Pipeline Overview
The following graphic gives an overview of the entire RNASeq workflow:
The green parallelograms denote the files that are either required (fastq) or generated by the pipeline.
This step generates the basic 10x Genomics Cellranger count analysis output.
Two specific QC plots are generated by the pipeline. One is used to look at the number of genes that only occur in a specific binned number of cells (cells per gene plot). The other plot is used to look at the number of cells that contain a specific binned number of genes (genes per barcode plot). https://satijalab.org/seurat/
Plots are generated using Scran to determine the performance of the Seurat filtering performed and visualize the cell cycle stages of the individual cells.
http://bioconductor.org/packages/release/bioc/html/scran.html
Using the mitochondrial filtered data as input, URD is used to calculate the optimal number of PCs to use. https://github.com/farrellja/URD
Silhouette plots are generated for each of the different cluster resolutions used in Seurat.
Velocyto is used to calculate the predicted change in gene expression
http://velocyto.org
clustViz is used to create a shiny interactive visualization that can be used to explore the clustered data
https://github.com/BaderLab/scClustViz
Cluster profiler is used to obtain the enriched KEGG Pathways and GO Biological Process categories for each cluster at each resolution.
https://www.bioconductor.org/packages/release/bioc/html/clusterProfiler.html
SingleR is used to annotate the cells in the sample with the inferred cell type. https://github.com/dviraran/SingleR
Many graphics are generated during the process, including a box plot depicting the p-values for the different cell type labels, T-SNE plot with the cells labeled by cell type, heat maps of the cells with the cell type labels, and marker gene expression plots.
