Today we’re sitting down with Mike Lin, a senior software engineer who joined DNAnexus in February 2012. Mike was previously at MIT and the Broad Institute, where he worked in Manolis Kellis’s lab on comparative genomics of protein-coding genes. He comes from a computing background, but started taking biology-related course work as an undergraduate student when he realized that bioinformatics looked like an exciting and growing field.
Q: What do you do at DNAnexus?
A: I’m a software engineer, working to build our new collaborative platform for analyzing DNA data in the cloud. Right now I’m particularly focused on scaling the back-end architecture to enable it to process thousands of human genomes per month.
Q: Where did you work before?
A: I recently finished my PhD at MIT and the Broad Institute, where I worked on machine learning techniques to study the evolution and function of protein-coding genes in the human and fruit-fly genomes.
Q: What was it about DNAnexus that attracted you to the company?
A: I believe the cloud platform we’re building at DNAnexus will accelerate scientific collaboration in genomics, and thus help fulfill its potential for human health and medicine.
Genomics is a highly interdisciplinary and collaborative field, but the “big data” it produces makes collaboration really difficult at times. For example, when I was involved in research at MIT, we had access to some amazing compute facilities there. But our collaborators at other institutions used their own compute clusters (and weren’t allowed on ours anyway), so we’d have to send massive files around by e-mail or FTP. Sharing analysis software was even worse, as getting someone else’s code to compile and run often led to hair-pulling frustration. These are big problems that impede progress of the entire science.
A secure cloud platform on which you can instantly share data and analysis software, and run analyses at scale without huge upfront investments in hardware and pipeline engineering, will really make a difference — and that’s what we’re building. Of course, there are also amazing commercial opportunities when you think about millions of people having their genomes sequenced in this decade.
Q: What have you learned since joining the company?
A: I’m continuing to learn a ton about working on a team of professional software engineers, which I didn’t have much experience with in the past. “Agile” sprint planning, continuous integration testing, distributed source control, and the human elements of keeping everyone focused and coordinated — all examples of important professional practices that are quite hard to get right, and weren’t really on the radar in academia.
Q: What would you do if you weren’t a software engineer?
A: If I hadn’t come to DNAnexus, I probably would have stayed in academia for postdoctoral research. Failing that, let me say airline pilot!
Q: What person had the greatest influence on where you are today?
A: I owe a great deal to my PhD advisor Manolis Kellis, who taught me not only everything about genomics, but also about the importance of taking risks and staying positive.
Q: Fill in the blank; there is probably a genetic link to _____
A: Circadian rhythms; I think there really are “morning people” and “night people.” I’m the latter.
Q: If you could have anyone in history’s genome sequenced, whose would it be? Why?
A: It’s hypothesized that there were actually a “Y-chromosomal Adam” and an “X-chromosomal Eve.” Genetic models of human populations would probably be quite a bit more precise if we could get our hands on those. And there’s the not-unrelated “Genghis Khan Effect.”
Q: If you had your genome sequenced, what would you hope to find (or not find)?
A: Like a substantial fraction of people of East Asian descent, I seem to have a lousy ALDH2 gene that severely curtails my participation in certain common social activities. In fact, I suspect I have not one but two bum alleles. I’d like to know for sure!
Q: Tell us one thing about yourself that nobody at DNAnexus knows.
A: I used to drive a purple car. It was a Volkswagen with a pearl-coat paint that on a cloudy day looked blue, but in the sunlight showed unmistakable purple streaks. Of course, I bought it on a cloudy day.