Add validation preventing ordering mip cases with samples as "unknown"
Closed this issue · 11 comments
As a production user,
I want there to be validation preventing customers from ordering cases which contain samples with status "unknown",
So that the corresponding mip analyses can successfully complete without errors.
Work impact
Answer the following questions:
- Is there currently a workaround for this issue? If so, what is it?
- No.
- How much time would be saved by implementing this feature on a weekly basis?
- This is a somewhat rare issue as most of the orders are placed correctly. But when this does happen, the analyses start and run but cannot complete.
- How many users are affected by this issue?
- All production users and customers. Analyses are also run unnecessarily as in the end, they cannot be completed.
- Are customers affected by this issue?
- Yes, the analyses they request cannot be completed.
Acceptance Criteria
- It is not possible to order a mip case containing all samples with status "unknown".
Question to @Clinical-Genomics/prodbioinfo :
Should MIP-cases with ANY status "unknown" sample be prohibited or MIP-cases with ONLY "unknown" sample
Just adding a note here: mip-dna can handle WGS samples with sex=unknown, it is exome samples where this is problematic. @jemten is it possible to add the same functionality to check gender in the analysis for exome samples?
I think @beatrizsavinhas is referring to setting the phenotype to unknown. I would have to double check but I think the problematic part is having all samples set to unknown. If the phenotype in one of the samples in a case is set to unknown and the other are affected and unaffected it should be fine. I would have to double check though. I agree that this is a check we should have in the pipeline as well. Currently the error message in MIP isn't clear at all.
Oops, completely misread this one! sorry @beatrizsavinhas 🙈
@Vince-janv @jemten We had another case today, closingshepard, and it cannot handle if there is ANY sample in a case with status unknown.
@Vince-janv @jemten We had another case today,
closingshepard, and it cannot handle if there is ANY sample in a case with status unknown.
Alight, that sounds bad. I'll look into it
Un update and correction on this: closingshepherd ran and completed with one of the samples having status unknown (we found that it was actually failing due to incorrect gender assigned for the father sample).
So I think the validation should prevent creating cases with ALL samples as unknown(at least one sample should have a affected or unaffected status).
@jemten agreed with the above statement. So I updated the acceptance criteria and move the issue to Technical refinement!
A new case in prod has showed up with this issue: stirringcalf
We will schedule a meeting to discuss this and https://github.com/Clinical-Genomics/cigrid-ui/issues/722 @jemten @karlnyr
After a meeting with @diitaz93, @beatrizsavinhas, @jemten and @karlnyr we decided to, for now, disallow ordering of mip-dna cases where all samples have the status unknown.
This rule has been implemented in the ordering validation project and will be merged to master in the coming weeks. #4001.
@karlnyr agreed to look inte mip-cases ordered with no affected samples to see if we can make the rule even stricter