Ted joined Fulcrum Genomics in November 2022. He describes himself as a reformed physicist because he has migrated from theoretical physics into complex systems, leading to eventually explore structural variants and develop experimental SV pipelines. He's stayed in flavors of biology ever since. We recently sat down to chat about work and life as a bioinformatics consultant.
What’s your area of expertise, and what excites you about your work?
Ted: I came into Fulcrum as a pretty strong software developer, in python in particular, but also in C++, so I’ve had compiled languages that were strongly typed as well… I do a lot of different things with that background. As someone who has changed fields after an advanced degree and post-doc, I have the curiosity and drive to learn new science in each project. You just pair me up with a subject matter expert and I have the experience to ask the right questions to get going, and the software development part is something I feel very comfortable with.
I do enjoy learning things, so learning the science is the fun. I like making stuff, like making a tool and knowing that it gets used, and especially when it gets used for helping people. One of our clients is working on a precision cancer vaccine and that feels really cool, maybe someone will get a medical treatment that helps them. That’s fun for me to think about.
What’s a common challenge in our industry that people don’t talk about enough?
It definitely feels like there’s a lot of rusty, cracked infrastructure, there’s a lot of essentially abandonware that everyone still needs, there’s a lot of stuff that’s built on the back of academic research, which - I want be careful - I value that. I come from academia! That’s definitely where these things get invented. But often when it’s time to scale it starts to fall down in terms of quality control. And there’s a lot of duplication of effort, so many groups solving the same problem from slightly different angles and maybe not focused on reproducibility and reliability and robustness. There’s not as much maintenance, because maintenance is not as exciting or career-wise as rewarding, oftentimes, as invention. It’s hard to find funding for this work.
What’s one tool, tip, or mindset shift that has made a big impact in your work?
I came in from python, which was designed to be scrappy, you can just throw things together, and you can make that work. It doesn’t even have typing! That was kinda bolted on afterwards. I’d never heard of linters, or auto-formatters, or type-checkers, or auto-actions in GitHub, things like that. I think I would have appreciated them starting out, but I know many people might not because they’re extra work.
But the amazing thing is the number of times these tools capture errors that maybe you wouldn’t have seen otherwise, things that would have just been in there giving you the wrong answer. That kind of mindset shift of slowing down to go fast and using tools to catch mistakes early, being systematic about that kind of thing, has been a transformation, particularly coming from academia.
What’s a recent project or insight you’re particularly proud of?
For one of our clients, starting from a draft tool they made, I wrote a tool that can look at splice sites and predict the potential RNA sequences that we might see as a result of the individual splice events. That project included new biology, math, and programming questions — all three of them were a challenge for this project but I think it came out pretty nicely! One of the things that aberrant splices can do is they can combine - there are cases where one splice may not be viable on its own, so you can have a case where an aberrant splice goes nowhere and does nothing on its own. But if you have two aberrant splices it can now pop back to somewhere correct and you have a new RNA sequence. My tool handled every combinatorial possibility, very efficiently. It was a discrete piece of work we finished and they’re using it and that feels so good. We finished pretty quickly too, in a month or so.
If you could give biotech startups one piece of advice, what would it be?
This isn’t biotech-specific, it’s true of startups across the board. Since at least the early 20th century, there’s been a lot of research into how to get the most out of workers. And the conclusion is surprisingly consistent: humans have a ceiling for sustainable productivity. Push beyond it for too long, and not only does total output suffer, but mistakes increase, especially in creative or complex work where productivity is harder to measure in the first place.
So my advice? Know the ceiling for your industry. If you push hard during a crunch, ok, but you must build in time to recover. Otherwise, you’re not actually getting ahead. You’re just burning out your team and reducing quality. There’s a whole literature on this, and it’s worth reading. The science is clear: long-term steady beats short-term sprint.
What’s something outside of work that inspires how you think about problem-solving?
How I think about problem solving is a very abstract question for me. The brain is a strange thing and I think mine is particularly weird. When other people talk about closing your eyes and visualizing something I thought they were just talking in metaphor because I can’t do it - I can’t close my eyes and see things. I am very verbal, I think in terms of sounds and words but I have close to 0% visual imagination.
I know my wife’s eyes are hazel because that’s the word she uses to describe them. I know my daughter has striking blue eyes because people always comment on them — sometimes even questioning whether she’s mine. It took someone pointing it out for me to remember that my own mother has blue eyes too. My visual memory is basically zero. I can remember things procedurally, which I think is why I program well. I can remember algorithms, procedures, ways to do things, words, descriptions. No pictures.
So for me, describing how I think about solving problems isn’t straightforward. It’s shaped by a brain that processes the world in a less visual, more structural way.