Introduction to scSpotlight

Seurat Analysis Workflow

scSpotlight uses the Seurat standard single cell analysis workflow, we recommend new users going through their pbmc3k tutorial to better understand each processing step. The demonstration dataset used is a 2,700 Peripheral Blood Mononuclear Cells (PBMCs) sample made publicly available by 10X Genomics. The matrix could be downloaded here.

A typical seurat analysis workflow is like below:

Invoke App

To start using scSpotlight, please use the command below and access the app via: 127.0.0.1:8081

scSpotlight::run_app(options = list(port = 8081, host = "0.0.0.0"))

The default mode of app is processing. User could also change mode to viewer, which will only allow illustrating dataset and querying gene expressions:

scSpotlight::run_app(options = list(port = 8081, host = "0.0.0.0"), runningMode = "viewer")

If one needs to load a very large dataset, use dataDir parameter to mount data directory and load Rds file directly.

scSpotlight::run_app(options = list(port = 8081, host = "0.0.0.0"), runningMode = "viewer", dataDir = "/path/to/data_directory")

Main Panel

Main panel of scSpotlight will be like:

Input Files

Under the “processing” mode, scSpotlight supports compressed file containing typical 10X cellranger output (Matrix Market Format):

$ tar -tvf PBMC_demo.tar.gz
drwxrwxr-x xzx/xzx           0 2024-01-11 21:11 PBMC_demo/
-rw-r--r-- xzx/xzx    67972057 2024-01-11 21:10 PBMC_demo/matrix.mtx.gz
-rw-r--r-- xzx/xzx      294722 2024-01-11 21:10 PBMC_demo/features.tsv.gz
-rw-r--r-- xzx/xzx       46443 2024-01-11 21:10 PBMC_demo/barcodes.tsv.gz

and processed Seurat object seurat_processed.Rds saved by [Seurat::SaveSeuratRds()].

If using compressed matrix file as input, scSpotlight will process data after loading without cell filtration and the dimension reduction will be shown.

Filter Cells

scSpotlight will automatically illustrate nGene (nFeature_RNA), nUMI (nCount_RNA), mitochondrial gene UMI percentage (percent.mt) and ribosomal protein gene UMI percentage (percent.rp) for each cell in VlnPlot panel. One could inspect the distribution and roughly evaluate the library quality. User could discard cells with too few or too many features detected and dying cells with high mitochondrial UMI percentage by adjusting parameters in the “Cell Filtering” block. The filtration step will also update highly variable genes (HVGs), the dimension reduction and the cell clustering result.