Change labman/labman/db/process.py generate_echo_picklist() call to _format_picklist() to pass in 20000 (nL) as max_vol_per_well

[charles-cowart]

From RodolfoSalido:

I quantified my Shotgun libraries and generated a pooling file using Lab Control. I only noticed one discrepancy between the pooling file generated by Lab Control and that generated by our Jupyter notebook. Our Jupyter notebook is designed to pool into a destination well until it reaches a certain volume (30µL) and then it cycles to the next well. The Lab Control pooling function seems to have a value of 60µL for the max_vol_per_well parameter instead of 30µL.

[charles-cowart]

From RodolfoSalido's output:

283 rows do not match between the two outputs, out of the total of 384. Of the rows that don't match, they appear to differ by 1/100th of a unit in transfer volume, and the destination wells differ. Awaiting confirmation from Rodolfo.

File A:
1,384LDV_AQ_B2_HT,B2,,262.69,NormalizedDNA,A1
1,384LDV_AQ_B2_HT,B8,,363.07,NormalizedDNA,A1

File B:
1,384LDV_AQ_B2_HT,B2,,262.68,NormalizedDNA,A1
1,384LDV_AQ_B2_HT,B8,,363.06,NormalizedDNA,A1

[charles-cowart]

Confirmed - here are the transfer volumes summed for each of the destination wells for each file:
From Rodolfo:

The real granularity of our liquid handler is 2.5 nL increments do 1/100th unit discrepancies are not a problem. If you sum all the volumes that go into any specific destination well, the addition shouldn’t exceed 30 micro-liters.

Confirmed by summing the transfer volumes in each file across their destination wells:

FIle 1:
A1 = 29825.000000
A2 = 29912.560000
A3 = 29770.010000
A4 = 24522.410000

File 2:
A1 = 59737.560000
A2 = 54292.360000

[charles-cowart]

It looks like the parameters set up in 'HTML_POOL_PARAMS_SHOTGUN' set 'blank-vol-' to empty string for the 'equal' pooling function. This in turn sets 'PoolingProcess.compute_pooling_values_eqvol().total_vol' to 60ul by default.

Perhaps changing the maximum total volume of the destination well to 30ul is a matter of setting the 'Pool all blanks at volume' property. Testing on maptest.

[charles-cowart]

Issues w/maptest when leaving this page and coming back to it. The robustness of this page needs to be addressed (See #361)

[charles-cowart]

Just to reiterate, I believe #361 should be resolved first, as it's not easy to get back results to confirm the correctness of a bug fix with the known beta issue.

[AmandaBirmingham]

TL;DR: change

https://github.com/jdereus/labman/blob/a6777cc8b29946c06599b5d3e78a5aeaf4541427/labman/db/process.py#L2559

to

return PoolingProcess._format_picklist(vol_sample, max_vol_per_well)

and this problem should be fixed. Explanation below.

---

I believe Rodolfo pointed us on the right road here with his reference to the variable name max_vol_per_well. I see PoolingProcess has this function:

https://github.com/jdereus/labman/blob/a6777cc8b29946c06599b5d3e78a5aeaf4541427/labman/db/process.py#L2497

Note the default, which is in nL:

https://github.com/jdereus/labman/blob/a6777cc8b29946c06599b5d3e78a5aeaf4541427/labman/db/process.py#L2505-L2506

60000 nL = 60 microliters (μL) :)

This method is called by generate_echo_picklist:

https://github.com/jdereus/labman/blob/a6777cc8b29946c06599b5d3e78a5aeaf4541427/labman/db/process.py#L2542-L2548

Here, the default is 30000 nl or 30 μL, which is the value the lab actually wants to use, so you'd think we'd be golden, except ... the max_vol_per_well value IS NEVER USED. Instead, _format_picklist is called without an explicit max_vol_per_well value:

https://github.com/jdereus/labman/blob/a6777cc8b29946c06599b5d3e78a5aeaf4541427/labman/db/process.py#L2559

Therefore the 60 μL default in _format_picklist takes over, and Bob's your uncle ...

[AmandaBirmingham]

From: Rodolfo Salido
Date: Monday, April 15, 2019 at 11:18 AM
To: "Birmingham, Amanda"
Cc: Charles Cowart , Daniel McDonald
Subject: Re: Shotgun Libraries pooling.

Hello Amanda,
We don’t need this parameter to be editable, we can have it hardcoded. Karenina has been using 20µL as her default and I’ve been pooling 30µL. We changed it from 60µL to a lower 20-30 µL because we changed destination plates and the newer plates hold less volume.

We don’t anticipate any future changes to our pooling procedure. After talking with Karenina, I believe that going with the more conservative 20µL max_vol_per_well will be the better choice.

-Rodolfo

On Apr 15, 2019, at 11:09 AM, Birmingham, Amanda wrote:

Hi—
I’ve looked into this, and I am 90% sure I know where the issue is coming from (I’ve put details on #438
). Making the code use max_vol_per_well = 30µL instead of 60µL should be the work of a moment (although updating the tests may take a little longer) ☺.

HOWEVER, I am also 90% sure there is absolutely no way for a user to change this value through the GUI—it is hardcoded. Rodolfo, is max_vol_per_well a setting that the lab ever has changed, or anticipates changing in the future? If so, we probably need to add it to the user interface somewhere and hook up the back end to take that input … Lemme know what you think.
Thanks,
Amanda