htseq-transcriptome
htseq script to count reads multiply mapping to a transcriptome assembly
Because a transcriptome assembly includes collections of isoforms from the same gene, any reads that would have mapped uniquely to a genome assembly - where there are multiple isoforms - would in a transcriptome assembly map multiply - once to each isoform.
Because these reads are discarded in the classic implementation of HTSeq-count, a different implementation needs to be developed to account for the legitimately multiply-mapped reads in a transcriptome assembly.
HTSeq-transcriptome provides this functionality. It is built with the same framework as HTSeq-count, and the same parameters are available (or will be added soon!). It therefore should be very easy to use for anyone who is familiar with using HTSeq-count.
The difference is that HTSeq-transcriptome has a unique way of assessing read bundles. Rather than discard all reads that are multiply-mapped, it first looks to see if the multiply-mapped loci all have the same gene_id. In this way, all reads mapping to different isoforms of the same gene will not get discarded, but are counted as a SINGLE count toward the gene's expression.