Erika Update #8 & #9: A first solid result, this definitely doesn’t work!

The first step is admitting you have a problem

Science involves a lot of mysteries, wrong turns, and dead ends. As they say, “if we knew how to do it, it wouldn’t be research.” It’s easy to fall into the trap of insisting that your science will work the way you envisioned when you first imagined the idea, back before you ever touched a pipette. A big part of effectively navigating a research project is knowing when to pivot. For me, a pivot is often preceded by realizing something’s not quite right or doesn’t match what I’d expect. This gives you a thread to pull on to unravel the mystery. 

From day one of the qtRNA project, there was something funny about which qtRNAs worked and which didn’t. Why did only three-of-the-four tRNAs from the literature work in my hands (Erika Update #2)?? Why are some serine qtRNAs totally fine, and others are super toxic and difficult to clone (Erika Update #3)?? Why can’t I made qtRNAs based on alanine and leucine, which, by all accounts, really shouldn’t care what their anticodon is (Erika Update #4)?? For months, the project was generating an ever-increasing number of head-scratching observations, and somehow I couldn’t quite figure out how they all fit together.

Finally, five months in, I got the first damning evidence that my measurement technique sometimes breaks down. In Erika Update #8 (linked below), I managed to isolate an example of a now-you-see-it-now-you-don’t qtRNA that looked functional on one day but non-functional the next. Then in Erika Update #9 (also linked below), I pulled on this thread. I went back and re-tested some of the qtRNAs I’d been puzzling over, and finally tracked the problem down: the way I was measuring qtRNA works great in most cases, but breaks down specifically for the tricky-to-understand toxic qtRNAs that were most important to study.

It was a relief to finally go from being suspicious that something was wrong to having clear evidence of a failure mode that needed to be fixed. I grew up a lot as a scientist from this moment. The unreliable thing I had been doing (optical-density normalized endpoint measurements) are pretty standard and work just fine most of the time. It was cool and slightly annoying to realize it was my job to identify when standard techniques will do, and when something more specialized is required. I felt like a person who had dutifully followed instructions, but the instructions were incomplete.

Erika Update #8 & #9 

Here’s the updates:

#8: 2018 5 31 - testing literature suppressor tRNAs - crosstalk

#9: 2018 6 2 - retesting literature suppressors - phage enrichment of Q-CAAA

In Erika Updates #8, I managed to isolate an example of a toxic qtRNAs that looked functional in my assay on one day, but looked non-functional on other days. Then in Erika Update #9, I pulled on this thread. I went back and re-tested some of the qtRNAs I’d been puzzling over, and I finally tracked the problem down. It’s really stupid, it goes like this:

• I was scoring “function” based on how much the bacteria glow divided by how much they grow

• One way to get a high score is to glow a lot. This was what I was expecting and this would equal function!

• The other way to get a high score is to grow very, very little!! 

And thus, if you have a qtRNA that’s very toxic it’s hard to tell if just slows down cell growth, or if it slows down cells and also makes them glow. The way I was gathering data (endpoint measurements) didn’t give me enough information to disentangle toxicity and function. To get the full picture every time, I’d need to switch to gathering data in a more sophisticated way (kinetic measurements).

Where’d it end up?

Erika Updates #8 and #9 were together a little under a month of work. During that month, only one data point ended up being published.

This is how much data got published from this ~month of work.

But don’t worry, it’s about to get a lot easier!

Want more?

If you want to follow along with this project, you can get updates by signing up through substack, or following me on linkedin or twitter.

If you have ideas for what I should cover in the blog post, suggestions for vocabulary to define, questions about the science, or other comments, please do reach out by twitter DM - I’d love to hear from you!