Deconvolution of 10x Visium data using the tools below
- Cell2Location
- RCTD
- Stereoscope
- Seurat
- SPOTlight
- Tangram
- CARD
- STRIDE
- SpatialDWLS
Location on Minerva
/sc/arion/projects/HIMC/nextflow/visium_deconvolution
Example usage
nextflow run main.nf \
--rna_h5ad /sc/arion/projects/HIMC/reference/single_cell_references/andrewLeader_plus_GSE131907_lung/andrewLeader_plus_GSE131907_lung.h5ad \ # Reference scanpy h5ad object
--rna_h5Seurat /sc/arion/projects/HIMC/reference/single_cell_references/andrewLeader_plus_GSE131907_lung/andrewLeader_plus_GSE131907_lung.rds \ # Reference seurat or RDS object
--outdir_final /sc/arion/projects/HIMC/himc-project-data/SAGN29/data/10X/SAGN29_40765_0_v5/deconvolution/andrewLeader_plus_GSE131907_lung/ \ # Save output location
--spatial_h5ad /sc/arion/projects/HIMC/himc-project-data/SAGN29/data/10X/SAGN29_40765_0_v5 \ # Visium dataset
--celltype_key Cell_type_final \ # Column in to use in Reference
--c2l_batch Sample \ # Cell2Locaton Batch argument OPTIONAL
--c2l_covariates Sample_Origin \ # Cell2Locaton Covariates argument OPTIONAL
--c2l_cellsPerSpot 20 \ # Cell2Locaton Estimated Cells Per Spot argument OPTIONAL
-w work/SAGN29_40765_0_v5 \
-c nextflow.config
To run just one tool specify the argument below
--tools RCTD
Or multiple tools (CAUTION can be buggy at times, would just run one tool instead of full user input
--tools 'RCTD, Cell2Location'
Previous commands for huge batch (HCC, NSCLC, CRC etc.) can be found in
batch_cmds/
The scripts are grouped by organ. For example NSCLC
# sample list can be found in
Lung.samples
# script to run per sample
run_Lung.sh
# batch command can be found here
run_Lung.cmd
Notes
- If want want to test the tool in an interactive session instead of submitting a job, change the config parameter
-c nextflow_local.config
- You can use the GPU nodes to run Cell2location
-c nextflow_c2la100.config
--tools cell2location
- Docker images were either created by me or found on DockerHub. Image locations can be viewed in
-c nextflow.config
- Repo of Dockerfiles can be found here