Runs velocyto on multiple 10x outputs from cellranger to project trajectory arrows onto an UMAP from a partially processsed merged/integrated seurat object.
Velocyto preprint: RNA velocity in single cells
Put the seurat object rds file in the input directory with .rds extension.
Add 10x paths into the "paths.tsv" file in the input directory.
SAMPLE:
location name
/PATH/TO/SAMPLES/10x_1 1
/PATH/TO/SAMPLES/10x_2 2
The "outs" directory from 10x cellranger would be a sub-directory of the "/PATH/TO/SAMPLES/10x_1". The name column references the postfix of the cell barcode used in the seurat object.
Add the paths in the omic_config.yaml for the GTF files.
Repeat masker downloaded from UCSC. Repat masker is optional to velocyto but it is required here.
Reference gtf annotation files from 10x. Human 10x gtf
The links are to hg19.
omic_config.yaml
GTF_ref:
/path/to/10x.gtf
repeat_mask:
/path/to/repeat_masker.gtf
Activate snakemake environment and run the following:
$snakemake --use-conda --cores=4 --printshellcmds --reason -np
$sbatch submit_snakemake.sh
Velocyto uses the basename of the path to the 10x location directory and this can cause problems. The script correct_CB.py corrects for this by using the "paths.tsv" file as lookup table to connect Velocyto output with seurat object cell barcodes.