Why RobotEvo?

(adapted from my PhD Thesis: RNA virus detection and identification using techniques based on DNA hybridization)

  • Programming automation of RNA extraction:

Usually, prior to proceed to the application of the DNA-hybridization-based technique, like RT-qPCR, the viral RNA need to be extracted. We used well established methods and commercially available kits based on columns (RNeasy Mini Kit or QIAamp Viral RNA Mini Kit, QIAGEN, Hilden, Germany) or magnetized particles (NucleoMag® VET kit) from MACHEREY-NAGEL, Durel, Germany) to achieve the separation, either automatically, using pipetting robots, or manually.

Especially useful was the combination of magnetized particles with a Freedom EVO universal pipetting robot from TECAN, Mannerdorf, Switzerland.

Using the provided software (Freedom EVOware) it was comfortable to write simple and specific pipetting protocols in a semi visual way. Unfortunately, writing more complex or flexible protocols (for example to accommodate arbitrary number of samples or minor modifications of the protocols) was very time consuming and error prone. You are compeled to use variables and program-control-flow structures like IF and LOOP. But you will find that there is a poor or no support of variables of different types, arrays, structural-programing and objects within the provided scripting language.

More important, the powerfull validation and visualization tools provided by the script editor are full sopprted only in relativelly lineal and simple scripts, considering only the “default” values of the variables, and thus, the default flow-path of the program, not detecting problems in the alternative paths, likely to be found in most runs.

To overcome these limitations and afford automation, a new software was written in Python, “RobotEvo”, which generates the scripts for the robot. This new Python library provide new layers of abstraction to offer a higher level programing model to allow a more direct programing of the steps needed in a typical biochemical/biological pipetting protocol like RNA extraction. The layers of the implementation are: a parser and a generator (module :file: `instructions.py`_ of the “low-level” instruction set directly usable by the provided Freedom EVOware software; a set of “modes” to provide the desired kind of output (human readable comments, separated instructions, EVOware scripts, etc., in module :file: `evo_mode.py`_; a model of the state of the robot to detect possible errors prior to the generation of the script by tracking what volume of what mix of reagents contains at each moment each reservoir or tip (module [Robot](https://github.com/qPCR4vir/robotevo/blob/master/EvoScriPy/Robot.py) – this is a _novel_ functionality impossible to achieve with the original software); low level pipetting instructions (like aspirate a specific liquid volume from a given vial into a tip); a higher level command set (like distribute some reagent into each sample, in module [protocol steps](https://github.com/qPCR4vir/robotevo/blob/master/EvoScriPy/protocol_steps.py)) to directly program high-level, more realistic protocol scripts including a base template for a full protocol; and, finally, a set of facilities to declare the reagents (module Reagent) and the labwares (like reaction tubes, racks of tubes, racks of tips, cuvettes, etc. in module [Labware](https://github.com/qPCR4vir/robotevo/blob/master/EvoScriPy/Labware.py)).

For the protocol for RNA extraction (module RNAextractionMN_Mag_Vet) the set of used labwares and reagents are declared first. Immediately an automatically generated check-list is presented to the human operator (a graphic user interface – from module GUI). After a possible adjustment of the predefined parameters (without any programming) the program go through a few high-level-defined protocol steps of distributeing buffers, mixing, washing, incubating, etc. generating a very detailed set of low-level instructions for the physical robot in a script to be imported and visualized in the TECAN Freedom EVOware software. The obtained script is very long but structurally very simple and well commented, which facilitates the visual control of each instruction prior to real pipetting.

  • modules.jpg