Question...
Wei mentioned that one difference between the performance of subread-align and subjunc is that subjunc has lower sensitivity. In other words if you compare the two on the same set of rna-seq like reads subread-align will align more of the data than subjunc. This seems to be opposite what I'm used to. For example if you align RNA-seq to a genome with bowtie and then with Tophat you'll have more reads aligned with Tophat almost guaranteed thanks to the additional alignments from spliced alignments. So intuitively something doesn't make sense. Why wouldn't the output of subjunc basically be the same as subread-align with the added alignments of reads that span junctions?
Wei mentioned that one difference between the performance of subread-align and subjunc is that subjunc has lower sensitivity. In other words if you compare the two on the same set of rna-seq like reads subread-align will align more of the data than subjunc. This seems to be opposite what I'm used to. For example if you align RNA-seq to a genome with bowtie and then with Tophat you'll have more reads aligned with Tophat almost guaranteed thanks to the additional alignments from spliced alignments. So intuitively something doesn't make sense. Why wouldn't the output of subjunc basically be the same as subread-align with the added alignments of reads that span junctions?
Comment