Entering edit mode
3.7 years ago
bataloffv
▴
20
Hi everybody! Probably a basic question. I have a bam aligned against a hg38 reference from Broad Institute. I also have a gtf annotation file. Everything loads and displays in IGV, but the search isn't aware of gene names from the GTF file. Actually the only way to use the IGV search box is to type chromosome coordinates. IGV user guide says that a feature track should not be indexed to be searchable, but the browser refuses to load a gtf of such size (1192M) without indexing it. How to make genes from a large gtf annotation file searchable by their names?
Thanks!
IGV should happily search tracks: http://software.broadinstitute.org/software/igv/?q=navigate#search
You're running it locally or using the web browser?
I run it locally. Apparently, the reference genome that I use, has no annotation track. So, according to the user guide you that you provided the link to, I should use a non-indexed feature track to be able to search genes in it, and the feature track is too large to be loaded without indexing.
Why do you say it's too large to load without indexing? If your computer has enough ram but IGV is still throwing out an error, you have to increase the allowable memory used by java. You'll have to google but from memory it was annoying but doable.
Maybe the solution would be to subset the genome/tracks?
Finally, is there something in particular you're searching for? Maybe there's a solution that we can think of that will allow you to subset both the genome and the gtf file (eg bedtools).
Thank you for the replies. I already tried increasing allowable memory (-Xmx12g instead of -Xmx4g in igv.sh). Still IGV demands the gtf to be indexed. With everything loaded (the gtf, a wgs bam, and a couple of vcfs) IGV uses just 1.6G of RAM. Actually, I already had found what I was looking for, at the time when I asked the question. But still it would be helpful to be able to load large feature tracks and have the search working on them. I guess somewhere should be a tweak, which adjusts size of loadable feature tracks, but I couldn't find it, maybe it's hardcoded in java code.
I think you should pose this question to the IGV developers. It's a good question!