Some reads hit on multiple queries; the total number of hits with bit score >40 in each dataset ranged from 18 (in TS) to 5 (in SL) to 1 (in HR and MR, respectively).
Some reads spanned more than one gene.
Some reads were found to map to intergenic regions.
Some reads sampled from low-abundance genomes may still not be filtered in the first step.
Some reads cannot be assigned unequivocally to an isoform since many isoforms share exons.
Some reads which did not align to RefBAC will be misaligned to RefEX.
Some reads, such as those with low mapping quality [ 11], can be filtered during summarization, while whole windows can be filtered afterwards.
Some reads possibly come from regions that are missing in the reference, confounding both the alignment algorithm, as well as the pairing algorithm.
Some reads will have overall low quality, and sequences with low average quality (for example, less than a threshold of ~20) should be removed.
Some reads which do not have any match to the respective database are placed under "No hit" node, and some reads that are originally assigned to a taxon that did not meet our selected threshold criterion are placed under "Not assigned" category.
We found the presence of some reads from both arms of the hairpin, which formed a duplex with 2-nucleotide 3'-overhangs when paired to each other.
