basilkhuder/Seurat-to-RNA-Velocity

high unspliced counts by velocyto

rtoddler opened this issue · 4 comments

Hi,
Thank you for your incredible guideline. I had a problem making loom file by velocyto. Here is the code I use:

velocyto run -b filtered_barcodes.tsv -o velocyto_output -m mm10_repeats_repeatMasker.gtf possorted_genome_bam.bam refdata-cellranger-GRCh38-3.0.0/genes/genes.gtf

I can successfully generate the loom file but the unspliced RNA is very high (unspliced:spliced=0.92:0.08). This is not expected since the dataset is generated from 10X 3' v3.1 single-cell library. Do you have any idea why I am getting this ratio?

I wonder if I used the wrong input for velocyto. Here are what I use:

  1. barcodes.tsv: unzipped file from CellRanger outs>filtered_feature_bc_matrix>barcodes.gz
  2. repeatMasker.gtf: downloaded from USCS table browser
  3. bam file: file from CellRanger outs>possorted_genome_bam.bam
  4. genes.gtf: mm10 mouse reference dataset provided by 10x

Any comment will be appreciated!

Hi there!

Are you using a human GTF file (refdata-cellranger-GRCh38-3.0.0/genes/genes.gtf) or was that a mistake in typing?

Thank you for your speedy response! I'm sorry. That was a mistake. I use mm10

Gotcha - thanks for clarifying. You're right though, the unspliced proportion is uncannily high. There doesn't seem to be anything wrong with the input commands, so the only suggestion I have is trying to re-run this without the repeat-masker file and see what proportions arise. If that doesn't change anything, you might want to try talking to the velocyto team (although, issues on their Github haven't been getting many replies as of late.)

Good to know! Yes, I will try to run it without repeat-masker. You are right...I browsed through ALL the issues on velocyto-team github and could not find the answers. I'm sooo glad that I find your tutorial (especially the part for extracting filtered cell ID). Once I figure out the loom file, I would definitely give it a shot. Might come back here for more questions tho because I'm really new to RNA velocity analysis.

Thank you again! Your guide is very clear and helpful.