Display gene information

[Peter-J-Freeman]

The Python library recovers gene information that we are not displaying. We must display stable gene data in the next build. Might be too late for Christmas 2019!

My preference would be to provide links directly to the gene information e.g. a link to the RefSeq gene record (not to be confused with RefSeqGene), Omim, Ensembl Gene, CCDS, genenames.

All available options can be seen on the REST API

[leicray]

This seems entirely sensible.

[leicray]

Previously, the links to reference sequences allowed the user to choose to display the sequence in GenBank for FASTA format, but the only format now available is GenBank.

[Peter-J-Freeman]

Agreed. To clarify, this is a direct link to Genbank fasta
The link format is
https://www.ncbi.nlm.nih.gov/nuccore/NM_000088.3?report=fasta

[leicray]

Are you saying that you agree to reinstate the option to view sequences in either GenBank or FASTA format?

[Peter-J-Freeman]

Good question. I have mixed opinions. The Fasta can easily be accessed through GenBank records, and the sequence is in the record. Is it necessary. I personally never use it.

[leicray]

Ok, let's just display the sequence in the much more useful GenBank format and perhaps reinstate the option of a FASTA display if users request it.

[Peter-J-Freeman]

Agreed, but this is a good thing to have logged. Good spot

[leicray]

I have just had another look at the gene information that is displayed via the Varsome link and it's really good. Linking out is actually an improvement because, as you say, it emphasises that Varsome is quite separate from VV.

[Peter-J-Freeman]

Agreed. It was a very popular embedded feature, but there are issues with VarSome which I think are not best practice for clinical use. I'm very hopeful that we can build something better. Step 1, Liam's work (or whatever comes next) with visualisation.

[TeriForey]

I removed the choice of genbank or fasta because it was the same number of clicks to the fasta either way, but made it quicker to get to the genbank sequence. Also, I've always found the genbank format much more helpful.

[Peter-J-Freeman]

I cannot argue with that. Agreed

[leicray]

Not sure if it belongs in this issue, but we ought to take a look at NMD Classifier to display info about predicted frameshifts that might trigger NMD.

[Peter-J-Freeman]

Here are the stable gene IDs we pull in from genenames.org

    "gene_ids": {
      "ccds_ids": [
        "CCDS11561"
      ],
      "ensembl_gene_id": "ENSG00000108821",
      "entrez_gene_id": "1277",
      "hgnc_id": "HGNC:2197",
      "omim_id": [
        "120150"
      ],
      "ucsc_id": "uc002iqm.4"
    },

Note: Omim and the CCDS may have more than one entry, i.e. are lists. My feeling is that we should link out to the various resources.

My thought is a table called Stable Gene Information where column 1 is the type e.g. HGNC ID, Ensembl Gene ID, and column 2 is the link which is also the ID (link with the ID displayed).

I might start making this Monday as I have a mega train journey to dundee. Comments please

[leicray]

In the context of eventually building a software tool to support decision making, I wonder if more links could usefully be pulled in from genenames.org. I'm thinking about links to ClinVar, COSMIC, Orphanet, etc. Take a look in the "Clinical Resources" section of the genenames output.

[Peter-J-Freeman]

Defiantly want to provide these links, but this is for gene data.

For the suggested resources, I really want to hit Variant Data. Also, pulling this stuff in from Genenames.org would require a database rebuild.

Variant data will be linked in a later build. I'm working with Brad to get SPDI up and running. We will use this to link to dbSNP and Clnvar. I'll take a look at COSMIC and see how the lonks are assembled.

I'm linking this back to requested features.

[Peter-J-Freeman]

Are we happy withe the gene table now displayed so I can close this please Raymond?