Transcript problems with GRCh38

[AlistairNWard]

Here is a link to GRCh38 file in gene COL9A3.

https://gene.iobio.io/?gene=COL9A3&genes=COL9A3&species=Human&build=GRCh38&affectedSibs=&unaffectedSibs=&rel0=proband&vcf0=https%3A%2F%2Fiobio.s3.amazonaws.com%2Fsamples%2Fvcf%2Fplatinum-exome.GRCh38.vcf.gz&tbi0=&bam0=https%3A%2F%2Fiobio.s3.amazonaws.com%2Fsamples%2Fbam%2FNA12878.GRCh38.bam&bai0=&sample0=NA12878&affectedStatus0=affected

1. When you land all the variants are gray. There is no impact assigned to any variant
2. Click on the transcript dropdown - lots of unnecessary whitespace and if you scroll down through the transcripts there are gaps which I don't understand

3. Click on the far left variant and you get this:

• Under pathogenicity there are symbols that look like gray no entry signs. I don't know what they are supposed to represent (and they're not in the legend). Also, there is one with no text next to it, so I'm not sure what I'm supposed to do with that
• The variant has no rsid number. We'll click on a different transcript in a moment, and you'll see this variant does have one, and this has no connection to the transcript

4. Click on a different transcript (you can just select one from the variant review card)
   • Focus on the variant is lost, so I have to go reselect it
   • Now you see there is an rsid

[AlistairNWard]

This still seems to be a problem, but maybe not something to worry about before paper release.

[AlistairNWard]

Some of the issues appear to be resolved. What remains is:

1. Why are there no impacts on the canonical transcript?
2. Why does the far left variant have no rsid on the canonical trasncript?
3. Changing transcript loses focus on the selected variant - but the variant info remains

[tonydisera]

The problem is that the canonical transcript for gene COL9A3 is incorrect. VEP is returning the annotations on variants for all transcripts of gene COL9A3, but the so-called canonical transcript 'ENST00000343916.7' is not included in VEP transcripts. This results in strange variant annotations in the Variant inspect card, like the ones mentioned above.

So first, there should be an error reported that the transcript provided is not recognized by VEP. And second, we should re-load the transcripts for this gene to correct the transcript problem. (See issue #164).

Here are the transcripts returned from the VEP annotation on a variant for this gene:

Gencode ENST00000335351 ENST00000462700 ENST00000466192 ENST00000466532 ENST00000467819 ENST00000469802 ENST00000469852 ENST00000472880 ENST00000481800 ENST00000490398 ENST00000649368

RefSeq NM_001853.4 NM_006602.4 XM_005260185.4 XM_011528545.1 XM_017027666.1 XM_024451812.1 XM_024451813.1 XR_002958453.1 XR_002958454.1

Here are the transcripts according to gene.iobio:

[AlistairNWard]

MANE would be a nice way to help resolve this. The MANE transcript for COL9A3 is ENST00000649368.1 (NM_001853.4). Which is odd because the Ensembl id doesn't match up.

[tonydisera]

This is strange. The so-called canonical transcript for COL9A3 in gene.iobio is ENST00000343916, which is confirmed in gnomAD.

So was this transcript renamed at some point to ENST00000649368?

[AlistairNWard]

Which is roundly rebuffed by Ensembl!! EST00000343916.3 isn't present at all.

[tonydisera]

Interesting @AlistairNWard! Thanks for looking into this. I think I'll download the GenCode gff (build GRCh38) and look at the transcripts for COL9A3. Let's hope that this has been fixed.