Identification of markers

Tutorial on single-cell gene marker identification

dbscATAC had identified 347,484 single-cell gene markers from 1,668,076 single cells spanning 1,028 tissue/cell types in 13 species.

Download the raw data

For the dataset GSE163697, its raw data can be download from https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE163697 Can run to get the Seurat rds files for all tissue samples:

perl Marker_download_rawdata.pl

Preprocess from the large matrix to cell tpye specific matrices

Can run:

Rscript Marker_split_into_celltype_RDS.R

Identify cell type specigic single-cell gene markers

Similar to scRNA-seq, scATAC-seq can also be utilized to identify tissue/cell-type specific gene markers by aggregating the accessibility signals across gene bodies and promoter regions. We inferred the regulatory activity of genes by calculating the chromatin accessibility near the gene body regions.
For each scATAC-seq sample, the large peak matrix was transformed into a gene activity matrix using the described method. After normalizing this gene matrix, the “FindAllMarkers” function in Seurat was applied to identify tissue/cell-type specific gene markers. To ensure the statistical significance of the identified differentially expressed genes, we set stringent thresholds: an adjusted p-value cutoff of < 1e-5, a minimum fraction of expressed cells in relative tissue/cell type as 0.3, and a log2 fold change threshold as 0.585.
can run:

Rscript Marker_identify_by_signac.R

#Visualization of cell type specific single-cell single-cell gene markers through the module "Search single-cell gene markers" in the home page of our webiste: