Researchers at the Mount Sinai Health System have created an automated script that takes your raw files from cell ranger and creates an integrated clustered Seurat object that can be used for visualization and downstream analysis.
The clustering pipeline availabletakes your folder with multiple raw matrix output files from cell ranger to create Seurat object. These Seurat object are then integrated using the approach as in the paper using a user defined parameter for the number of PCs used for clustering the resolution needed for the clustering. There 2 parameters are passed as arguments with the script. The output of this script is an integrated model that is stored a san .rds file in your working directory.
The visualization app available atlets you load the .rds file created from the above script and displays the summary of the integrated model (number of UMIs, table of cells per cluster per group/sample, number of samples used and the name of the project). It lets the user plot TSNE, Heatmaps, feature plots, dot plots for the user defined gene of interests interactively. It also lets the user perform downstream analysis on the dataset – defining cluster markers, perform differential gene expression, reclusters a specific cluster and subset the cluster based on multiple different filters. Besides this, it also lets the user save the newly sub-clustered or reclustered data as an .rds object and the user can easily download the average expression table from the dataset. It does include most of the functionality needed to analyze scRNAseq data. There are some recommendation to incorporate various other methods that are being added to the app currently. We would be delighted to add more features as per need and request.
Availability – Clustering script – https://github.com/mamtagiri/sc_ClusteringPipeline Visualization app- https://rcg.bsd.uchicago.edu/ibdgc/seurat/