Pypetting

Pypetting provides a simple set of wrappers to write complex worklists for the Tecan EVOware pipetting software. Unfortunately it does not entirely avoid interacting with the original software, but it reduces the necessary interactions with it to a minimum.

Installation

Install using pip with pip install pypetting.

Pypetting has only been tested with Python 3.10.0.

Examples

Define a position on the grid

from pypetting import GridSite

site = GridSite(grid=36, site=1, carrier="MP 3Pos")

Define labwares

from pypetting import Labware

greiner96 = Labware(name="96 Well Griner", rows=8, columns=12, spacing=1)
trough100 = Labware("Trough 100+25ml", 8, 12)

Pipetting from one column to the next

from pypetting import GridSite, Labware, aspirate, dispense, write_gwl

pipetting_site = GridSite(36, 1, "MP 3Pos")
labware = Labware("96 Well Greiner", 8, 12, spacing=1)

ALL = 8 * [True]

worklist = [
    aspirate(pipetting_site, 1, ALL, 50, "LB CD ZMAX", labware=labware),
    dispense(pipetting_site, 2, ALL, 50, "LB CD ZMAX", labware=labware),
]

write_gwl("worklist.gwl", worklist)

Place plate in the incubator

from pypetting import GridSite, Labware, transfer_labware, return_plate, write_gwl

src = GridSite(36, 0, "MP 3Pos")
dest = GridSite(68, 0, "StoreX 22Pos")
labware = Labware("96 Well Greiner", 8, 12, spacing=1)

worklist = [
    transfer_labware(src, dest, labware),
    return_plate(1, 1),
]

write_gwl("worklist.gwl", worklist)

Writing very complex workflows

The EVOware software is not designed to write and maintain very complex workflows. We therefor suggest writing the workflows using a minimal wrapper within EVOware that dynamically generates worklists with pypetting and executes them. This can be done as follows:

1. Call an application to generate the workflow, store the return value of the application in a variable
2. Execute any workflows generated by the script
3. Jump to step 1 if the return value is non-zero

This workflow requires the Python script to keep track of the current state, which can be done by storing the current state in a file, or by using a database. We recommend designing the overall workflow as a finite state machine, where each state transition can be executed as a separate worklist. This allows testing each state transition in isolation, which will result in a more robust and maintainable workflow.