I’d like you all to meet my new friend, the Biomek FK – a very expensive piece of kit. She’s a robot. She’s meant to make my life a lot easier, and my experiments more reproducible. She can take the contents of a stack of 96-well plates and distribute them into a dozen replicates of 384-well screening plate in about the same time it would take me to manually pipette two rows.
And she looks a lot sexier doing it, believe me.
Better living through robotics
I just had the opportunity to use her services, and I am smitten. You guys ever ten-pin bowl? As the gleaming bank of arrayed yellow tips levitates over the mother plate, preparing to swoop down and pipette up a discrete amount of siRNA, that’s what she reminds me of: the apparatus that picks up all the pins and sweeps the fallen ones away when you’ve just choked on the strike. The machine that erases your embarrassment in a clatter of lost moments to clear the way for future triumphs.
High through-put is not all it’s cracked up to be, though. Working with this so-called ‘labor-saving device’ has been the most exhausting thing I’ve yet done in the new lab. Sure, the actual automatic aliquoting only took a few hours. But the mother plates still have to be created by hand: it took two of us most of a day, after which my thumb and wrists were trembling with fatigue. And the thing about generating dozens and dozens of plates is that they still need to be labelled, sealed, spun, frozen and thawed, manipulated in bulk, treated with kid gloves. Like preparing a meal for thirty people instead of two, you worry that some quality has been sacrificed.
And ultimately, of course, once the experiment is done (by further robots), the images will need to be analyzed by eye. We’ve got people working on the algorithms, but for most cell biological readouts, no computer can yet beat the human eye. And then there’s the mound of data that has to be annotated and stored: spreadsheets with fifty columns, that sort of thing. To say nothing of working out What It All Means.
Genome-side screens are amazing, and if they work, they can generate tons of interesting data. But I will admit to a bit of wistfulness for my previous stint, when I worked, painstakingly and lovingly, one just one protein. I knew its quirks and habits intimately, and the hypotheses were linear, not trapped in an infinite matrix of possibility.
Still, I love a challenge. Together, the FK and I are going to change the world. One well at a time.
Oh my. Yes, I remember the days of 15x 96-well plates for crystal screens, all of which had to be examined manually.
At t+1 day, t+3 days, t+1 week, t+1 month. . . ad tedium.
I think it might be prudent to buy stocks in the companies developing image algorithms now. I mean it’s absolutely amazing what they’ve managed to achieve, but once they crack it, we can all go home. 🙂