/singleCell

scRnaSeq Analysis for 10X Genomics based on Seurat, Scater & SingleCellExperiment packages.

Primary LanguageR

SingleCell RnaSeq Analysis

You will find bash and R scripts in two distincts directories.

  1. xenomeCreation.sh create the genome reference based on 10X tutorial that will be used by Cell Ranger.

  2. align10x.sh run Cell Ranger alignment for the several samples.

  3. filterMouseReads.sh remove reads that mapped on mouse and create clean fasta with reads specific to human.

  4. qualityControl.R make plots to check quality.

  5. filterCells.R filter single cells (based on previous generated figures) Seurat Object and save it to Clean.rds file on disk. Thresholds are stored in files containted in the data directory.

  6. DE-speudoBulk.R retrieve individual Clean.rds files from previous step and do speudo-bulk differential expression analysis between two conditions.

  7. clusterCells.R from the rds object containing filtered seurat object for each condition , you apply seurat normalisation and save in an _integrated.rds file... ( can be 2 or several conditions)

  8. plotclusterCells.R load the integrated.rds file..

  9. scfGSEA.R fSGEA for speudo-bulk RnaSeq.

  10. tsne.R plot tsne or umap for several experiments.

In bash/triggers, there is bash scripts that call (trigger) other bash scripts...

  • trigger_filterMouseReads.sh call filterMouseReads.sh for several samples.

  • trigger_main.sh calls sequentially bash scripts in core directory that call the core main R scripts qualityControl.R, filterCells.R , DE-speudoBulk.R and scfGSEA.R for several samples/conditions.

Velocity Analysis

Mix of R and pythons stuffs.

  1. veloctyo.sh : You create loom files using dir from 10x samples using velocyto.

  2. plotNestedCluster.R : tsne.R adaptatation. Will combined samples, run Umap Tsnse (need to extract list of genes differentially expressed per cluster) save files to disk to be reused by python scripts latter.

  3. loom_combine.py : combine several loom files. (loom file are generated from dispached 10x samples per experiment)

  4. velocyto.R : Read the loom, transform to seurat, RunPCA, applied all seurat stuffs...and finaly run RunVelocity

  5. velocyto2.R : Read the loom and finally save h5ad file that can be use as starting point for velocyto.py.

  6. velocyto.py : start from h5ad file and plot trajectories , velocity related plots.