Bioinformatics QC page & simplfy analysis finalise

[erikwaskiewicz]

Minor Software Update

Description

• Change the drop down where sample analyses were closed to have a series of drop downs so that less mistakes are made
  • Added send back button with breakglass
  • Close sample button now pops up a modal
  • Modal will do checks that used to happen on submit as a ajax calls, will only carry on to next step if the checks pass/ will limit options if they fail
• Added a bioinformatics QC step before results are released
  • idea is that they can be auto uploaded here but only visible to scientists once QC is done
  • added permission so only bioinf team can do the QC
  • Added form to duplicate info from AutoQC, in future would like to automate this
  • Added bioinf QC form to simplify the bioinf QC step
  • This hasnt been fully tested so might be a bit clunky, might try to drop the patient name at the same time as deploying this to simplify further
• Simplified worksheet status so it'll hopefully be quicker to display

Outstanding:

• tests have been broken - fixtures changes and more model functions mean logic has moved, will need overhaul
• Hasnt had really detailed testing, particularly bioinf QC bits

Justification

Not changing the logic, just how the database functions

---

Testing

Description :

Ongoing, see above, just an interim code review please

Data Location:

N/A

---

Checklist

< Place an X in the appropriate boxes, if not applicable state N/A >

• [ ] Is the test data easily accesible?
• [ ] Has the pipeline/software been run end to end by the tester?
• [ ] Has the pipeline/software been code reviewed on github?
• [ ] Has the pipeline/software version number been incremented?
• [ ] Are all relevant doccuments updated e.g. SOP, ReadME, Pull request

Additional checklist for pipeline updates (if applicable)

• [ ] Is the sensitivity or other key performance metrics unaffected by the changes?
• [ ] If the environment changes has this been deployed and tested in place?
• [ ] Has the pipeline version in the variables file been updated e.g. development > master

[Istaisa]

Doing this review as a comment per chunk of testing. User testing - bioinformatics QC checks (note have not checked underlying code yet)

User Groups / Display

• Users not in the QC group can see what needs bioinformatics QC but cannot click on to the runs
• Users in the QC group can click on the runs and perform QC

COMMENTS/ISSUES

• Add instructions to the docs on how to set up the QC user group locally

Performing QC Checks

• I really like this section I think it's going to make failing more straightforward and querying better in the future
• Have tried combinations of run pass all samples pass, run pass some samples fail, whole run fail
• If we reopen QC when analysis has already started it keeps all the scientist checks which is good

COMMENTS/ISSUES

• Can't edit the patient name within the bioinformatics QC page (but losing it altogether might be the easiest way to solve this)
• You can use the same AutoQC primary key for different runs - do we want this or should we force it to be unique?
• AutoQC URL needs changing: should be 10.59.210.245:5000/run_analysis/
• Samples aren't finalising properly from the bioinformatics QC checks. The status in the table changes from bioinformatics check 2 to bioinformatics QC fail, but when you go in to the sample you've still got the check interface open and can't download the report/XML

[Istaisa]

User testing - updated pass/fail interface (note have not checked any of the underlying code yet)

General Interface

• Can't finalise at all without checking patient demographics and updating name
• Can't pass RNA without checking all the fusions
• Can't pass DNA without doing all the DNA checks
• Send back check button works as expected

Passing Samples

• If everything agrees after 2 checks it closes nicely
• If last 2 checks agree after 3 checks it closes nicely

COMMENTS/ISSUES

• Will finalise if user 1 and user 3 are the same person, or if user 2 and user 3 are the same person which means it can finalise with some variants only having been checked by 1 person

Failing Samples

• The status of a sample awaiting a fail is now 'IGV check 2' and the button's less obvious - this might be the intended behaviour though?
• 2 fails works correctly
• pass then fail then fail works correctly (yay!)
• fail then pass then pass works correctly
• fail then pass then fail then fail works BUT checker 3 and 4 can be the same person (see above - same thing with 2 checks rather than 2 checks by independent users)
• scientists can't reopen a completed fail sample - i can't remember if we only added that functionality in on complete samples for a reason

COMMENTS/ISSUES

• Overall very clean but would be good if it was 2 independent matching checks not just 2 matching checks
• This is more of a GUI one - but would it be possible to make some of the tickboxes a little bit bigger? e.g. on the demographics page tickbox. Lots of dexterity required on a big screen!

[Istaisa]

My unit tests do work - don't know if that's concerning if they're meant to have broken