Strange spliced/unspliced percentages with cell hashed data

[carbycrab]

Hello,

Thank you for creating such an innovative tool! I am using velocyto upstream of scvelo to analyze my 10x single cell sample and am encountering some strange results. scvelo returns majority unspliced (86%) and minority spliced (14%) with my dataset, which are way off from the 15-25% of unspliced intronic reads in La Manno et al 2018. I was wondering whether my abnormal spliced/unspliced percentages might have to do with the way I am running velocyto or the nature of my dataset, which has undergone cell hashing. This is my code:

samtools sort -t CB -O BAM -o mysample/outs/cellsorted_possorted_genome_bam.bam mysample/outs/possorted_genome_bam.bam
velocyto run10x -m GRCm39_rmsk.gtf mysample gencode.GRCm39.basic.annotation.all.gtf -vv

When I run another 10x dataset that did not use cell hashing through the same pipeline, I get normal values for spliced (82%) and unspliced (18%). Do you know whether this is a cell-hashed sample issue and whether I need to run velocyto differently? Thanks so much for your help!

[learning-MD]

Hi, wanted to follow-up on this, as I have several hashed samples that I'd like to use velocyto for. Any chance you figured this out, @carbycrab, or anyone else have suggestions before running hashed samples? Thanks!