Multidispense mode enabling

[Youssefabdelazziz]

Is there an existing issue for this?

• I have searched the existing issues

Feature description

Dear Dr. Fine,

Thank you for developing the Pyalab package. it's an incredibly insightful tool for Integra automation. I am excited to integrate it into my master’s thesis at TUM.

Currently, I see that the transfer mode is supported. However, adding a multidispense mode (either with different volumes or constant volumes) would be highly beneficial. My research involves high-throughput experimental design and machine learning, where randomized dispensing plays a crucial role. Implementing multidispense functionality would significantly reduce dispensing time and improve workflow efficiency.

In my current setup, using single transfer mode, the experiment takes approximately 3 hours, which is impractical for real-world applications. A multidispense feature would greatly enhance speed and automation.

Could you kindly guide me on potential workarounds or plans to support multidispense functionality in Pyalab? I would truly appreciate your insights.

Looking forward to your response.

Best regards,
Youssef Abdelaziz
TUM Master’s Student

[ejfine]

Hello Youseff,

There's a minimal implementation of "multidispense" at the moment. It was used to be able to have more control over the pre/post aspirate volumes (which Transfer doesn't allow) and doesn't actually dispense multiple times yet.

Can you provide some more details of your workflow to be able to better estimate the implementation effort? i.e. source and destination labware types, deck, pipette type, pipetting height

[Youssefabdelazziz]

Hello Dr. Fine,

I’m currently working on high-throughput screening and optimization workflows where I assess 10 reagents simultaneously using a randomized experimental design. This setup demands multi-dispensing of different and repeated volumes across multiple wells.

While the single transfer mode works well for basic tasks, the overall execution time exceeds 2.5 hours depending on the complexity of steps, which becomes a major bottleneck for high-throughput experiments.

To help illustrate the current implementation, I’ve included a simplified code trial for MultiDispense that outlines the setup and pipetting logic.

Setup Summary
Deck Layout: StandardDeckNames.THREE_POSITION.value
Pipette Type: VIAFLO SINGLE 1250 µl
Pipette Tips: 1250 µl GripTip Sterile Filter Low Retention
Source Plate: 96 Well F-Bottom Plate
Destination Plate: 384 Well F-Bottom TC-Treated Plate
Liquid Level Detection: Ideally handled via surface detection. No manual height tuning applied in this version

---

Code:
from pathlib import Path
from pyalab import (
Deck, DeckLayout, DeckPosition, LabwareOrientation, Pipette, Plate, Program,
SetInitialVolume, StandardDeckNames, Tip, MultiDispense,
AspirateParameters, DispenseParameters, PipettingLocation,
)

Define the labware

source_plate = Plate(name="96 Well F-Bottom Plate", display_name="Source Plate")
destination_plate = Plate(name="384 Well F-Bottom TC-Treated Plate", display_name="Destination Plate")

Define pipette and tip

pipette = Pipette(name="VIAFLO SINGLE 1250 µl")
tip = Tip(name="1250 µl GripTip Sterile Filter Low retention")

Program creation

program = Program(
deck_layouts=[
DeckLayout(
deck=Deck(name=StandardDeckNames.THREE_POSITION.value),
labware={
DeckPosition(name="B", orientation=LabwareOrientation.A1_NW_CORNER): source_plate,
DeckPosition(name="C", orientation=LabwareOrientation.A1_NW_CORNER): destination_plate,
},
)
],
display_name="Multidispense",
description="Transfer from source to destination",
pipette=pipette,
tip=tip,
)

Set initial volume for A1 in source plate

source_section_index = program.get_section_index_for_labware(source_plate)
destination_section_index = program.get_section_index_for_labware(destination_plate)

program.add_step(
SetInitialVolume(
labware=source_plate,
section_index=source_section_index,
column_index=0, # A1 = column 0
volume=50
)
)

Multi-dispense step

program.add_step(
MultiDispense(
source=PipettingLocation(
labware=source_plate, deck_section_index=source_section_index,
column_index=0, upper_left_row_index=0
),
destinations=[
(PipettingLocation(
labware=destination_plate, deck_section_index=destination_section_index,
column_index=0, upper_left_row_index=0
), 25),
(PipettingLocation(
labware=destination_plate, deck_section_index=destination_section_index,
column_index=1, upper_left_row_index=0
), 25),
],
aspirate_parameters=AspirateParameters(
liquid_speed=5,
post_delay=1
),
dispense_parameters=DispenseParameters(
liquid_speed=5,
post_delay=1
),
)
)

Save the program

program.dump_xml(Path("/home") / "jovyan" / "docs" / "Multidispense.iaa")

[ejfine]

Thanks for those details!

You mentioned Pipette Type: VIAFLO SINGLE 1250 µl, but to my knowledge that pipette cannot be used on the Assist Plus. Did you mean the larger volume D-One Pipette (Part #4532)? Or are you wondering if it's possible via Pyalab to use a single channel Viaflo pipette?

[Youssefabdelazziz]

Thank you very much for your response.

Apologies for the confusion; what I meant was "the VIAFLO SINGLE 300 µl".

I would like to inquire whether it is possible for Pyalab to perform multidispense with different volumes, or even repeated volumes using the multidispense feature due to the time limitaiton of the experiment.

[ejfine]

Gotcha. Yes, I'll take a look at it in the next few days

[ejfine]

ah, I see --- in the Integra Library XML files, the D-ONE pipette file is called VIAFLO SINGLE 300 µl! that makes sense why you were using that term now

[ejfine]

@Youssefabdelazziz - are you currently using the extra adapter to raise up plates in Position A of the deck? I'm wondering if that'd be required within pyalab for you to start playing with it, or if it could wait until later

[Youssefabdelazziz]

Thanks so much for bringing this up, that’s very thoughtful of you!
We’re not entirely sure if it’s necessary for small tips at this stage. We hadn’t really considered it yet, but we might revisit and adjust it later on.

[ejfine]

I'm working through some of the ability to fully support D-One pipette at the moment. Are you currently using two different tip types within the same Method? Or could supporting that be a follow up feature?

[Youssefabdelazziz]

Hey,
Thanks alot for following up.
Currently, we are not considering changing the pippette tips type in the current protocol and I think it won't be the case for practicality reason.

[ejfine]

Do you use any mixing at the moment? And it looks like you are using a delay after the aspiration and dispense (or was that just random example code)?

[ejfine]

I created the ability to define a program with D-One tips (use the DOneTips object instead of just Tip for your program definition). And updated SetVolume and Transfer to be able to accept a row_index parameter for use with the D-One.

Those things all pass in simulation for me in Vialab, but I don't have access to a D-One to fully test it out.

Can you try it out and see if it works as expected for doing simple transfers? If that all looks good, then I'll know it's a good foundation to start adding the D-One multi-dispense capability on

See #34 for details

[ejfine]

@Youssefabdelazziz - did you have a chance to test the new code with your D-One to confirm it functions in reality as it does in Simulation?