AN ENGINEER IN BIOLOGYLAND
“Biology? That’s unusual.” This is something I’ve been hearing a lot lately, usually from my friends when I tell them about the courses I’m taking this term. I’m in the home stretch of my Engineering Physics degree, which is the point where I’m supposed to choose some technical electives and become a specialist. Most Engineering Physics students take traditional electives, like fluid flow, power transmission, or aerodynamics. But there is a growing group of students, like myself, who are looking to apply their engineering knowledge to less-traditional areas, and biology is perhaps one of the most popular.
But how exactly can an engineer do anything useful in biology? That was a question I tried to answer last summer when I went to go work at the BC Cancer Agency’s Genome Sciences Centre (GSC), in Vancouver. The GSC is unique in that it has an embedded engineering group, with two full-time engineers, who do on-site technology development for the life scientists at the GSC. Technology development is quite a broad term, but what it means in this context is designing and building devices to make life science experiments faster, better, and less expensive.
On my first day of work I was plunked into an environment full of life science and life scientists, neither of which I had much experience with. My challenge was to build useful new devices for this lab, and fortunately all of the life scientists were enthusiastically on board with this idea. So I plunged ahead into the work term, starting with meetings between the senior scientists and engineering group about what they wanted done. Immediately, I noticed some differences in the approaches the two groups took.
I found that the engineers often wanted to approach a problem from the ground up, and do things completely differently, which is understandable because we wanted to create novel devices, not just make slight improvements. Life scientists, on the other hand, were extremely concerned with keeping experimental conditions as constant as possible between the old method, and the engineers’ new method. This is also understandable because the GSC is (for the most part) a production sequencing and mapping facility – it is simply not an option to have production grind to a halt because the new method turns out to be inequivalent to the old method.
Sometimes there would be disagreement, but for the most part the engineers and life scientists were able to find a balance between massive overhaul and “not messing too much with the experiments, “ and you’d end up with a really useful device that does all the things the old method did, but faster, better, and for less money. I think that the project I worked on for most of the summer is a great example of how collaboration between engineers and life scientists can result in something that neither group would be able to accomplish alone.
My project involved agarose gel. If you’ve ever prepared agarose before, you know that it can be time consuming, boring, and messy. If you’ve never prepared agarose before, you’ll have to take my word that it can be time consuming, boring, and messy. For most people who use agarose, a few gels a day is the most they’ll probably ever need, and preparing it by hand (add agarose powder to buffer, microwave, stir, add water, microwave, stir, add water, microwave, stir, add water, cool to 60 degrees, pour) is an acceptable way of doing things. This is woefully inadequate at the GSC, where (in the past) two technicians had to spend their entire morning preparing and pouring the day’s 18 gels. When I arrived, a decision had just been made to expand to 54 gels per day. Well, hand preparation is a good technique on a small scale, but on a large, production scale this is a problem crying out for technology development. I undertook the project to build an automated gel-pouring device.
The biggest concern the life scientists had was ensuring that gel made with the new device was exactly equivalent to handmade gel; in other words, “we found a way to make this experiment* work with handmade gel, let’s not mess with it too much.” So I had to be certain that this device would spit out gel with the correct temperature, volume, percent agarose, and pour speed. Of course, if this device was going to become part of the gel production “pipeline,” I would have to ensure that it would be reliable as well, because a failure would shut the whole pipeline down.
The first part of the device was a commercial mixing unit, designed for preparing agarose. We decided to purchase it outright because a German company had already gone to the trouble of engineering it, so why duplicate their work? In this unit, you could simply dump a large amount of (measured) agarose powder and buffer in, and it would automatically heat, dissolve, and mix over 30 batches of gel at once, and hold it all at a particular temperature to boot! This was pretty good, but I hadn’t done any engineering yet, we had just spent a lot of money. The real trick was getting the gel from the mixing unit to somewhere useful.
For this I designed a custom pump system, which was essentially a large syringe. When the syringe aspirates, gel is drawn in from the mixing unit. When the syringe dispenses, gel gets pushed out of a dispensing hose. The syringe could be moved up and down by a computer-controlled motor, which had very fine speed and position control. With this setup, I was able to get highly repeatable volume and pour speed for the gel.
Of course, everyone knows (and if you don’t you’ll have to take my word again) that agarose solidifies when it cools. There were quite a few valves and hoses between the mixing unit (where the gel was kept heated) and the end of the dispensing hose. If you simply let agarose run through these elements, without thinking about its temperature, then pretty soon you’re going to end up with a system jammed full of solidified gel, which isn’t much use to anybody. To get around this, I used heated hoses, and cloth heating elements on the syringe and valves. A particularly nice feature of these heaters was their adjustable temperatures, so that the gel was kept at the same temperature no matter where it was in the system. Thus, no matter how long the system was left sitting, the temperature at the outlet hose was always the same ideal pour temperature.
The final part of the system was control, so I rigged up a simple, sealed control box next to the hose. After creating a batch of agarose in the mixing unit, the operation of the system was quite straightforward – press the dispense button, and 350mL of 60ºC agarose was dispensed from the hose in exactly 10 seconds. Press dispense again to repeat, up to about 35 times. The gel was exactly what you would get from handmade agarose, minus the repetitive weighing, mixing, microwaving, water replacement, and cooling. With only one technician, three times the number of gels could be poured in the same amount of time.
Actually implementing this device at the GSC was an adventure in itself. Life scientists, it seems, are accustomed to buying commercial instruments that have been extensively tested before ever being put on the market. As an engineer, however, I know that it’s very hard to get a prototype like this to work perfectly the first time; there’s always initial troubleshooting and “teething problems” that are smoothed out within a week or two. When these bugs invariably came up, it would lead to a tense day where the life scientists were forced back to handmade gel and I would be hurriedly trying to fix the bug so the machine could be used the next day. Fortunately, these bugs were quickly ironed out, but I found it interesting how the life scientists and engineers had different expectations for bugs in a prototype. Life scientists (as end users) understandably viewed bugs as a bad thing, while as engineers, we seemed to view finding and eliminating bugs as an inevitable step in making a reliable product.
Collaboration with life scientists and technicians was particularly important in making this device useful to them, as I changed quite a number of things because of their feedback. One example is the pour speed: too fast, and the gel splatters, too slow, and the gel can solidify before it’s all poured, sometimes leading to uneven slabs. Another example is the presence of bubbles in the gel. I won’t go into the details of the solution to that problem, but the feedback from life scientists on exactly what size and number of bubbles could be tolerated was invaluable.
In my time at the GSC I also observed some other interesting differences between the engineers and life scientists. To a life scientist, placing a massive shaker on a cantilevered table seems completely reasonable, but to an engineer it screams “resonance” and “broken table.” On the flip side, the colour of a table top seems cosmetic at best, and irrelevant at worst to an engineer. To a life scientist, a black tabletop can mean being able to see bubbles in gel, while a white tabletop makes seeing those bubbles nearly impossible. In general, it seemed like there were a lot of things that seemed basic to engineers that weren’t obvious to life scientists, and there were a lot of things obvious to life scientists that would never have crossed an engineer’s mind.
The automated gel machine is now used on a daily basis at the GSC, and is (as far as I know) bug free and reliable. The shaker is on a table with proper supports, and the tabletop is black, and all is well in the world. In my short four months at the GSC, I certainly found that engineers and life scientists approach technology differently. However, by working together and combining expertise, I think that we were able to create some truly useful new devices.
*Slab electrophoresis for physical mapping, in this case.