Fig. 2: CNA identification of a glioma sample BCH869 by XClone and performance comparison between methods. | Nature Communications

Fig. 2: CNA identification of a glioma sample BCH869 by XClone and performance comparison between methods.

From: Robust analysis of allele-specific copy number alterations from scRNA-seq data with XClone

Fig. 2

a UMAP of BCH869 with different types of cell annotation: cell clustering by transcriptomes and CNA tumor clone ID identified by the original paper, CNA region identified for each clone and the inference clonal tree in the original paper. b–d The heatmaps show CNAs identified using XClone’s RDR and BAF modules, and their combination, for BCH869. Each row in the heatmap represents a cell, and each column a genomic location: genes for RDR and combined heatmaps, and 100-gene bins for BAF heatmaps. Colors in the heatmap indicate the inferred copy number state at each location, reflecting genomic heterogeneity among cells. e presents a heatmap visualization of CNAs produced by Numbat. In this heatmap, each row typically represents a single cell, and each column represents a genomic location, nearly chromosome arm scale fragment. f–h Evaluation of the performance of benchmarked methods in identifying copy number gain, copy number loss, and loss of heterozygosity on the BCH869 glioma sample (at the gene scale). i Volcano plot showing negative log10 P-values against average log2 fold changes (FC) for DEGs between Clone1 and Clone2-4 in chromosome 14. DEGs (Adjusted p-value < 0.05 or 0.2) are colored points above the horizontal dashed line. Significant DE genes expressed higher in Clone1 are highlighted in red, and lower in Clone1 in blue. j Obvious DEGs expression pattern in Clone1 and Clone 2-4. k RPS6KA5 gene expression with SNP chr14:90897944_G_A. In the boxplot, the x-axis is the two groups with different genotypes of chr14:90897944 SNP: “G/-" in three clones 2, 3, 4 (262 cells); “-/A" in clone 1 (227 cells), where “-" stands for copy loss. The y-axis is the RPS6KA5 gene expression that the normalized counts (by library size with scale to 10,000 followed by log1p transformation). The box center line, boundaries, and whiskers denote the medians, the first and third quartiles, and 1.5 times the interquartile range, respectively.

Back to article page