DNAnexus Snapshot: Five Minutes with Andrew Carroll

Andrew CarrollIn this series of Snapshot blog posts, we profile some of our fascinating team members to give you a better sense of who we are.

This week we’re talking to Andrew Carroll, a scientist who has been with DNAnexus for the past two years. Andrew has instrumental roles in collaborating with our academic customers as well as building the tool portfolio for the platform.

Q: What do you do at DNAnexus?

A: My title is scientist. On a day-to-day basis, that involves working with our partners: understanding what they want to do scientifically, and helping them put that into DNAnexus and run it. For example, I work with Baylor College of Medicine on one set of projects relating to Mendelian diseases. I helped them take the pipeline for that analysis and get it up and running within DNAnexus. At the end they get a set of genetic variants in their cohort, and their scientists evaluate those to see which might be causative. My other main responsibility is to find open-source academic tools and port them to DNAnexus so that anybody will be able to run them. I find these by interacting with the community — talking to people we work with, reading the literature, attending conferences. Then I assess the tools myself to see which generate good, robust results.

Q: Where were you before?

A: I completed a postdoc at Lawrence Berkeley National Laboratory. Before that I was getting my PhD in molecular biology at Stanford. I was in Chris Somerville’s lab doing research relating to bioenergy and plant biology.

Q: What attracted you to DNAnexus?

A: I really enjoy tricky computational work and computational challenges. I like working on any hard problem relating to information. Through Stanford I knew one of the DNAnexus founders, Arend Sidow, and realized that this company offered an opportunity where I could come in and have a lot of freedom to work on hard problems and have the solutions be applied in real research. Being here has exceeded my expectations. These partnerships have been incredibly rewarding and include work I might not have gotten to do anywhere else, such as run 3 million CPU hours for a project. That’s pretty remarkable.

Q: What have you learned since joining the company?

A: Since I come from a plant background, I’ve learned an immense amount about operating in the human health world. Things that people care about in these cohort studies, such as keeping data private, were just never a concern in plant biology. Also, I feel that we have some of the best engineers in the world on the team and getting to interact with them on a daily basis has greatly improved my computational ability.

Q: If you weren’t a scientist at DNAnexus, where would you be?

A: I would be a scientist somewhere; it’s way too in my blood to ask scientific questions. If I couldn’t be in biology, I’d be involved in the field of large-scale machine learning and neural networks. I’m fascinated by the way brains process information.

Q: Fill in the blank: There is probably a genetic link to ______.

A: Personality. I find it quite interesting that in the evolution of humans there must have been strong pressure on maintaining certain personalities. There were much smaller networks of people early on, so there would have been pressure to maintain compatible groups of personalities and diversity of skills and strengths.

Q: If you could have any person’s genome sequenced, whose would it be?

A: I wouldn’t choose a person, because the things that make a single person really special and unique and notable aren’t in our genes. I’d go for something novel, perhaps something in the tree of life that’s underrepresented. I don’t think we have a fern genome yet, so it would be good to fill that gap.

Q: Tell us one thing about yourself that nobody at DNAnexus knows.

A: I’m actually scared of heights, but I try not to let that change anything I do. When I hiked Half Dome in Yosemite National Park, the first time I looked at the cables I thought, there is no chance I’m doing that. But the second time I went I made it to the top, and it was terrifying but not nearly as bad as I thought it was going to be. It gave me an extreme sense of accomplishment.


DNAnexus Snapshot: Five Minutes with George Asimenos

DNAnexusIn this series of Snapshot blog posts, we profile some of our fascinating team members to give you a better sense of who we are.

This week we’re talking to George Asimenos, our director of science and clinical solutions. George came to DNAnexus when it first started, and has played a pivotal role in building the scientific and engineering foundation of the company ever since.

Q:  What do you do at DNAnexus?

A:  I’ve been with DNAnexus since the very early days, so I’ve done pretty much everything — from searching for office space, assembling desks, and unloading the dishwasher, to hunting for executives, writing code, and watching that code unfold on the cloud. These days I’m focusing more on science, exploring ways in which DNAnexus can help its customers reach new scientific frontiers and accelerate DNA-based medicine.

Q:  Where were you before?

A:  I was previously a student at Stanford University, doing a Ph.D. in computer science. It’s where I fell in love with biology and genomics.

Q:  What attracted you to DNAnexus?

A:  I could not say no to such a great founding team: my Stanford thesis advisor, Serafim Batzoglou; my Stanford reading committee chair, Arend Sidow; and my best Stanford student colleague, Andreas Sundquist.

Q:  What have you learned since joining the company?

A:  Things move very quickly at a startup company, and I’ve learned to welcome change. Where your office is located, what your logo looks like, who you report to, what your product does and who you build it for. As the ancient Greek aphorism goes, “Panta rhei” — “everything flows.”

Q:  If you weren’t a scientist at DNAnexus, where would you be?

A:  I would be a food critic, ideally an inspector for the Michelin guide. Tasting menus of the world would rejoice!

Q:  Fill in the blank: There is probably a genetic link to ______.

A:  Aging. It’s complicated, but DNA is certainly implicated. Hopefully we can make advances in understanding how they are related during this lifetime.

Q:  If you could have any person’s genome sequenced, whose would it be?

A:  I think it’s time to sequence some great foods and figure out what genes make for a tasty olive oil, a succulent tomato, and an aromatic oregano (ideally combined into a salad).

Q:  Tell us one thing about yourself that nobody at DNAnexus knows.

A:  Throughout Stanford’s SURG101 (an anatomy class where students, in groups of 10, dissect a whole human cadaver over the course of several weeks), my team ran behind during the week devoted to arm dissection. We were able to finish only the left arm of our cadaver. So I went back into the lab on Sunday, and spent all day dissecting the right arm myself — a life changing experience discovering the awe-inspiring complexity of the human body.

At Beyond the Genome, Talks Focused on Single-Cell Genomics and Informatics

Beyond the Genome, a top-notch conference hosted by Genome Biology and Genome Medicine, took place at the beautiful Mission Bay Conference Center right here in San Francisco last week. There were lots of great highlights from the event.

#BTG13 We were pleased to see the opening day focus so heavily on single-cell genomics. Nicholas Navin from MD Anderson Cancer Center gave an overview of this burgeoning field, sharing recent studies and walking through challenges such as assembly and distinguishing real biological variants from technical artifacts. This wasn’t new territory for us: one of the user-contributed apps on DNAnexus is a tool called SPAdes, a genome assembler designed specifically for single-cell studies. It has garnered lots of attention with single-cell and metagenomic studies for its ability to properly assemble DNA reads that other assemblers have trouble with. Looking ahead, scientists like Sunney Xie from Harvard spoke about moving toward studies of DNA and RNA from the same single cell.

The final day of the conference was devoted to informatics, so of course we were especially eager for those talks. Speakers highlighted best practices in genome assembly methods, comparing several different available tools. One of the reasons we think it’s so important to offer as many of the leading solutions as possible through the DNAnexus platform is precisely for this reason: scientists should be able to effortlessly compare the performance of various tools on their data sets. We also gave ourselves a pat on the back during a talk from Chris Dagdigian at BioTeam; he noted that Amazon has a multiple-year head start in cloud computing compared to most other providers. We looked long and hard before selecting Amazon Web Services as our cloud source, and it was nice to hear from an outsider that we made the right choice!

The final keynote came from genomics luminary David Haussler at UC Santa Cruz. He spoke about the new Global Alliance, encouraging data sharing and standardization methods to facilitate that. He also cited the need to get rid of silos that are keeping research data separate from clinical data. In a look at cancer genomics data, Haussler said cloud computing was necessary to conduct analysis in such large-scale projects, and that choosing a cloud-based solution allows users to bring computation to the data, rather than slowly and painfully moving these massive data sets to the tools. We heartily agree.