cells due to the fact that introns are regarded as transcripts so the read counts for cells changed the cell order in validation;
2. for the common cells between the two datasets, both spliced and unspliced counts on default setting dataset are larger than those on include-introns dataset - don't know why because the E/N/I flag (Tag: RE:A) should be the same for all reads in the BAM files from the two options. Did Velocyto take additional tags from cellranger include-introns on intron reads into consideration and discarded some? But why did spliced count also decrease with include-intron flag?

Thanks! Jon

[lijxug]

Any updates? Same question here.