search

Clinical-Genomics / scout

VCF visualization interface
https://clinical-genomics.github.io/scout
BSD 3-Clause "New" or "Revised" License
152 stars 46 forks source link

Provisionally solved #3937

Closed dnil closed 8 months ago

dnil commented 1 year ago

There is a wish from users to be able to stratify cases on the confidence of resolution. Sometimes, variants of unknown or tentative significance are reported out, with caveats regarding their lack of use in prenatal testing, cascade carrier testing etc, asking for further functional experiments or segregation testing.

Stratification based on current Solved status plus ACMG or Variant tag might be possible, but it's felt a distanct status would be useful.

northwestwitch commented 1 year ago

More and more I feel that we should allow users to use custom tags on cases..

dnil commented 1 year ago

In a way they can already with the cohort tags..

northwestwitch commented 1 year ago

In a way they can already with the cohort tags..

True. Would that be enough to satisfy the request of this issue?

dnil commented 1 year ago

Well, if one would add tags for the current statuses plus the new one it might, but we would probably have to script the addition of the current tags in that case. I think the main idea is to be able to pull nice live graphs showing these as one area of a graph where the others are unsolved(archived) and solved.

northwestwitch commented 1 year ago

Well, if one would add tags for the current statuses plus the new one it might, but we would probably have to script the addition of the current tags in that case. I think the main idea is to be able to pull nice live graphs showing these as one area of a graph where the others are unsolved(archived) and solved.

Custom tags looks like an easier task I think. Could have many applications as well. One could define a set of tags that are standard (all the existing statuses) and then users could define their own

dnil commented 1 year ago

I'll bounce that around a bit; we would create a bit of extra work for the analysts, if they have to remember to add that as an extra tag in addition to the rest. But the flexibility of it is alluring. Status levels are pretty rigid: we will introduce complexity and risk of bugs with every new one.

northwestwitch commented 1 year ago

Status levels are pretty rigid: we will introduce complexity and risk of bugs with every new one.

Agreed!

dnil commented 1 year ago

Ok, I've chatted around a little. If we add the new status tag to the mark as causative dialog, or change status dialog, so the user is directly prompted with the choice, not having to remember to do it later, it would be an ok added effort for the gains in statistics stratification.

The tag could be prepopulated with perhaps "Provisional" and "Diagnostic", (I think we can phase out that status later), and at some point we should update the Dashboard to stratify the statuses with these.

It is very tempting to prepopulate this dialog choice with "Diagnostic" IFF there are ACMG variants class 4 or 5, or variant tags at LP and above (KP, P, LP, NVP, PPP, ...), and "Provisional" IFF there is a class 3, tag PP or US. We can't require the existence of such a variant today, as we still can't classify all variants (UPD, SMN, ...) but at some point perhaps.