Skip to main content
Fig. 8 | Genome Biology

Fig. 8

From: GoM DE: interpreting structure in sequence count data with differential expression analysis allowing for grades of membership

Fig. 8

A Structure in Mouse Atlas sci-ATAC-seq data (n = 81,173) inferred from topic modeling, with K = 13 topics. B Topic model fit to kidney cells (n = 6,431) with K = 10 topics. C, D Gene-based enrichment analysis of differentially accessible peaks for the kidney cell topics shown in B, in which peaks are linked to genes using Cicero [95]. In A, the cells are grouped by tissue, and replicates (for bone marrow, large intestine, lung and whole brain) are shown as separate tissues. Numbers in parentheses next to each tissue give the number of cells in that tissue. In D, marker genes for S1 (topic 4) and S3 (topic 5) proximal epithelial tubular cells are highlighted in red (see Table 1 of [96]). “Mean l.e. LFC” is the average l.e. LFC among all peaks connected to the gene, restricted to l.e. LFCs with lfsr < 0.05. Log-Bayes factors greater than 200 are shown as 200 in the volcano plots. See Additional file 1: Fig. S9 and Additional file 6: Table S7 for more gene enrichment results. In B, the cells are subdivided into 5 groups (a–f) only to improve visualization. See also Additional file 1: Fig. S10 which compares the topics in B to cell-type predictions based on clustering [97]

Back to article page