Nanopore data to gtf/gff file

Question

[I0110]

Hi,

I want to use Nanopore PCR-cDNA sequencing to detect known and novel isoforms in Arabidopsis under certain growth conditions. I am thinking to use minimap2 for mapping, alignment and error correction. If I want to get a gtf or gff from the mapped long reads, which software should I use? Or does minimap2 output a gtf/gff file?

Thanks!!!

[Devon Ryan]

Why not use stringTie? Note that minimap2 doesn't output a GTF.

[I0110]

Just not sure if the long reads will not be assembled properly since I only used stringTie for illumina reads before. But I will try that first. Thanks!

[I0110]

StringTie2 is just released and appears to be a good solution for this problem: https://www.biorxiv.org/content/10.1101/694554v1

[sonsunjirachote]

Is stringtie compatible with nanopore data? I always use minimap2, graphmap2 and it generate only SAM output....

Answer 1

[colindaven]

Long shot here, but you might get a first approximation using gmap with a GFF3 output mode.

It's very good, but if you are looking at error rich uncorrected reads it will mess up the ORFs due to sequencing artifacts.

That would be my choice, otherwise you can try to convert bam etc to GFF3 (perhaps via BED etc?).

[I0110]

Can we use minimap2 to correct the error first and use the bam file output (I believe they should give a bam output) for GMAP GFF3 output mode? But then why we don't just do the minimap2 > stringtie as Devon suggested? Is there an advantage to use GMAP GFF3 output mode? Thanks!