Skip to main content
Fig. 4 | Genome Biology

Fig. 4

From: ReadZS detects cell type-specific and developmentally regulated RNA processing programs in single-cell RNA-seq

Fig. 4

Comparison of ReadZS with pseudobulk-based approaches for APA detection. A The ReadZS has unique power to discover regulation missed by peak callers. In both of these examples called by ReadZS as significant, the change in relative peak height suggests the different groups of fibroblasts use alternate PA sites at different rates, yet these genes were not called by Sierra as undergoing DTU. Top: Rpl13a has cleanly separated peaks but was not called by Sierra. Bottom: the jagged peaks of reads in the 3′UTR of Rab2a might hinder peak calling-based methods. B Proxy measurements for true and false positive rates of MAAPER and ReadZS, from comparisons of different pairs of cell types from HLCA P2 and P3, evaluated at different alpha values (corrected p-value cutoffs for significance): “true positive rate” proxy (TPR)—red squares; “false positive rate” proxy 1 (FPR1)—blue circles, and “false positive rate” proxy 2 (FPR2)—green triangles (see “Methods” for calculation of these metrics). Missing points indicate that there were not sufficient significant genes or windows to calculate the proxy measurement. Full set of plots is in Additional file 1: Fig. S5. C Comparison of the total memory and run time (measured in GB•hours) for running ReadZS or running MAAPER on the same data. Each row represents a set of data, specifically a comparison between two cell types subsetted from the combined HLCA P2 and P3 datasets. ReadZS GB•hours were calculated automatically for each Nextflow run in the Nextflow Tower interface. MAAPER GB•hours were calculated by multiplying the memory allocated to each job with the number of hours required to run the data. D SCGB3A1 is one of many genes not called by MAAPER but identified by ReadZS as undergoing cell type-specific RNAP. The window overlapping SCGB3A1 was called as significant by ReadZS in the comparison between macrophages (from HLCA P2 and P3) and alveolar epithelial type 2 cells (from HLCA P2 and P3)

Back to article page