Opentrons User Interview with Adam Ericsen of Emory University

Adam Ericsen researches HIV host-virus interactions using fully automated genotyping, microbial sequencing and immunological assays. Here's how.

Adam Ericsen is an Assistant Professor and the Assistant Director of the Genomics Core lab at Emory University’s Yerkes National Primate Research Center. Ericsen’s research program focuses on understanding modifiers of host immune responses, a system of complicated, complex steps he has innovatively automated—thanks to the three Opentrons pipetting robots that he began integrating into his lab’s toolkit in April 2018.

We talked with Adam to find out how those robots help automate his entire lab.

Adam Ericsen (front) in his lab with his 3 OT-2 and lab members Nathan Gullicksrud, Steven Plonk, and Yuhui Wu (l-r). CREDIT: Adam Ericsen
Adam Ericsen (front) in his lab with his 3 OT-2 and lab members Nathan Gullicksrud, Steven Plonk, and Yuhui Wu (l-r). CREDIT: Adam Ericsen

Opentrons: Please tell us a little about your background.

Adam Ericsen: I earned my Ph.D. in pathology from the University of Wisconsin at Madison. By training I’m an immunogeneticist and infectious disease researcher, and my work focuses heavily on understanding host-virus interactions in HIV infection. I primarily use macaque monkeys as models for this research.

Opentrons: What kind of research do you do?

Adam: For the past decade, I’ve been looking at how genetic and environmental factors modify disease progression in virus infected animals.

Opentrons: What’s a typical day in the lab for you?

Adam: Our lab runs a barrage of genotyping, host and microbial sequencing, and immunological assays. At a bird’s-eye level, we work to understand how disease progression is modified, both genetically and non genetically, in hopes of identifying actionable biological paths toward slowing, or altogether stopping, the progression of AIDS in HIV-infected people.

Our lab is very new, and it’s not easy to train technical staff into highly diverse assays. I needed our lab to excel at whole blood processing, buffy coat and plasma fractionation, DNA isolation (from any and all sample types), quantitative molecular assays, and flow cytometry. Ambitious? Maybe... but to do meaningful research I needed my small lab to function like a big one.

A common protocol for us is to spin blood and separate it out into several components. It wasn’t easy to get to where we are now, but getting a fairly routine blood processing workflow up and running was imperative. Training a new lab—my new team—to do assays "just like this" was terribly time consuming and stressful. I constantly reminded myself that time spent training my team was time well spent, but my resolve was short-lived.

And that’s where the Opentrons robot fit in: I needed my training, my experience, and my hands to be in the lab, even when I couldn’t be.

Opentrons: Why did you choose the OT-2?

Training a new lab—my new team—to do assays "just like this" was terribly time consuming and stressful... and that’s where the Opentrons robot fit in: I needed my training, my experience, and my hands to be in the lab, even when I couldn’t be.

Adam: I had heard of Opentrons, and around the time I started the lab, the OT-2 was coming out. I was attracted by the open-source aspect of the platform: knowing that I could add, modify, and enhance core functionality at any time allows my small lab to buy once and, with the magic of code, add significant additional value. As one of my techs framed it, "We write the code, and we watch it move."

We ended up combining SQL, C++, PHP, JS, HTML5, and Python to build out a custom web-based front-end for managing all interactions with our Opentrons protocols (we couldn’t have done it without the expansive documentation for the API). I can be sitting in an airport in Portland while controlling one of our OT-2s in Atlanta; gimmicky for sure, but it’s just an example of what we have been able to accomplish.

Perhaps more importantly, the robot interfaces with our sample and reagent inventory and harnesses the Google Speech API. Since we dove into the deep-end, we’ve been having a blast "automating everything," as my folks put it. Together we wrote a ton of modules in Python, and each completed protocol run phones home to our lab server to ensure that all dependencies (like labware definitions, deck positions, volumes, etc.) are up to date.

We’ve embraced the OT-2 platform to do every assay that I would have otherwise trained a new technician to do by hand. In fact, we don’t do anything by hand anymore. I joke that once we get the robots to write grants and manuscripts, we’ll do all our work from the park.

We’ve embraced the OT-2 platform to do every assay that I would have otherwise trained a new technician to do by hand.

Opentrons: What specifically do you use the OT-2 to do, and how does it fit into your workflows?

Adam: We have entirely automated PCR, Fluidigm, flow cytometry, blood processing (post-centrifugation), and nucleic acid isolation assays. When the need arises we design custom labware, which my guru wife 3D prints. All our labware is microchipped, and anything loaded onto the robot is scanned in by barcode or RFID: the robots check with the server to make sure the right samples are where they are supposed to be.

We can even fully process whole blood that’s been spun down first. Nathan Gullicksrud, in my group, came up with a way to hold blood vials at certain heights so that we can easily extract plasma and buffy coat with ease. That means we can get the layers to fractionate, isolate the plasma, clean up the white blood cells, and that’s it—they’re prepared. To be clear, we’re talking about as many as 96 simultaneous vacutainers being processed in a few hours with limited user intervention. We rapidly scaled to a throughput that would otherwise have been entirely unrealistic.

With the OT-2, we save a ton of time. Period.

We rapidly scaled to a throughput that would otherwise have been entirely unrealistic. With the OT-2, we save a ton of time. Period.

Opentrons: Please tell us more about how you customized the OT-2 to fit that workflow.

Adam: I coded an entire interface to connect with the OT-2. We also did several hardware modifications—including drilling holes, and putting microchip scanners and photosensors in every slot position. The robots speak: "Attention. Attention. The run has been paused. For your safety, please close the front window," or "Transferring DNA from column X in the DNA input plate to…" As a result, we can do just about anything—from whole blood with all outputs it produces, to fully automated DNA protocol.

My goal is to free my folks up to do more fulfilling activities. Sitting at a Biosafety Cabinet for several hours, every day... technical expertise in a research laboratory can be so much more than that. The Opentrons platform lets me be ambitious, even overzealous, with the scope and scale of research endeavors that I am willing to entertain as a new PI.

3D printed RFID sensors placed under the OT-2 deck
3D printed RFID sensors placed under the OT-2 deck. CREDIT: Adam Ericsen

Opentrons: What was it like to get your Opentrons robot up and running?

Adam: Opentrons is the first free roaming liquid handling robot that I’ve used. I wrote the core Ericsen Lab code on top of the Opentrons API, and as we encountered bugs or issues, we consulted with the Opentrons support team. We pushed through issues, revised or re-imagined our approach as needed, and worked closely with the support team. We have yet to fail outright.

I have a great team. As a new investigator, I feel incredibly fortunate. That said, no one in my group came with programming expertise. I basically told them that I envision building an automated, smart, and aggressive research program. I put them in front of the robot and said, "we’re going to figure this out together." Instead of training someone into the manual aspects of conventional assays, we wrote code to do it on the OT-2. The process of training someone into the lab became the process of working together to train a robot to successfully perform basic assays and complex experiments. That, in turn, gave my team the expertise in the assays that we need to be great at. My goal was to build a team of enthusiastic and hard-working creatives that could successfully perform an otherwise intimidating and complicated assay or experiment without having ever attempted it previously. And with Opentrons I believe this approach is going to save me time because they’ll learn from the robot, and the connectedness of the lab will safeguard research data, maintain records, and ensure that we can stand behind the work that we produce.

If a robot isn’t open-source in my lab, then it’s just a replacement for a pair of hands—and I can get hands cheaper than that.

Opentrons: What did you expect the process of lab automation to be like?

Adam: I never entertained failure. Buying a robot for a research laboratory isn’t like buying a popular consumer-focused device. Initially, I encountered a lot of frustrated investigators (i.e. through their YouTube rants) that had no idea what it would take to get an Opentrons robot up and running. There is certainly a learning curve, but we are also talking about controlling a production laboratory robot using Python. In case it isn’t clear, this means a young lab with limited resources can outperform large research programs at a fraction of the cost. My robots sit next to my lab’s flow cytometer, which cost 10x what an OT-2 costs.

Opentrons: What else is good or bad about your OT-2?

Adam: It’s great! The OT-2 can do simple things, but it can also do things that no other platform can do.

Opentrons: Had you used any lab automation before your OT-2?

Adam: No, actually. But as soon as I heard you guys ran on a Python backend, I thought, "I could do that!" My wife and I were doing 3D printing, and we were surprised at how similar the learning curve was.

Opentrons: Are you planning any other automation?

Adam: Well, we just ordered two more OT-2s, so definitely yes! The OT-2 supports both the speed and integrity of what we’re doing. Having it be open-source is enabling, because limitations that exist with proprietary platforms are deal-breakers. If a robot isn’t open-source in my lab, then it’s just a replacement for a pair of hands—and I can get hands cheaper than that.

The OT-2 has been a part of every day in the lab since we received it 18 months ago. It was a significant early investment—which I could have invested in training staff, but felt the robot and its platform were a more valuable use of our resources. We do highly advanced assays; with a robot, we just have to swap some pipettes.

Opentrons: Is there anything else you’d like to say about using the OT-2?

Adam: One of the funny things we did that has been incredibly fulfilling is we’ve incorporated the Google speech API to have the robot announce things such as when it runs out of a liquid, or when it needs tips, and what test it is starting. It speaks in this friendly female voice. Sometimes it randomly changes to male and startles us. But I love that we could even do that!

The OT-2 is not a robot taking my job. Rather, it’s transforming my job into something much more fulfilling. We also invested in a benchtop flow cytometer, next to our OT-2. New labs generally don't invest this much money into new equipment, but I was able to gain all this functionality with the robot, so it wasn’t expensive. It can take months to train someone to work with a cytometer, but with the OT-2, we can easily train people to do blood sampling in ways they haven’t done before. Now we're in a position where it’s basically point-and-click. The OT-2 has quickly become an invaluable, integral member of our lab team.

The OT-2 has quickly become an invaluable, integral member of our lab team.

This interview has been edited for length and clarity.