Proteins, particularly antibodies, have tremendous research and clinical value, but growing and purifying them can be challenging. To isolate proteins of interest and produce sufficient yields for ongoing study, labs must carefully remove genetic material, extraneous proteins, and other molecules. The process is not unlike panning for gold.
Once a protein is isolated, researchers can characterize its structure, function, interactions, and other traits. However, just getting to this point can be time consuming, expensive, and create a significant bottleneck in the drug discovery process.
Automating a complex, error-prone process
Historically, protein purification has been a manual process. Lab workers, often PhD scientists, hand-pipette samples, operate vacuum manifolds, and perform other rote tasks, including centrifugation and multiple filtration steps.
This approach is error-prone and time consuming. Even more concerning, it keeps highly skilled scientists chained to an assembly line rather than designing new antibodies and performing other critical functions. On the other hand, protein purification is challenging to automate. As an example, an immuno-oncology (I-O) company’s protein purification workflow starts with 24-well plates, containing unpurified samples, which are barcoded, taken through a centrifuge, and filtered. Magnetic beads are then added and the sample is incubated for two hours. The beads are then washed and the samples filtered into 24-well collection plates. Following additional filtration and shaker incubation, the samples are placed in 96-well storage plates. Ultimately, a portion from each well is transferred to read plates, and total protein amounts are recorded. Looking to accelerate this process, the company looked to implement automation software; however, there were several challenges. The process uses a liquid handler several times, increasing the risk plates could be deadlocked.
Deadlocks happen when there are more plates than spaces for them. For example, one plate needs to go from the incubator to the liquid handler, and another needs to go from the liquid handler back to the incubator and the two block each other, stopping the system.
To maintain a linear progression and ensure critical liquid handling steps are only executed when plates are ready, a global script – a list of commands – was developed that divides the workflow into three distinct parts:
1. Sample plate introduction to shaker incubation
2. Incubation to liquid handling for first reformatting from 24-well to 96-well filter plates. from 96-well filter plate to storage read plates
3. First reformatting to final steps
By creating a single sequence that spans the first two liquid handling steps, the system restricts how many plates can enter the system, so the samples don't over-incubate or under-incubate. The global script evaluates multiple variables simultaneously, such as the number of initial production plates, plates in the system, and available shakers.
By breaking the process down so seamlessly, the script can temporarily disable any of the three workflow segments as needed, preventing deadlocks and prioritizing plates that are already in the system. The scheduling software enforces a linear progression for liquid handling procedures, with each sequence executed in order and completed before another sequence can begin. This helps simplify run design, alleviating the need for additional scripting.
Automation software adjusts multiple variables on the fly, resetting based on plate movement. Software tracks plates as they move through different parts of the system, delays them as necessary to ensure effective processing, and releases them to continue processing.
Fruits of automation
Software can transform a cumbersome, labor-intensive protein purification process into an automated one, dramatically increasing walkaway time for protein researchers. This new workflow produces fewer errors and doubles throughput. Manually processing four 24-well production plates took approximately nine hours. The automated work cell completes the same run in four hours and 15 minutes. In addition, the system can keep working on its own.
The system can do more because it’s not restricted to an eight-hour day. It reduces human error because instead of a person hand-pipetting samples from plate to plate, everything is handled by robots, and scientists don't have to worry about losing sample integrity. All samples are handled consistently in terms of timing, temperature, and other factors.
The system also manages consumables more efficiently. Tips are automatically provided, as needed, and sent to waste, as required. The system can calculate how many tips should be used for each plate. Rather than automatically replacing a tip box that’s partially full, the software tracks progress and uses all tips, saving money and the environment.
At the end of the run, the system can produce the critical data the team will need to move forward, showing them how much purified protein they have generated. With the system now fully automated, researchers can focus more on antibody design and other vital tasks and let the system purify antibodies on its own.