I am trying to align and perform reference-guided transcript assembly for all publically available EST and RNA sequencing reads based on a new reference genome for my species of interest. I've done this with several sets of Illumina reads so far using HISAT2 followed by StringTie to assemble novel transcripts from the aligned reads. There are a couple of datasets consisting of Ion Torrent (and 454) that I want to do essentially the same thing with. However, as I understand it, I should not use these programs for processing non-Illumina data.
From what I have read, GMAP is able to handle alignment of both Ion Torrent and 454 reads, and I have used it to align these sets to my reference genome. GMAP can output both SAM and GFF3 formats, but the GFF3 yields a unique annotation for each read, which is not what I want.
I have found a few programs by Googling (i.e. MIRA) that can handle Ion Torrent and 454 reads, but it seems like these are meant for de novo assembly and not reference-guided assembly like I am doing.
I am looking for a package that will essentially perform the same function as StringTie for reference-aligned Ion Torrent reads. Is there perhaps anything that would take GMAP-generated SAM alignments as input? Or is it possible to actually use StringTie for this purpose by changing some of the default parameters optimized for Illumina?
Another option could be to find some quick and dirty way to "collapse" alignments overlapping on the reference strand into single transcripts or transfrags so that they could be more manageably annotated in GTF/GFF format (similar to how PacBio iso-seq transcript derivatives are collapsed into a single annotation). What would be a possible tool for accomplishing this?
