Opentrons User Interview With Northwestern University

Michael Rourke, a doctoral chemistry researcher at Northwestern University, is advancing organic chemistry research using the OT-2 and the Opentrons Protocol Designer.

Inspired by the potential to increase accuracy and precision while freeing his team from mundane repetitive tasks, Michael Rourke led his Northwestern University organic chemistry lab to transition to lab automation—and improve lab safety—using the Opentrons OT-2 liquid handling robot and open-source platforms through the user-friendly Opentrons Protocol Designer.

Michael Rourke
Michael Rourke. CREDIT: Michael Rourke

Opentrons: Tell us a little about your background.

Michael Rourke: I'm a synthetic chemist through and through. I have a degree in biochemistry from the University of Colorado, where I researched synthetic chemistry, mostly modification of carbohydrates. Now I'm at Northwestern as a PhD candidate in the Scheidt Research Group doing synthetic methodology. When I got here, I was working in an MBRAUN glove box with high-throughput photoredox catalysis in 96-well plate format.

Opentrons: That sounds like it’s a lot of manual transfers.

Michael Rourke: It is. Say you take two substrates, A and B, and combine them to make AB—with several variables you want to change (you might want to do six photocatalysts, four solvents, four bases, and so on, for example). It starts to add up pretty quickly. And the more sophisticated it is, or the more variables you consider, the smaller the boxes get on a plate. And when your focus gets smaller, it’s not as easy to say, “I'm in this region of the plate doing this right now; where did I start and stop?”

Opentrons: So how were you managing that complexity before you got the OT-2?

Michael Rourke: We were doing it with a manual pipette, and using cards between the wells to track our progress. It’s, frankly, subject to human error. Additionally it can be quite uncomfortable to hold a pipette in a glove box for hours. I had a review with my advisor and told him the high-throughput technology was working, but our accuracy and precision had potential for improvement.

Opentrons: Is that when you first considered automation?

Michael Rourke: Yes. My advisor agreed I should shop around to see what was available, and my boss was excited. But when you bring back a quote for a quarter-million dollars, you're not going to get far in an academic-granted center! Our biggest fear, though, was the durability of the pipettes. We’re using the Opentrons Electronic Pipette, which is made of carbonate based polymers. Their potential to degrade in organic solvent was the biggest foreseeable challenge or difficulty for our workflow. We saw an article about a Michigan professor who had an OT-2 robot in a glove box, and we figured out the pipettes we would need for the OT-2 were similar in pricing to the ones we were already using. In January 2021, we pulled the trigger, ordered an OT-2, and put it in an inert atmosphere glove box.

Opentrons: What was your experience setting up your OT-2?

Michael Rourke: I had measured it to be sure it would fit, but that was the moment of truth—and it did fit! I have the USB feedthrough because I didn’t know how well the WiFi would work in the encapsulated space, but I could have saved some money and not done that because once I got it in there and sealed it the WiFi worked pretty well.

The Scheidt Research Group's OT-2 in a glove box. CREDIT: Michael Rourke

Opentrons: How is the OT-2 working now?

Michael Rourke: It’s holding up. I did degrade one pipette slightly the first time I used it because I didn't have a necessary air gap at the top. But I’m working at about two-thirds capacity for organic solvents, so 150-200 microliters in the pipette. I put our own plates—Paradox 96-well reaction plates designed for photoredox catalysis—onto the robot and used the Opentrons Labware Library to work them into the protocol. Adjacent to the robot I have a blower to blow down the plates. And I use some reservoirs. So, I try to maximize the number of vertical columns  specifically with reaction solvents, so that I can go to a reservoir. I’m doing many of  my reagents in 2-dram scintillation vials. We’re all legitimate organic chemists here, with minimal experience in Python, so we’re all using the drag-and-drop Protocol Designer interface to make our protocols.

Opentrons: How are the OT-2 pipettes working for you?

Michael Rourke: We don’t have any problems with drawing and injecting. A lot of the chemistry I'm doing right now is with organic solvents such as acetonitrile, tetrahydrofuran, dimethyl formamide (DMF)—as well as dichloromethane, which is notably volatile. We coupled this with liquid chromatography mass spectrometry (LCMS), as well as gas chromatography mass spectrometry (GCMS), so I can put a standard in it before I run the reaction, or I can take it back in and have the robot take from the scintillation vial and add a known amount to each little well.

Opentrons: What is one of the most noticeable benefits of automating pipetting for you?

Michael Rourke: It's important to have consistency. Everybody who's ever set up plates by hand has had that moment where they said, “Did I just add to that one?” Now, everything’s completely reproducible. That’s the biggest benefit.

Opentrons: How would you improve the OT-2?

Michael Rourke: Our pitch to our boss was that automating pipetting was about throughput, but it’s also about accuracy and precision. There's nothing worse than spending a day or so doing a reaction and then leaving something out. The Opentrons Protocol Designer codes entries by color, but color printer cartridges are not always available in academic labs. So we print off a screenshot of the starting deck condition and we write on it using ChemDraw to ensure accuracy. It would be nice if you could add a number or letter code as a description in Protocol Designer to track a solvent or reagent in a vessel but— that’s a small thing.

Opentrons: How does the workflow process differ from how you used to do it by hand?

Michael Rourke: Every workflow is custom almost every time, so everyone in the lab designs our own stuff on the fly. You have to be creative in how you lay it out. I’ve never set one up without a microchannel, but some of my coworkers still only use a single-channel pipette; and although those are more accurate and precise than manual pipetting, it’s similar to the pace and procedure we use by hand. I did 300 reactions in two and a half hours with a combination of single and microchannel operations.

Opentrons: How does letting a robot handle all that pipetting in the glove box improve your throughput?

Michael Rourke: It’s definitely increasing our throughput. We used to do 96 in a day. Now, I can do 300 in a day. Certain things are never going to change about what we do in the glove box, but prepping everything on a large scale, all at once, you can take a day and set it all up and then come in the next day and run them. I used to only be able to do one plate a day, and now I’ve run as many as four, and if we had all the reagents and everything, we could probably get up to eight in a day, or let it run overnight. And rather than having to stand at the glove box for five days in a row and come in over a week or two for the analysis, now I can screen as the initial step, and then set aside time to analyze all the data points.

Opentrons: Anything else?

Michael Rourke: I became a chemist to analyze data and come up with new ideas, and to share ideas with colleagues and younger students. When I was in the glove box, thinking in words was too much. You could only chant numbers as you counted under your breath, and no one would talk to you because you would lose count. It's just counterproductive to creativity and happiness. But the mundane task of filling wells is something that we are not required to do anymore. No one has done a plate by hand since we got the robot. I’m excited that this summer, undergraduates are going to be trained on, and use, the OT-2.

I think that in time things will change because a lot of things in labs could be automated. For example, we're very proactive about safety here at Northwestern: organic solvents are flammable, and using them is a real risk. The ability to transition reaction setup to an automated system in a controlled environment such as an inert glove box has obvious advantages. Additionally, we have labware enabling the robot to do preparatory scale reactions in 20 mL and 40 mL vial format. As organic chemists, we have a huge push to ensure safety, and in the near future I foresee laboratory automation making a significant contribution to risk mitigation in both industry and academic labs.