Y24-080 scRNA Research - Aggregation may need to be bypassable

[andrewsparkes]

Description As PSD we need to understand whether the users will want to be able to bypass aggregation in the scenario where there are fewer source plates. See additional context for aggregation process. In existing Chromium pipelines the SSRs will bypass the aggregation process and create a Cherrypick plate directly (exact method unknown, possibly via fake SS cherrypicking), if there is only one or fewer than 10 source plates. i.e. it is not worth the effort of creating the intermediate aggregate plates. Rather than have the SSRs have to hack the creation of the Cherrypick plate we would like a clean configured solution to create the Cherrypick plates direct when we have a smaller number of source plates (when <10).

Who the primary contacts are for this work SSR (TBC which SSR does this in existing chromium pipeline)

Knowledge or Stake holders R&D Abby and Lesley

Additional context or information Aggregation is a 2 step process limited by the liquid handler robots capacity for a maximum of 10 source plates. It is a process intended to combine small numbers of samples from multiple source plates into 1 Cherrypick plate. The Cherrypick plate is then the start point for a Library prep pipeline. i.e. up to 96 individual samples from 96 separate source plates.

Step 1: create an Aggregation plate from up to 10 source plates. (LRC HT 5p cDNA Input or LRC HT 5p cDNA PCR XP source plates into LRC HT 5p Aggregate plates) Step 2: create a Cherrypick plate from up to 10 aggregation plates. (LRC HT 5p Aggregate to LRC HT 5p Cherrypick plates)

Possible solutions

1. shortcut aggregation: Still using Aggregation submission. Allow creation of a Cherrypick direct from 1-10 source plates (LRC HT 5p cDNA Input or LRC HT 5p cDNA PCR XP source plates into an LRC HT 5p Cherrypick plate). Would need pipeline configuration modifications and a new variation of the robot bed verification to allow this.

2. Sequencescape cherrypicking: Allow creation of Cherrypick plate from sources using Sequencescape cherrypicking submission. This is outside Limber and requires SSR intervention to set up the Cherrypick.

3. others?

[KatyTaylor]

Some information on the existing 10x pipelines here - https://ssg-confluence.internal.sanger.ac.uk/display/PSDPUB/Chromium+10X+Pipelines+Documentation

Includes this paragraph:

An aggregate submission is made on the LBC Stock plate

• The aggregation pipeline combines multiple LBC Stock plates into an LBC Cherrypick plate
• They looked into doing normal cherrypicking within the Sample Prep team, using the SS UI, but a problem was that the cDNA was already amplified so this didn’t fit the normal process
• The cherrypicking robots are in the Sample Prep team, so they can’t use the SS cherrypicking UI for other teams (add more robots that can be used for SS cherrypicking?)
• If they could use the SS cherrypicking UI, this could help the process as they could do the admin / paperwork for the cherrypick ahead of time, so there wouldn’t need to be a pause

I'm not sure what the significance is of the cDNA being already amplified - a lab process constraint? I get the thing about the Tecans though.

Liz H is our SSR contact for 007c.

[KatyTaylor]

Comment from Abby - looks like improvements to this part of the process are not planned until 'Phase 3'.

For phase 2 (~due around September) we were just planning to take the exisitng process as is. For phase 3 (unlikely to happen this year) we would look at improving it. For this particular step, for phase 3 we would actually plan to remove the need to aggregate at all by increasing the number of plates that can be picked at once during cherrypicking (could be done on the current protocol/liquid handler). This is something that could apply to both 007c and the existing pipeline (with the necessary LIMS changes). We can also readress the amplified cDNA issue preventing the use of the Tecans however I think because of the step before aggregation (cDNA clean up), this may force us out of the sample prep lab anyway due to a lack of appropriate instrumentation.

[andrewsparkes]

'Existing process as is' involves fake cherrypicking by SSR, and an invisible (to LIMS) control in H12.