GPL-710 As developers we want to ensure that when the destination plate is created that the appropriate data is generated in Sequencescape

[stevieing]

User story
When the destination plate has been created send the destination plate barcode to sequencescape and create a plate with information for each sample pulled from either dart database, mongo database or the warehouse and also generate a COG UK barcode.

Acceptance criteria

• LIMS endpoint accepts a destination plate barcode
• Destination plate is generated in Sequencescape, with following additional data:
  • Wells
  • Samples
  • Sample metadata - Root Sample ID, Result, COG UK ID, RNA ID (this one is new, as would normally be able to track this back via the source plate)
• Plate purpose TBC (see comment below)
• Any error information associated with the run(s) that we can get from the DART database is recorded in Sequencescape (this will actually happen in the Event warehouse as events)

Additional context
Could generate COG UK barcodes earlier, e..g in step 1, for filtered positives, and store in MLWH

See diagram with options here - https://app.lucidchart.com/lucidchart/c029ead4-12c9-4baf-acb2-a2a3fbead2fc/edit?page=0_0#?folder_id=home&browser=icon

[KatyTaylor]

I think that currently, the resulting plates from the cherrypicking are:

• LHR RT (96 pipeline, non-tailed)
• LTHR RT (96 pipeline, tailed)
• LHR Cherrypick (384 pipeline, both)

It sounds like with the new Beckman robotics cherrypicking, we won't know which pipeline it's going down at this stage.
So we need to decide / ask what plate purpose we should be creating for the destination plates.

Options:
If we go with LTHR RT, they can go straight down the 96 pipeline as normal but there will be a change in process for the 384 - quad stamping 4 LTHR RTs instead of LHR Cherrypicks

If we go with LHR Cherrypick, the quad stamping for 384 can work as normal, but for 96 we will have to either switch the plate type / stamp into a new one, or modify the Limber pipeline

Or create a new one?

(I think going with LHR Cherrypick would probably be best, because the name makes sense for both situations)

See https://github.com/sanger/General-Backlog-Items/issues/77 as is relevant

[emrojo]

Tasks:

• Add a new blueprint to Lighthouse for Beckman cherrypicking, with one endpoint receiving a plate barcode
• Connect to DART(SQL server)
• Create query in DART Access a view in DART to get all wells belonging to a destination plate_barcode from request and join with all source wells (Waiting for more details about it)
• Create a query in Mongodb to get all samples information from the samples at destination wells
• Combine info into a Heron plate creation message
• Make sure COG-UK ids are created and sent to the warehouse
• Send message to SS to create plate, using plate purpose LHR Cherrypick LHRT Cherrypick