Fig. 2

Genomic profile of unmapped reads across 10,641 samples and 54 tissues. Percentage of unmapped reads for each category is calculated as a fraction from the total number of reads. Bars of the plot are not scaled. Human reads (black color) are mapped to the reference genome and transcriptome via TopHat2. Unmapped reads are profiled using the seven steps of ROP protocol, described below. (1) Low quality/low-complexity (light brown) and reads matching rDNA repeating unit (dark brown) were excluded. (2) ROP identifies lost human reads (red color) from unmapped reads using a more sensitive alignment. (3) Hyper-edited reads are captured by hyper-editing the pipeline proposed in [17]. (4) ROP identifies lost repeat sequences (green color) by mapping unmapped reads onto the reference repeat sequences. (5) Reads arising from trans-spicing, gene fusion, and circRNA events (orange color) are captured by a TopHat-Fusion and CIRCexplorer2 tools. (6) IgBlast is used to identify reads spanning B and T cell receptor gene rearrangement in the variable domain (V(D)J recombinations) (violet color). (7) Microbial reads (blue color) are captured by mapping reads onto the microbial reference genomes