SNPedia link not loading

[andreaswallberg]

Dear developers,

I started up my oldish Scout instance from last year for teaching, and it mostly seems to work fine, with the exception of SNPedia links not showing up in the Summary box after clicking a variant.

Do you know if SNPedia has changed API/interface recently? Anything I can do to fix this?

It seems like SNPedia is normally loaded via a HTML page:

blueprints/variant/templates/variant/variant.html

Some more, possibly quite stupid questions.

Changing the HTML code in variant.html does not appear to lead to any visible changes on the actual page after restarting the server.

Is content for the scout app cached somewhere so that I need to force a reload from source for edits to take effect?

Is there some other resource I should modify rather than the variant.html page if I want to force a visible change on that page / debug my SNPedia issue?

I run scout using gunicorn, if that makes any difference.

Cheers!

[andreaswallberg]

Another issue from before, which could possibly be related is the following:

After adding an HPO gene panel to a case, selecting "Use HPO list for clinical filter" and pressing "Clinical HPO SNV variants", get a huge error list on the following page:

Genes not included in gene panel versions in use when loading this case (clinical list):['ABCC2', 'TRMU', 'A2ML1', 'RFT1', 'NBAS', 'AKR1D1', 'MYD88', 'IRF2BP2', 'SLC7A7', 'VKORC1', 'MGAT2', 'VWF', 'HLA-DQA1', 'CFH', 'PEX14', 'MTTP', 'DPM1', 'FGG', 'STIM1', 'F2', 'NUMA1', 'GALT', 'SOS2', 'F10', 'DLD', 'DPM2', 'LZTR1', 'NRAS', 'NLRC4', 'VPS33A', 'CYP7B1', 'CTC1', 'PARN', 'MCM10', 'NBEAL2', 'B4GALT1', 'CFI', 'PEX13', 'TFAM', 'ZBTB16', 'RIT1', 'TERT', 'STAT2', 'EPB42', 'PEX6', 'GATA6', 'EFL1', 'MAP2K1', 'FGA', 'LYST', 'PEX3', 'PEX16', 'HSD17B10', 'IKZF5', 'WRAP53', 'SOS1', 'PIGA', 'NHP2', 'CBL', 'STX11', 'PEX2', 'PTH1R', 'MST1', 'RAF1', 'KLKB1', 'SPTA1', 'GP1BB', 'USB1', 'XIAP', 'PML', 'F5', 'RINT1', 'NABP1', 'DPAGT1', 'PROC', 'ANK1', 'PEX11B', 'SLC25A15', 'LMAN1', 'RRAS2', 'LARS1', 'GNE', 'HAVCR2', 'PEX10', 'HBB', 'F7', 'RRAS', 'TCF4', 'SLC37A4', 'F13A1', 'PEX1', 'RASA2', 'RTEL1', 'HRG', 'SERPINE1', 'IDH2', 'SLC25A13', 'GATA2', 'SHOC2', 'PET100', 'THBD', 'SERPINC1', 'GNA14', 'DDOST', 'SRD5A3', 'MRAS', 'DOLK', 'SLC4A1', 'PRF1', 'ALG6', 'PTPN11', 'TBL1XR1', 'HADH', 'F9', 'MPI', 'F13B', 'GFI1B', 'PMM2', 'ALG12', 'GGCX', 'TINF2', 'FIP1L1', 'UNC13D', 'STAT3', 'ORAI1', 'PROS1', 'PGM1', 'GP1BA', 'STAT5B', 'KRAS', 'TEK', 'GP9', 'SPTB', 'MCFD2', 'F12', 'SERPINF2', 'GPR35', 'CACNA1S', 'IDH1', 'CD46', 'ABCD3', 'RBM8A', 'ATP6V1E1', 'UQCRFS1', 'NOP10', 'RARA', 'SSR4', 'ANO6', 'AGGF1', 'PEX26', 'FAH', 'TERC', 'OSTM1', 'ACD', 'BCOR', 'COG8', 'ALDOB', 'RYR1', 'ATP6V1A', 'ACAD9', 'MPDU1', 'PEX5', 'BCS1L', 'ATP6V0A2', 'HMGCL', 'COG4', 'NPM1', 'SLC30A10', 'ITGA2B', 'AMACR', 'PRKAR1A', 'F11', 'FGB', 'AGA', 'PEX19', 'AHCY', 'NR1H4', 'COX16', 'ALG2', 'FLNA', 'BRAF', 'PLG', 'NGLY1', 'F8', 'ITGB3', 'DKC1', 'STXBP2', 'HSD3B7', 'PEX12', 'DAAM2', 'SERAC1', 'HLA-DQB1']

It seems like my database as not been correctly loaded when setting it up, using the version of scout from last year. I will try to roll forward to the current version and see what differences these make.

Lastly, how important is the OMIM API key for expected correct configuration and scout behavior?

I don't have one and they never seem to get back to me with one :-/

[northwestwitch]

Hi @andreaswallberg I remember there was some issue with SNPedia, see issue #2460. It's fixed now. As for the other message, it's also fixed (see issue #2760 and #2679). Have you tried installing the latest version of Scout?

Regarding the OMIM key, it is important for assigning OMIM diagnoses to a case and for the gene-phenotype associations. OMIM people could be quite slow to answer if you apply for a key from this page: https://www.omim.org/api, but if you contact them explaining what's the problem (from here: https://www.omim.org/contact) they should answer in a couple of days or so!

[andreaswallberg]

Many thanks @northwestwitch !

I did a git pull and got the same errors as with the older code. I will look into all your feedback!

Regarding the OMIM key, we got away with not assigning an OMIM diagnose last time, so as long as it is not needed to build the database itself "correctly" at this point (i.e. no broken genes etc), we can do without it. Waiting for their response :-)

[dnil]

Hi Andreas! Good fun that you are using Scout for teaching - we love that. I'll second the previous speaker in that we have many new features for you in the current version, but I quite understand wanting to keep things stable.

Lets see if we can help you with the current questions. My time for today is quickly running out; I guess I will just answer a few from the top of my head and hopefully you will get more from others - or later.

• OMIM API key is essential for populating the db with diseases, and populating the generic OMIM morbid gene panel. From a teaching perspective, it will ground the cases in medicine a little more firmly while the bioinformatics is perhaps similar. For us, they usually answer requests within a couple of days. You may have some luck asking a colleague at the clinical platform or hospital for one? You will only need to use it for the download / update.

• We have migrated from SNPedia to the venerable dbSNP. @northwestwitch might remember more why, but basically we changed out the links to ones like https://www.ncbi.nlm.nih.gov/snp/rs76157638. It is all fixed in the latest release.

• The external link out urls are generated in a different function (links.py), but changes to the html pages are generally instantaneous to the deployment. Your setup may have you fooled though. If you install on docker with a static web page directory, you would have to rebuild your docker image. If you install into a conda or other virtual environment with 'pip', you may need to install again (or install with 'pip -e') to see your changes. Obviously there could be additional caches/proxies depending on your setup, but I guess you are looking at a fairly straight connection to the gunicorn server?

• The HPO term to gene conversion is in principle not dependent on OMIM, but the Monarch Initiative / JAXA. They have changed their servers not too long ago, so if you tried to update phenotypes with an old version that may have gone wrong? Also we may or may not have been checking for OMIM API keys before trying to download the phenotype files before. Not so in the later versions of Scout though. A completely or partially failed download of gene to phenotype terms may explain your issues with the HPO panel? Its a little hard to say, as it could just be the case panel being out of sync with the db from the warning above. 😊

[andreaswallberg]

• OMIM: I hope they will get back to me. I will reach out to the clinic/hospital/SciLifeLab if not. Do you know if we have users of Scout in Uppsala?
• SNPedia: that was unfortunate as we used the risk table on the right hand side to compare risk for hetero and homozygotes: https://www.snpedia.com/index.php/rs1799963 I can't see something similar in dbSNP, but I guess you can't please everyone :-)
• Lastly, we have very simple pedigrees in our three cases and only 1-2 handcrafted variants per case. We are not emulating WGS or WES here. Scout works well for us to bring home some messages about inheritance patterns and molecular pathology via links to OMIM, which is why I am sensitive to changes that break my carefully crafted lab questions haha :-)

[andreaswallberg]

Yay! Open source FTW!

[dnil]

Feel free to PR - in that case I would suggest adding the SNPedia link just next to the NCBI one. If you like it, chances are others do as well!