Fig. 2

Results on simulated datasets. The figure depicts the yields (a), speed (b), and memory footprints (c) of the aligners on the simulated datasets. It is worth noting that for deSALT and Minimap2, both the results with and without gene annotations are shown (indicated as “deSALT+GTF,” “deSALT,” “Minimap2,” “Minimap2+GTF,” respectively). a Each of the subplots indicates one of the four metrics (Base%, Exon%, Read80%, and Read100%, respectively) of the aligners on the datasets in a specific sequencing model (PacBio ROI, PacBio subreads, ONT2D, ONT1D, PS-ONT, and NS-ONT, respectively). In each subplot, the blue, green, and red lines respectively correspond to the results of the datasets from randomly selected genes of human, mouse, and fruit fly, and the orange lines correspond to the results of the dataset from all the protein coding genes of human. Moreover, the shapes (reversed triangles, rectangles, circles, and rhombuses) indicate the datasets in various sequencing coverages. In b and c, each of the subplots indicates the speed (#Base/m) and the memory footprint (in GB) of the aligners (in 8 CPU threads) on the datasets simulated from human (randomly selected genes) with a specific sequencing model. The datasets in various sequencing coverages are shown separately, and the bars in different colors refer to various aligners. Also refer to Additional file 1: Table S8 for a more comprehensive assessment of the alignment speeds and memory footprints in various numbers of CPU threads (1, 4, 8, and 16 CPU threads)