IGV merging features in Pact genome

[yaaminiv]

I visualized various P. acuta feature tracks in this IGV session (all features link to gannet so it should be viewable by others). Looking at the tracks, I noticed that introns were overlapping with CDS, which doesn't make any sense. I think this might be because IGV is taking introns from different transcripts that overlap with eachother and smashing them together.

If the introns, CDS, exons etc. for each transcript aren't mashed together, it would be easier to figure out if these tracks are valid. Is there something I need to do when generating tracks to ensure they're processed properly in IGV?

[kubu4]

Intron tracks should be generated from gene/CDS info. Is that how this was constructed?

[yaaminiv]

@kubu4 No, I pulled the intron information straight from the genome annotation file we have in this notebook:

I did the same thing for *M. capitata* and I didn't see any overlapping introns for that annotation.

[kubu4]

Were those annotations created by the same people with the same software?

I'd start by visualizing the original GFF. However, I'm guessing IGV uses coordinates based on gene ID. So, it will merge stuff that have overlapping coordinates with same feature and gene ID.

[yaaminiv]

Were those annotations created by the same people with the same software?

The genomes were constructed and annotated by separate groups but with the AUGUSTUS software.

The original GFF is a bit convoluted, but I'll see if I can pull out what I need, make sure the scaffold names match, and load it in IGV.

[yaaminiv]

However, I'm guessing IGV uses coordinates based on gene ID. So, it will merge stuff that have overlapping coordinates with same feature and gene ID.

I think you're right. Just looking at the CpG track, it merged two CG motifs on separate lines into one object. I'll hold off on going to the original GFF for now. I pulled the information from there, so my guess is that it will also look smushed. This is not ideal!