Allow common normal (pseudodeficiency, modifier) variants to load via managed variants?

[4WGH]

Hi, we would like to "register" some managed variants. Where can I read how do they work? How can I submitt them?

Sorry to bother you but I can not find the information myself in the Scout user guide :(

[northwestwitch]

Hi! You add them either manually one by one or by importing a file.

Go to this page: https://scout.scilifelab.se/managed_variant and if you want to add one by one use the Add variant option, otherwise Load variants to load the file.

If you don't have so many it is perhaps easier to save one at the time.

I don't think we have anything about them in the documentation, we should add an howto!

[4WGH]

thanks! I assume is hg37 coordinates?

who can see them? only CMMS? can they be filtered according to panels?

---

Från: Chiara Rasi @.> Skickat: den 9 april 2021 10:55 Till: Clinical-Genomics/scout @.> Kopia: Michela Barbaro @.>; Author @.> Ämne: Re: [Clinical-Genomics/scout] Managed variants - question (#2508)

Hi! You add them either manually one by one or by importing a file.

Go to this page: https://scout.scilifelab.se/managed_variant and if you want to add one by one use the Add variant option, otherwise Load variants to load the file.

If you don't have so many it is perhaps easier to save one at the time.

I don't think we have anything about them in the documentation, we should add an howto!

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/Clinical-Genomics/scout/issues/2508#issuecomment-816532452, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AC7R5EOQSR5PBBCOSRFPPILTH26H3ANCNFSM42UQK3SA.

[northwestwitch]

Yes, hg37, we are thinking to introduce a genome build for managed variants (or at least a liftover) for when we switch to the new build but it's not in place yet.

regarding you other questions, please @dnil correct me if I'm wrong: I think everybody can see them and at the moment they can only be filtered by the items that are in the filters panel (so no gene panels):

[dnil]

Should also be in a howto; we did take that verbally on the Scout user meeting when it was introduced, but everyone can't be there every time so no excuse for having a short text.

It's whatever-build-you-use-coordinates right now (hg37 for you if you want to use it in Solna) - we will update with a proper build handling soonish. They are global right now, as per discussions on said meetings. I like that, since all will help out collecting, but that could be changed upon new agreement!

Panels is a good idea, I will gladly add that to the filter, but I won't change anything until it sees some use. It was really urgent to have it implemented for some time, and then only a handful of variants showed up during the first months..

[4WGH]

thank you Daniel, we at CMMS are making a list of variants we want to upload so we are going to use this function 🙂. Therfore I'm asking so many questions 🙂. We try to avoid to see not case relevant variants therefore the question about the panel. Good to know is going to be a possiblity in the future to make them panel filtered.

One more question: Some variants are common polymophims (with a clinical relevance for us) This make me think that some of these variants are now not even uploaded in Scout. Once we make them "managed variants" will they be uploaded? Ingengerd had to manually look for them in a case as they were not uploaded and this is what we hope to overcome making this managed variant list 🙂

Soon you will have a ready "howto" just putting together all your answer to my question 😁

---

Från: Daniel Nilsson @.> Skickat: den 9 april 2021 11:08 Till: Clinical-Genomics/scout @.> Kopia: Michela Barbaro @.>; Author @.> Ämne: Re: [Clinical-Genomics/scout] Managed variants - question (#2508)

Should also be in a howto; we did take that verbally on the Scout user meeting when it was introduced, but everyone can't be there every time so no excuse for having a short text.

It's whatever-build-you-use-coordinates right now (hg37 for you if you want to use it in Solna) - we will update with a proper build handing soonish. They are global right now, as per discussions on said meetings. I like that, since all will help out collecting, but that could be changed upon new agreement!

Panels is a good idea, I will gladly add that to the filter, but I won't change anything until it sees some use. It was really urgent to have it implemented for some time, and then only a handful of variants showed up during the first months..

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/Clinical-Genomics/scout/issues/2508#issuecomment-816540499, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AC7R5EI6JD4RN7MBKPQBL2TTH27Y5ANCNFSM42UQK3SA.

[dnil]

😁 Yep - we can just refer to this thread instead!

I don't quite understand the part about avoiding seeing variants not in your panel: these are not yet connected to a case, so seing them in this list should not be too stressful? The "managed variants" page will not link to cases. You can find any matches on individual case pages, and even then they will be hidden unless you open the little tab with "Matching managed variants" exactly as "Matching causatives from other cases". See pictures here: https://github.com/Clinical-Genomics/scout/pull/2155

I'm very close to going on a rant about incidental findings policy, the autonomy principle and the ASHG vs ESHG line of dealing with secondary actionable findings, but I think you are well aware of this.

Yes, managed variants will always load - for cases loaded after the variant was added, obviously. It gets the same priority bypass as MT-variants, clinvar P/LP etc.

[dnil]

I wrote something brief in the docs as well; will go live with the next release. Feel free to ask again if there is anything unclear! Looking forward to many interesting variants being managed! 😄

[ielvers]

Hi :) When should the managed variants start appearing in the list of called variants? I added a variant through the https://scout.scilifelab.se/managed_variant submission page this spring and I'm noticing it's still not appearing in the list of SNVs in newly run cases.

Variant: chr14:88407888A>G (c.1637T>C, a pseudodeficiency variant in GALC) In case 21236 I can see in IGV that the index case and one parent are both heterozygous for this variant, but when I search for GALC in the SNV list no variants appear. I can also see the variant in IGV in case 21231, but it doesn't appear in the list of SNVs. These two cases were run in August and September, respectively. According to my notes I submitted the variant on May 3rd. Did I make a mistake when submitting the variant?

Thanks!

[dnil]

Hi again!

The managed variants feature is enabled, though it hasn't seen that much use, so could very well have issues still. Works on test variants though, so let's check what happens here.

First things first, I do not see this variant in 21236 (gladurchin) SNV VCF files, neither in the selected nor full "research" set. There is no way this variant has made it above 40% in the population? Otherwise, it may be a caller issue. But yes, at least at GnomAD it is now at 43%. That would mean Scout never sees it at all. Options to fix this would include changing the "normal population" cutoff for MIP, or letting MIP ask Scout for the current set of managed variants, and include them. Or just vote this one pathogenic in Clinvar if you have data on it - its without conflict, two star and multiple centers there? Does MIP much as Scout already let those through @henrikstranneheim @jemten?

[dnil]

Or like Modifier, if its not actually Pathogenic on its own.. 😊

[ielvers]

I thought the point of the managed variants was that they would be included, even though they normally wouldn't meet the criteria for inclusion? This variant is not pathogenic or a modifier, it's a pseudodeficiency variant that has no effect in the patient but in the standard lab test it will look like there is very little enzymatic activity. So knowing if the patient has this variant would help us interpret biochemical data (and rule out that specific disease). I have at least one more variant like that, a common variant that we sometimes need to check to interpret the patient's biochemistry. So this is not a one-time issue. :)

[dnil]

They would be included if they reached Scout. But I don't think anyone imagined 43%. Your use case is valid of course, so perhaps we can work with the pipeline developers to fix a pass through - either using managed variants or like a pseudo deficiency database? I don't know how common those GAA and ARSA variants are, but given that those are effectively "normal variation" I guess there is no limit really.

[ielvers]

The PEX6 3'UTR variant I added last week has an allele frequency of 0.65 (https://gnomad.broadinstitute.org/variant/6-42931627-GTTTA-G). It's not disease causing in itself, but it alters the expression of the gene and if present on only one allele it will lead to allelic imbalance. If a patient is heterozygous for a pathogenic variant and has this overexpression variant on the same allele (and not on their healthy allele) they will actually be sick even though it's a recessive condition. So we need to detect it (now we search manually if we find a potentially pathogenic variant in PEX6).

The other variant I want to add is a quite common too, minor allele frequency 0.35 according to gnomAD (https://gnomad.broadinstitute.org/variant/2-234668879-C-CAT). It can be seen as disease causing or not depending on the definition of disease (Gilbert's disease). This is the most common variant causing the condition, and knowing if the patient is homozygous for this variant will help interpret biochemical data and sometimes rule out a more serious condition.

I need to know that I write the variants correctly when adding them. Should a deletion of positions chr6:42931628-42931631 be added as that or as chr6:42931627GTTTA>G?

[dnil]

I've talked to Anders and Henrik. The simplest solution would be to up the hard cutoff to like 0.7. The immediate downside is increased compute time, increased time for the case uploads for the production personel, and also some uncertainty about how the compound scoring in GenMod will cope with additional low scoring "compounds". They will give it a try at some point, possibly already for the MIP 11 release depending on time.

The variant notation should be exactly as in VCF files / what you see on the variant summary, so as you have before. The coordinates are (obviously, I guess) on the reference genome, so will encompass five bases (like pos 42931627 end 42931631 ref GTTTA alt G) or one base (234668879 234668879 C CAT) in your examples.

[ielvers]

Thank you! I'm looking forward to hearing what solution y'all come up with.

Regarding the variant notation, I don't know how it would be written on the variant summary since the variants never appear there which is why I want to add them :) But good to know that they should be written as GTTTA>G and not TTTA>nothing.

[dnil]

Good, I'll reopen this issue and transfer it over to MIP now; as far as I know everything is prepared in Scout. This way you can follow the development!

[henrikstranneheim]

Resolved