Subject to verification, it appears that the volume reduction is triggered by https://github.com/sanger/sequencescape/blob/master/app/models/cherrypicking_pipeline.rb#L15 and clearly not in https://github.com/sanger/sequencescape/blob/master/app/models/pipeline.rb#L139. The actual calculation occurs in https://github.com/sanger/sequencescape/blob/master/app/models/cherrypick_request.rb#L19.

Confirmed that the code in app/models/cherrypicking_pipeline.rb::CherrypickingPipeline.robot_verified! is executed on robot verification through Sample Management (Lab View) > Bed Verification.

Once this has been completed, then a Robot verification completed and source volumes updated. event appears at the bottom of the Batch page.

The destination plate is created and updates as soon as the Approve Plate Layout step has been completed.

The confluence page Cherrypicking options in Sequencescape has been updated with this info.

Calculations (and explanations) for volumes can be found in app/models/cherrypick/volume_by_micro_litre.rb and sibling files. This is also where validity checks are for volume being defined. This is likely to be particularly helpful for Y24-383.

Given the table of results from test cherrypicks below, there appear to be some bugs:

Assuming simple arithmetic (which might not be the case), there are some unexpected final volumes.

For plate SQPD-10857 with an initial volume of 50 ul, a 10 ul pick with no buffer, results in a plate of 40 ul and 10 ul - this is expected.

For plate SQPD-10859 with an initial volume of 50 ul, a 10.71 ul pick with 5 ul buffer results in a plate of 39.29 ul (correct), but 15.0 ul in the destination plate - this is expected to be 15.71 ul.

Similarly, for plate SQPD-10861 with an intial volume of 50 ul, a 14.29 ul pick with no buffer results in a plate of 35.71 ul (correct), but 10.0 ul in the destination plate - this is expected to be 14.29 ul.

Plate SQPD-10864 replicates the issue identified in the training SS spreadsheet where the volume pick is less than 1 ul is 1 ul - less than the expected 5 ul required by the Robot Minimum Picking Volume.

Plates SQPD-10866 and SQPD-10868 attempt to replicate the inverse, to get 1 ul of buffer picked. In both cases 50 ul of well volume and no buffer is picked - which appears correct but would result in a greater quantity than expected - which is also correct.

This is largely complete, only the questions below to Liz C are outstanding. The responses to these will tell us if there are any bugs in the calculations:

1. I would like to confirm how the Robot Minimum Picking Volume is meant to work. Should this minimum volume apply to just the wells (v in the worksheet), buffer (b in the worksheet), or both?
2. a. Is the resulting volume in the destination plate important?
   b. If so, if v10.71, b5.00 is cherrypicked (Concentration Pick, Volume Required: 15), should the resulting volume in the destination plate be 15 ul or 15.71 ul?
   c. Likewise, if v14.29, b0.00 is cherrypicked (Amount Pick, Minimum Volume: 10), should the destination volume be 10 ul or 14.29 ul?

A bit of extra detail after reading the code that updates the source well volume:

• When the cherrypick pipeline pages are clicked through, picked_volume on the well_attributes table is set, for the target well.
• When the bed verification happens, this value is retrieved and subtracted from the source well's current_volume (also on well_attributes.

@StephenHulme can you confirm if this is correct?

Given the table of results from test cherrypicks below, there appear to be some bugs:

Thanks for this investigation, it looks great. Looks like there is a bug in the setting of destination plate well volumes.

Plate SQPD-10864 replicates the issue identified in the training SS spreadsheet where the volume pick is less than 1 ul.

That plate is missing from your table, could you add it? Could you describe a bit more about the "issue identified in the training SS spreadsheet where the volume pick is less than 1 ul", please?
I think the issue is that the volume pick is less than the specified 'robot minimum picking volume' - is that correct?

One more question - was the one mentioned above the only issue raised by Liz in her spreadsheet, or are there more to look into?

Thanks.

That plate is missing from your table, could you add it?

Sorry, copy-paste error. Plate SQPD-10864 has been added to the table.

Plate SQPD-10864 replicates the issue identified in the training SS spreadsheet where the volume pick is less than 1 ul.

Correct, the volume pick (1 ul) is less than the specified 'robot minimum picking volume', in this instance of (5 ul). I think a typo crept in there...

was the one mentioned above the only issue raised by Liz in her spreadsheet[?]

As I understand the accompanying email there were two concerns associated with the spreadsheet. The first (I think, though could be wrong) was confirmation of volume and total micrograms changing, rather than a specific bug. The second concern was a potential bug with robot minimum picking volume which I have been able to replicate. While it appears clear that this behaviour is incorrect, there is currently fairly complex logic regarding the minimum picking volume making it challenging to identify the correct behaviour without additional information (see question 1 above).

• When the cherrypick pipeline pages are clicked through, picked_volume on the well_attributes table is set, for the target well.

Yes. More specifically, when the Finish Cherrypick page is reached (but before Release) concentration, current_volume, buffer_volume, requested_volume, and picked_volume on the well_attributes table is set for the target well (for a Concentration Pick when contration and volume data is available).

• When the bed verification happens, this value is retrieved and subtracted from the source well's current_volume (also on well_attributes.

Correct. See screenshots below (identical wells, top 3 are source plate, bottom 3 destination).

When the Finish Cherrypick page is reached:

After bed verification:

Thanks for the answers to my questions. Looks like we're just waiting on some clarification from Liz on your 2 q's then (v or v+b, and destination volume).