alexdobin / STAR

RNA-seq aligner

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

--soloFeatures for RNA velocity

radiasso opened this issue · comments

Hi Alex, I'm very new to github so please forgive me if I'm maybe not following the exact protocol for issues and questions.
I've been using STAR for aligning single cell RNA data for several years now, never had an issue.
Recently I've been asked at work to add RNA velocity to my analyses and as far as I can see, all tools will need spliced and unspliced counts.
I believe I can generate them through STAR, but I'm not sure what I should change or add to my code for the alignment?

This is the code I use (the genome is slightly changed because I needed to add just a couple of entries for fluorescent colors, otherwise nothing else is different than the standard cellranger genome):

STAR 
--genomeDir=/ext-HD2/roberto/Tools/lib/cellranger-3.0.1/refdata-cellranger-mm10-3.0.0-brainbow/star_chromo_150721 
--readFilesIn=/my-fastq-files.gz/ 
--runThreadN=12 
--soloType Droplet 
--soloCBwhitelist /ext-HD2/roberto/Tools/reference/Brainbow/3M-february-2018.txt 
--soloUMIfiltering MultiGeneUMI 
--soloCBmatchWLtype 1MM_multi_pseudocounts 
--soloUMIlen 12 
--sjdbGTFfile=/my-genes/genes.gtf 
--readFilesCommand zcat

Am I right in thinking I can add --soloFeatures Gene GeneFull_Ex50pAS Velocyto, or a variation of it, to generate this additional info about velocity? I normally get features.tsv, matrix.mtx and barcodes.tsv as final results of the alignment (Solo.out/Gene/filtered), and I load and work on them through scanpy and so on...
I've been trying to figure this velocity thing out for I don't even know how long but I can't find one clear tutorial anywhere. I'm guessing I'd need either scVelo or Velocyto after all this to continue the analysis...

Thank you so much for your help and sorry again!

Hi @radiasso

Am I right in thinking I can add --soloFeatures Gene GeneFull_Ex50pAS Velocyto, or a variation of it, to generate this additional info about velocity?

yes, you need to run STAR with Velocyto option.
To load the results in Scanpy, please check this thread:
#774 (comment)

Hi, @alexdobin, thank you for your confirmation.
I tried adding --soloFeatures Velocyto to the code. It ran apparently without issue until the very end, when it "started Solo counting" and "Starting Solo post-map for Velocyto".
At that point it raised: Segmentation fault (core dumped).
The exact same code had worked for normal STAR alignment already in the past, I just added --soloFeatured now.

I'm using STAR version=2.7.5a_2020-06-29 in Ubuntu 22.04.3 LTS (work server with 2Tb of space and 64Gb of RAM). Since it's a work server I can't uninstall and reinstall everything. STAR has always worked great for these alignments, I really hope I can add the Velocyto option somehow.
Any ideas why it raises that error?
Thank you again for your help!

EDIT: I'm an idiot, I re-ran with --soloFeatures Gene Velocyto and it worked perfectly, generating both the Velocyto raw folder (features, barcodes, matrix) and the classic Gene/filtered. I'll go look for the comment you linked to see how to load this data in Scanpy!