Y24-376 - Pooling algorithm changes

[KatyTaylor]

User story As a user of the scRNA Core cDNA Prep pipeline, I would like some changes to the pooling algorithm (allocating samples into wells on the LRC PBMC plate) to increase the flexibility and robustness of the pipeline.

Who are the primary contacts for this story Katy, Abdullah, Andrew

Who is the nominated tester for UAT Abby

Acceptance criteria Pooling algorithm (creating LRC PBMC Pools plate) is adjusted so that:

• [x] 'Number of pools' is pulled from the submission/requests (for each study-project group), rather than being calculated.
• [x] All pools within a study/project/donor group have the same number of samples (or differ by one if not exactly divisible).
• [x] The requested number of pools can vary between study-project groups in the same run.
• [ ] The result of the algorithm is shown to the user in a success message after the plate is created (see format in additional context).
• [x] Plate creation fails with an informative error if any of the following conditions are not met:
  • [x] Total number of pools on the destination plate must be between 1 and 8 (inclusive).
  • [x] Number of samples in each pool must be between 5 and 25 (inclusive).
  • [ ] Each pool should have unique donor IDs.

References Follows on from:

• https://github.com/sanger/limber/issues/1909

'Number of pools' won't be present in the request metadata until https://github.com/sanger/sequencescape/issues/4423 has been done. However, this story can still be done first - we would need to do the request metadata migration as part of this story.

The 'number of samples per pool' field added to request_metadata as part of https://github.com/sanger/sequencescape/issues/4337 is now obsolete - so we could change the name of this field rather than adding a new one?

Additional context Output of the pooling algorithm is (e.g.), to display to the user: Study/project/donor group	Pools (num samples)	Well
Study X - Project A - Donor group 1	4, 4, 3	A1
Study Y - Project B - Donor group 2	10, 10	B1

[KatyTaylor]

Have linked Abdullah's draft PR where he was playing around with algorithm changes to solve "All created pools have the same number of samples (or differ by one if not exactly divisible)."

[KatyTaylor]

• User interface ideas to prompt conversation - https://docs.google.com/presentation/d/1abZ2bPjxToNOuLkUsct2fpSFzojz3rYTC6Re7BgIRW0/edit?usp=sharing

• Pooling algorithm “rules” - https://lucid.app/lucidchart/481ca556-9b51-4a07-b6e2-47e29d05f6c5/edit?viewport_loc=-264%2C48%2C2505%2C1267%2C0_0&invitationId=inv_bc2e91f5-40ea-4d18-95ef-622184c31735