Fig. 6From: Isoform prefiltering improves performance of count-based methods for analysis of differential transcript usageTPR and FDR for DEXSeq-noaggreg, with various degrees of isoform-guided annotation filtering. The performance evaluation is stratified according to the degree of isoform dominance (the difference in relative abundance between the two most highly expressed isoforms of a gene, indicated in the panel headers as e.g. [0.000,0.333)). Excluding lowly expressed transcripts from the annotation catalog before defining the DEXSeq-noaggreg counting bins improves the control of FDR, with negligible loss (or even a slight gain) of power. The total number of genes (n) and the number of genes affected by DTU (n.ds) in each category are indicated in the panel headersBack to article page