The Worklist - Providing Lab Automation Instruments Instructions to Carry Out Experiments

There are many ways to create a protocol for controlling lab automation instruments, but the strategy of the worklist is currently how labs are formatting and launching their experiments. The concept of the worklist is quite simple, but can appear complex. I've seen many different worklist formats, from very simple to very complex. Here are some thoughts behind automation worklist development and some best practices to consider.

Experiment Design

Laboratory information management system (LIMS) is in all labs these days, unless the lab happens to be still keeping data on pen and paper. This is a general term for the databases to hold all the lab experiment data, and typically this is where your experiments will be launched. Users will go here to design experiments, and ideally they will return there to view the data that their experiment generated.

Manually running experiments still requires some sort of experiment setup, typically described and recorded in an electronic lab notebook (ELN). This is often integrated directly with LIMS these days, so that scientists can provide rationale behind their experiments and note anything interesting that happens. Let's say that a chemist wants to set up a chemistry reaction experiment in which two liquids will be mixed at given concentrations. They would traditionally use the lab notebook. Here is an example of what an experiment setup could look like there. Don't worry about the chemistry aspect, it's just an example to show how a scientist would go about calculating the amounts of the two reagents to be combined in this experiment.

Worklists

A worklist format is itself a structured set of experiment instructions, and that structure is set by the requirements of the experiment being run. The data structure of the worklist itself and file type depends largely on the software that will be utilizing the instruction set, and can vary from .csv (my favorite) to .xml, .txt, .json, etc. Note that in this .csv example provided, the columns represent experiment parameters and the rows represent lines of instruction within that parameter set. Imagine that our scientist is formatting a worklist by hand based on the experiment that they designed in their lab notebook. In this experiment, they want to combine different compounds into specific reaction wells of an experiment plate at very specific concentrations. Here is an example of what both the experiment design and the resulting worklist would look like.

You can see that the worklists is structured feed necessary parameters into an experiment method, which is designed to read the instructions and process experiments accordingly. Here is some more information about how the automation instruments are programmed to respond to these parameters that the worklist feeds in. The scientist in this case had to do some calculations based on desired concentrations to get these transfer volumes. The great thing about using computers to generate worklists is that they can readily handle these calculations for you.

The real power in this strategy comes with utilizing software solutions to receive the required experiment inputs and generate instructions based on that expected content provided by scientists. so the ideal lab IT setup (LIMS) will allow scientists access to their electronic lab notebook through a web portal for experiment design, and then LIMS is able is able to generate a worklist file based on the experiment provided. Some labs (like ours) go further to develop software interfaces with LIMS to allow that worklist file to be sent over to an instrument queue for processing, and then the user receives email notification when the automated system has completed their experiment. Here is an example of how a software system could translate experiment inputs into an appropriate worklist to drive the resulting experiment.

Dynamic Instrument Control

Worklists can also be set up to allow enormous flexibility of device and labware usage. For example, we often use worklist values to drive the incubation of the reaction plate at a given temperature. Required inputs to this process are incubation time and incubation temperature. The method is programmed to respond to all possible scenarios, and robotic devices are scheduled to act accordingly.

So when a worklist executes that specifies incubation time greater than zero, the system knows that a incubation is desired at the provide temperature. In response, we send commands to warm the incubation chamber to that temperature and then engage robotic arms to facilitate the plate transport for that time period. However if the time period specified is zero, then the incubation chamber and robotic arms are not called to action, left idle to handle other tasks.

In a similar way, worklists can be used to dynamically select the labware. Our compound management system supports plating of the library into any plate format specified by the worklist, and into one of 4 plate types that we handle. The scientist simply defines the plate map and the plate type when they create their experiment, and the resulting worklist is addressed by our automation instruments to fulfill the custom order.

Experiment Reports

Another type of structured file that needs to be considered in this process is the experiment output data, which is typically structured as an Experiment Report (output report) in the way that LIMS can expect and interpret. This allows scientists to update their experiment with the experiment status and result data. In our case, we write reports to capture any errors that the machines may have encountered when carrying out experiments. In this way, we know to take a closer look at the data rather than taking accuracy for granted.

This type of connectivity between automation instruments and scientists is totally software driven, and the field of research IT is exploding. These systems require tremendous forethought and painstaking development, but the efficiency of helping scientists discover new life saving drugs more quickly is priceless.