Y24-035 Targeted NanoSeq - update configuration for pipeline and add submission for ISC prep

[andrewsparkes]

User story As PSD we would like to fix the targeted nanoseq pipeline configuration to better reflect the reality of what the lab users are doing, and allow them to loop back to the LTN AL Lib and process different subsets of samples as many times as they like.

Who are the primary contacts for this story Sara W and team Andrew S

Who is the nominated tester for UAT Sara W team

Acceptance criteria To be considered successful the solution must allow:

• [x] update config for Targeted NanoSeq to better reflect reality
  • only re-ISC end route is required, remove other duplex seq route
  • shearing may or may not be done at the beginning, leave the two routes for that part
• [x] allow the submission for re-ISC to be done on the LTN Lib PCR XP plate rather than require a cherrypick into an LB Lib PCR-XP plate - this is likely to require the plate to be added to the acceptable plate purposes for the re-ISC submission template (SS config change and rails console fix in databases), plus add the relationship for transfer from LTN Lib PCR XP plate into LB Lib PrePool plate
• [x] add an automated close off of the library prep requests at creation of the LTN Al Lib plate (use :state_changer_class: StateChangers::AutomaticPlateStateChanger and :work_completion_request_type lines in the purpose config) - [Edit: maybe Ok to use Charge and Pass as currently. But we MUST close these before continuing.]
• [x] add an automated submission at creation of AL Lib Dil for a customer request 'name TBC' [Edit: see comments below - have decided manual is the way to go, given the submission the name: Targeted NanoSeq ISC Prep, but TBC]
• [x] automatically close the new customer request at creation of the LTN Lib PCR XP plate
• [x] improve code that identifies correct request to use when inserting poly-metadata on creating the LTN Al Lib Dil plate, as it currently only looks in aliquot.request, needs to look in the receptacles.requests_as_source for the newly created active requests [CORRECTION: no change required as it looks in aliquot.request on the CHILD well, which is right]
• [x] update the integration suite test for this pipeline to match what the users are doing (subsets of samples and a loop back)

Dependencies None

References This story has a non-blocking relationship with:

• 1628

Additional context Currently the centre part of this pipeline where they loop back to the LTN AL Lib plate is not tracked with a submission. Plus all samples are tracked as moving from LTN Al Lib plate through to LTN Lib PCR XP despite only a subset physically being transferred through. Plus they are editing the Hamilton driver file to remove the unneeded samples manually.

[andrewsparkes]

Options for manual vs automatic submissions:

1. Create a manual submission on the AL Lib for the samples. Then check at Customer upload for Dil creation that all wells with requests are present in file. i.e. changes to file upload validations:

• requests for library prep must be closed off on the plate (user must have charged and passed the library creation requests)
• must be a single active Submission of correct type on the plate (i.e. check SSR has created the Submission)
• count of wells with values in file must match count of active requests on plate
• locations of wells with values in file must match to active requests on plate (ie. A1, C3, D2 in both) with no omissions or extras

Then create the Dil plate, transferring only those wells with requests (partial transfer). Changes:

• labware creator to extend or use partial submission code to allow subset of samples to progress
• WellFilter to be partial version that selects based on active requests of right type in requests_as_source field on well

Pros:

• SSR is in control. And is handling sending out and receipt of Customer file anyway.
• Uses existing Limber code and infrastructure, cheaper to develop.

Cons:

• extra overhead for SSR to create the 'ISC Prep' Submission for each loop.
• MUST close off Submissions from previous loops before going back to the Al Lib plate or could get in a mess.
• SSR must create the new Submission before lab uploads the Customer file to create the Al Lib Dil plate.

2. Create auto submission on upload of the customer file Based on which wells have values. Wait for requests to be created by SS, then create the Dil plate. Is this a viable option? Does Heron do this? Not exactly. Heron uses SubmissionPlatePresenter which displays submission options to choose on source plate. This is not the same, as we need a partial submission based on the file upload contents. i.e. in the labware creator rather than the Presenter.

If we do it at file upload in the labware creator, we will need to do extra validations:

• Are there already open requests on the plate? create Error - these MUST be completed or failed or cancelled.

We will then need to filter the wells AFTER the submission is created to make the transfers. i.e. steps in the labware creator are:

• validate the file uploaded
• create the submission based on wells with values in the file
• wait for the submission to be made [THIS MAY TAKE SOME TIME- what if we timeout??]
• create the empty child Dil plate
• filter the wells by request to create the transfers for the Dil plate

Cons: This idea of creating a Submission inline with a labware creation in Limber is not currently done anywhere else. Reason being submissions can take time to process (as async jobs in SS) so labware creator would be keeping you waiting as it polled SS for updates on submission creation, before it could create the child plate. How to code for a time out?

Summary Favouring the manual Submission option due to complexity of automated version. Should check with SSR if they are Ok with the extra Submission needed for each loop. Mitigated by less hacking required in other parts of this pipeline after these changes.

[andrewsparkes]

Integration Suite test branch here: https://gitlab.internal.sanger.ac.uk/psd/integration-suite/-/merge_requests/129

[andrewsparkes]

For deployment:

Need to check that the LTN plate purpose is added to the acceptable_purposes for the reISC request type by the deployment (record loader has update for this so may be fine). Add it manually after deployment if not set up.

In rails console: check and add the LTN purpose to Reisc

rt = RequestType.find_by(name: 'Limber REISC')
rt.acceptable_purposes
pp = PlatePurpose.find_by(name: 'LTN Lib PCR XP')
rt.acceptable_purposes << pp

[andrewsparkes]

Waiting on lab team for UAT

[andrewsparkes]

Meeting for Monday 17th June to discuss UAT testing

[andrewsparkes]

UAT testing is completed, no additional concerns. Happy to deploy to live. Now just waiting on users to agree a date (as this is a change to an existing pipeline).

[andrewsparkes]

To be deployed to live on Thursday 25th July at 1pm.