DPL-757 As PSD I want to create a first draft design for RVI new project (multipick) that will reflect lims architecture and required components to solve the current list of user requirements obtained.(S=M; C=L)

[emrojo]

Description
As PSD I want to create a first draft design for RVI new project (multipick) that will reflect lims architecture and required components to solve the current list of user requirements obtained.

Who the primary contacts are for this work
Andrew Sparkes
Stuart McHattie
Eduardo Martin

Knowledge or Stake holders
Other people that may have specific knowledge about this work or have a stake in how it is implemented. e.g. John Smith is an expert on x

Additional context or information
diagram

The goals for the RVI consolidation work can be broken down into three parts:

Sample selection
Sample selection will likely be based on use of sample metadata and may require use of an algorithm.
Alternatively sample selection can be dictated by results from previous rounds of sequencing and analysis (e.g. detection of a novel virus with poor coverage in a Bait capture experiment would trigger shotgun metagenomics). - non-automated process

Currently actual sample selection logic and required metadata is unknown and will likely change. I (TW) believe this service/s and logic should be supported within GSU rather than inside PSD. Although - we will likely need to provide an interface and/or support.
For now we should consider the what data is required as input for the plate consolidation service - as this will form the output of the sample selection process.

Plate consolidation service (Cherry Picking)
This service will pick a predefined list of samples from a set of source plates onto one or at most 2 destination plates.
Samples will have been pre-stamped into shallow well plates

Requirements

• Source plates & samples may have already be cherrypicked and/or sequenced on other studies/pipelines/cost codes (Heron) - history must be preserved.
• Cherry picked destination shallow well plates may go on to storage in -80C freezers.
• Destination shallow well plates may be subjected to further rounds of automated or manual cherry picking to select samples for specific pipelines and studies. i.e. destination plates may become sources plates
• Bed verification is a requirement.
• Separation of this service from SequenceScape / Limber is desirable. Preferably via RabbitMQ - although this maybe PostMVP.
• A front end UI to guide the lab user through the process and flag up errors

The data inputs to this service interface are expected to be:

• A set of sample ids, source plate barcodes, and well positions
• Number of randomised wells to be left empty on the destination plates for controls (do we need to specify control type too?)
• Destination plate types
• A study id to record the picking operation under
• User who is requesting this
• Submission data in order to automatically submit the destination plates to a given sequencing pipeline (Could be post MVP - if this is possible manually?)
• Volumes?

The data outputs are:

• Destination plate barcodes and contents (inc controls)
• Destination plate types
• Data about any submission made.
• Information about success / failure / errors
• Events to log the operation in MLWH Events DB

Sequencing logic / Downstream RVI pipelines
RVI Samples will likely be sequenced using different methods/pipelines. Most likely in Limber/SequenceScape but currently unknown.

This may be flagged at the stage of the initial cherry picking in terms of defining the submission/s for the destination plate/s (for example plate will be sequenced using Heron and Bait capture and might be described as ‘parallel sequencing’).

Alternatively sequencing might undergo iteration (‘iterative sequencing’: in this case samples would be sequenced using one (or more methods) depending on the results they would then be selected for sequencing with another method (this might require a further consolidation step.)

Whole destination shallow well plates may go on to be submitted to specific pipelines and studies.

Source samples may have already been sequenced on other studies/pipelines/cost codes (Heron) - history must be preserved.

Outputs

• A rough diagram for the top level components of a minimal flexible cherry picking sample service (include connections between applications & databases )
• A set of user stories for implementing such a service (include consideration of error handling and error paths, changes required to the Biosero interface / platform)
• One research story for each sequencing pipeline to be implemented (see RVI diagram)

[TWJW-SANGER]

Deprecating this in favour of a hopefully smaller research story DPL-779