RefSeq transcripts

[LeifGlitch]

The issue relates to the variant here:
https://scout.scilifelab.se/cust003/25437-1/83303506f9a882543d5ee4bae4b8661a

Hi!
For this variant the HGVS description is missing for the MANE transcript/disease associated trancripts; NM_030948.3. In this specific case it was a near miss to identify this as a known published pathogenic variant.

The right HGVS description is visible in the list of transcripts further down the page for the newest version. PHACTR1; NM_030948.6, c.1553T>A | p.Ile518Asn.

Is it a bug that makes it not appear in RefSeq transcripts?

best/Leif

[northwestwitch]

Weird, we'll check!

[dnil]

Thank you for reporting! We'll check, but from previous experience it feels likely this is due to the annotation being made on ENSEMBL transcript, and the oooold ENSEMBL 37 release having marked ENST...415087 as the best match to some version of NM_030948. This would in all likelihood be different with MANE on hg38.

On a sidenote, perhaps time to submit it to ClinVar to have it safely caught - a little sad that it isn't there already as previously published.

[dnil]

@northwestwitch also note the panel "disease associated transcript" is lagging a little with NM_030948.3 vs this case with NM_030948.6.

[northwestwitch]

@northwestwitch also note the panel "disease associated transcript" is lagging a little with NM_030948.3 vs this case with NM_030948.6.

That might be why. I haven't checked the code yet, but by looking at the variant the transcript with NM_030948.6, c.1553T>A is not canonical - and that's also why it doesn't end up in the table above, additionally.

Perhaps we could relax the disease associated transcript matching check to not to include transcript version, if that's the case. But I'll check what the code does now!

[northwestwitch]

Ok, as Daniel was guessing the issue will be resolved switching to build 38. The thing is the transcript in build 37 is missing refseq_identifiers, see the difference between 37 (top) and 38 (bottom) in our database:

And unfortunately (I mean, it's by design) the code collecting the most important transcripts to be shown in the top table is only picking those that are either canonical or have Refseq IDs:

scout/scout/server/blueprints/variant/controllers.py

Line 86 in 3a7705f

if "refseq_identifiers" not in tx and tx.get("is_canonical", False) is False: