In wguo-research/scCancer: A package for automated processing of single cell RNA-seq data in cancer

r h.i.r h.ii Cell malignancy estimation

In order to distinguish malignant and non-malignant cells, we infer copy number alterations (CNV) from tumor single cell RNA-Seq data referring to the method of R package infercnv. Then we calculate a smoothed malignancy score based on the CNV profile.

Following is the malignancy scores distribution plot for observation cells in the sample (blue) and reference cells (grey). By detecting the bimodality in the malignancy score distribution,

if(!is.null(results$malign.thres)){
    cat("we get the bimodal boundary is nearly `", format(results$malign.thres, digits = 3, scientific = T), "` (red dash line).", sep = "")
}else{
    cat("we cannot think the distribution is bimodality.", sep = "")
}

results$malign.plot$p.malignScore

(Hi-res image: view)

Here is the t-SNE plot colored by malignancy score (left) and type (right).

plot_grid(results$malign.plot$p.malignScore.Point, 
          results$malign.plot$p.malignType.Point, ncol = 2)

(Hi-res image: left, right)

Here is a bar plot showing the relationship between cell cluster and cell malignancy type.

results$malign.plot$p.malignType.bar

(Hi-res image: view)

The estimated cell malignancy scores and types can be found in the column Malign.score and Malign.type of the table file cellAnnotation.txt.

After this step, scCancer saved following results files to the folder 'malignancy/': * Estimated CNV profile of reference cells: inferCNV-reference.txt. * Estimated CNV profile of sample cells: inferCNV-observation.txt. * Malignancy scores of reference cells: refer-malignScore.txt.

h.ii <- h.ii + 1

wguo-research/scCancer documentation built on May 26, 2024, 9:12 p.m.