Scripts for single-cell 'omics analysis
Full pipeline in seurat 3.0 including
- optionally filter cellranger raw BC matrices on QC cut-offs
- removes ambient RNA (SoupX) (experimental)
- removes doublets (doubletFinder)
- integrate samples (using Seurat 3.0 algorithm)
- tSNE feature plots and diagnostic QC plots
e.g.
time Rscript ./seurat_pipeline_seurat3.0.R --dirs_project_10x "c('/nfsdata/data/sc-10x/data-runs/181119-serup-pancreas/1-5000_cells/','/nfsdata/data/sc-10x/data-runs/181119-serup-pancreas/2-5000_cells/','/nfsdata/data/sc-10x/data-runs/181119-serup-pancreas/3-5000_cells/','/nfsdata/data/sc-10x/data-runs/181119-serup-pancreas/4-5000_cells/','/nfsdata/data/sc-10x/data-runs/181119-serup-pancreas/5-5000_cells/')" --dir_out /projects/jonatan/pub-perslab/181119-serup-pancreas/ --flag_datatype sc --flag_organism hsapiens --prefix_data serup_panc_2 --prefix_run seurat_2 --n_cells_loaded 9000 --n_cells_recovered NULL --use_filtered_feature_bc_matrix T --nCount_RNA_min 2000 --nCount_RNA_max 50000 --run_SoupX F --nFeature_RNA_min 0 --nFeature_RNA_max 25000 --percent.mito_max 1 --percent.ribo_max 1 --rm_sc_multiplets T --vars.to.regress 'c("percent.mito","nCount_RNA")' --nAnchorFeatures 2000 --n_comp 50 --use_jackstraw T --res_primary 1.2 --res_to_calc "c(0.8,1.2,1.5,2.0,2.5)" --feats_to_plot "c('nCount_RNA','nFeature_RNA','percent.mito','percent.ribo')" --feats_plot_separate T --RAM_Gb_max 250 &> ./serup_panc_2_seurat_2_runlog.txt
Rscript ./seurat_pipeline_seurat3.0.R --help
- Currently the pipeline normalizes the raw counts using LogNormalize. Add sc-transform?
- Add UMAP visualisation