What is Lab Automation Engineering and Who's a Lab Automation Engineer?
Intro
I was inspired by an article written by fellow lab automators over at the Automated Lab to write this post. Lab Automation is a niche form of automation engineering used to enhance biological workflows using hardware and software technologies. In this post, I’ll talk about what Lab Automation Engineering is, discuss how lab automation is used in different facets of research and drug development, and my own journey in the space as well as what I specialize in.
What is Lab Automation Engineering?
To start, let’s focus on the latter part of this title: automation engineering. If you take a look at the timeline created by my colleagues at the Automated Lab, you’ll see that automation engineering started on the assembly line. According to Madeline Wolf, who I had the pleasure of meeting at SLAS this year, the first industrial robot was invented in 1937. The Robot Gargantua was able to stack wooden blocks. Not to be confused with a manned crane, this robot was programmable. This set the stage for robots in manufacturing, like the Unimate, used on the Ford manufacturing lines in the 1960s.
These made the scale up of manufacturing processes possible, not just because of their ability to work faster and longer than humans, but they were also able to handle hostile environments, like intense heat, which was necessary for certain steps of the car-making process. The people who maintain and run these robots are known as automation engineers. Assembly lines have become what the general public often associates automation with the most, but you only need to look at that phone in your hand to see other forms of automation engineering at work.
Automation engineering is also heavily associated with computer and data science. Making tools that automate budgeting, clearing your inbox, scheduling, spam call filtering,or metro-card autoload are some simple examples of automation in everyday life. Software to support the industrial robots I mentioned earlier allows remote programming and monitoring of manufacturing processes. AI is also automation, as you can teach a machine learning model to automate the analysis of large datasets like those for medical or financial institutions. This can greatly reduce timelines because of faster data-based decision making.
In a lab setting, every workflow has different needs. You still have automation at work when manufacturing drugs, for instance. This automation may be more streamlined and regulated as opposed to automation in early stage research and development. In this environment, dynamism is required to adapt to all of the experiments. Madeline goes on to compare clinical vs. non-clinical and academic vs industry labs, so please check out the Automated Lab.
Automation in the Lab
Automation in a lab refers to the software and hardware components that work together to make an automated workflow possible, Here are some aspects of an automated lab:
Automated liquid handlers like the Tecan Fluent and Hamilton STAR
Small volume liquid dispensers like the Tecan D300e and Dispendix I.Dot
Plate Readers and Confocal Microscopes like the Tecan Spark and Revvity Operetta
Automated incubators like the Liconic incubator
Barcoded tubes and plates and barcode readers
LIMS software
Data management and analysis software
ELNs (electronic lab notebooks)
Scheduling software
This list doesn’t include specialized tools that can be used in an automated workflow for common lab protocols:
Robocolumns enable protein purification on the deck of your liquid handlers. Specilaized tools like the Te-Chrom scales up protein purification to a 96-well format, When paired with software from Synthace, programming high throughput protein purification has never been easier.
The Sartorius Ambr 250 automates fed batch experiments without having to scale up to a large bioreactor. As someone who went into the lab at 3am to feed cells, this is something I wish I had back then.
BlueCatBio BlueWasher performs centrifugal plate washing, greatly reducing dead volume, which improves assay CV. When paired with a stacker, this useful tool can help with those finicky wash steps. Even works with cell-based assays!
MSD Electrochemiluminescent technology can scale up Western Blot analysis, which isn’t automation friendly when done traditionally.
High Content Imaging and analysis can give you mounds of data using a fraction of the consumables and analysis time.
Bioreactors: For manufacturing biologics, bioreactors that can handle thousands of liters of solution are necessary for the scale of commercialization. Automation ensures that the proper environmental conditions and feeds are administered as well as minimizing contact with the outside environment to mitigate contamination.
When thinking of automated solutions, it’s not always a 1:1 translation from the manual protocol. There are countless solutions available, you just need to know who to ask. Another question that needs to be answered is: “What kind of workflow is best for your lab?” There isn’t a standardized model for what lab automation looks like, it’s not a monolith. I’m not going to recommend a fully automated workflow for a company that works with patient samples, for instance, because of the amount of dead volume needed for a fully automated workcell. An automation engineer knows how to apply technological principles to biological workflows to create a myriad of solutions. Referring back to the title of this post, “What is a lab automation Engineer?,” I hope I can tell you a little bit about my journey and why that title fits me.
RA to Lab Automation Engineer
I started working in the industry at a bioprocessing CRO. It was a lot of hard work, but I got a taste of the different facets of how a lab worked. My main function was lab work, but I also managed the lab, ordered consumables, negotiated purchasing contracts, and did some light accounting work. I ran feb batch experiments to optimize antibody production in cell lines. Now, if you’re familiar with process development, you know that some feb batch experiments need to be fed every 4-6 hours. This means I was in the lab at all hours of the day sometimes. We had bioreactors, which automates the fed batch process, but only for one fed batch at a time and volumes starting at 3L. I remember going to a bioprocessing conference and seeing a demo of the AMBR. It was the first time I was exposed to a technology like that. If you don’t know, the AMBR miniaturizes the bioreactor, so you can try many different conditions for your fed batch experiments. By automating the feeding schedule, scientists could sleep at night while their cells were taken care of.
From there, I made a pivot to a virology company screening small molecules. Most of the screening was done by hand: we reconstituted the lyophilized powder, made dilution plates with a multichannel, and set up our cell based assay with the same multichannel. The automation came with the assay harvest: we used a Biomek NX from Beckman Coulter to transfer supernatant from 4 96-well plates into a 384 well QPCR plate. From there a macro written by a data scientist was used to parse the data into 4 96 well plate data files. We also used a data management system to make the data easier to find by tying our assay data to a molecule entity. Our work resulted in an IND filing, and had two separate assays run by two operators each to provide a robust data package for that filing. A highlight was using the Biomek to run a 10000 compound single point screen. This made me realize that scientists should not go into a 384 well plate by hand, as I definitely missed some wells aspirating with my single channel. That’s what hit picking is for!
I had a brief stint at a non-profit research company, where I was able to create a high throughput array ELISA. This involved using the Scienion SCIflexxarayer, a small-volume liquid dispenser, to print proteins into 384 well plates and other liquid handlers and readers to execute the ELISA testing for pan-Lyassavirus antibodies. I also used tools that were hacked on site, like a chromatography system that enabled purification of multiple proteins at a time. This was also my first exposure to the Agilent Bravo, where I explored resin containing tips for automated protein purification.
From there, I went back into the virology space while also getting exposure to functional genomics. I screened host directed therapies(HDT) to find a pan-respiratory molecule. A host directed molecule inhibits a cellular function related to the lifecycle of the virus. Remember that viruses hijack the cell's processes in order to replicate. The theory behind HDT is that one cellular function is involved with the life cycle of multiple viruses or “one drug for multiple bugs.” In order to find the drug target, CRISPR knockout screens are performed to knockout specific genes from the cell. Then the cell was infected with a virus: if a cell didn’t get infected compared to the control cell, then it’s likely that that gene is involved in the viral lifecycle. If the absence of a certain gene has “anti-viral” properties for multiple viruses, then that gene has a better chance of being involved in the lifecycle for multiple viruses.
I came in after the CRISPR screening to screen molecules that targeted specific cellular functions. We had a room full of liquid handlers, and I was given free reign to use them and implement automated assays. The starting molecule was usually a Cancer drug, which shut down certain genes in cancer cells. The screening campaign required screening the same molecules in multiple assays, which is where the automation shined. Because we used the same equipment for our assays and similar protocols, assay performance was consistent across the half dozen different assays and 3 operators screening the same compounds. This was proven through the pEC50 correlation between molecules across assays.
Our data pipeline was also more advanced than other teams in our lab. We used a data science tool to consolidate assay metadata and raw data into structured data that could be uploaded into a data management system. There was some copy and pasting involved, but we were not running EC50 calculations ourselves. Once the curves were generated, we could easily search data across molecules and assays. This allowed the chemists to make decisions as soon as the data was uploaded.
Taking this a step further, I was a key player in an integration project. If you read my post last week, I talked about LIMS systems and how they are necessary to structure data. In collaboration with IT and vendors, we created a workflow that integrated a sample management, LIMS, and data management system with our liquid handlers and plate readers. This took out the need for any copy, pasting, or data export/import. The result was fully automated data analysis and sample tracking. The only thing that was missing was barcoding for our assay plates, so I worked with IT to make a barcoding system to enhance the sample tracking part of the workflow.
As a freelancer, I’ve kept the concepts alive and am learning more every day. I’ve networked like crazy and put myself out there. Writing this blog has given me an outlet to show that I belong in this space and that I know what I’m talking about. Despite never having the title, I know I’m an automation engineer because of the work I’ve done. I’ve been designing experiments and thinking about how automation can improve those workflows throughout my career. Lab Automation Engineering is such a new space that many companies either don’t know how to level one correctly, or they do know and don’t level them correctly to save money. Take a look at this career ladder and look at the work you do. Are you a research associate, or are you actually an automation engineer?
Conclusion
By exploring the definition of lab automation engineering, its applications in research and drug development, and my personal experiences in the field, I hope I provided a comprehensive overview of this specialized area. Whether you're curious about the technology, the career path, or the impact of automation on scientific progress, I hope this post offers valuable insights and a deeper understanding of what it means to be a lab automation engineer.
Special shoutout to the team at The Automated Lab, they’re putting together a great resource for folks that want to learn more about lab automation engineering. Definitely check them out.