Computer scientist Aaron Watters has a strong sense of curiosity. Over the years, that curiosity has led him to working on projects from finance to the chemical industry. Now, he is a senior software engineer at the Flatiron Institute’s Scientific Computing Core, or SCC, where he helps scientists in all disciplines manage, process and visualize their data.
Watters is currently collaborating with a group of developmental biologists at the Flatiron Institute’s Center for Computational Biology (CCB), though he’s previously worked with other CCB scientists on gene expression, as well as with other centers, such as the Center for Computational Astrophysics (CCA), where he created visualizations of neutron stars and galaxy simulations. Watters has a doctorate in computer science from the University of Pennsylvania and a bachelor’s degree in mathematics from the University of California, Davis.
Watters recently spoke to the Simons Foundation about his work and the importance of data visualization.
What are you currently working on?
I’m partnering with researchers at the CCB and Princeton University on a project studying mouse embryos to determine how they develop at very early stages. Using a technology called light sheet microscopy, the researchers take 3D images of a mouse embryo as it’s developing to capture how two cells can divide and multiply into hundreds. This helps them see not only how the embryo’s cells develop but also their lineage — the parent, grandparent and so on of each cell.
My involvement is that I create ways to present these data, such as through developing tools and visualizations. This makes it much easier for the researchers to see what’s going on and what the data can reveal. For example, I recently created a viewer that lets the researchers scroll through the images of the embryo over time, at different stages of development. This viewer allows them to pick out individual cells from that embryo and rotate their view, so they can see the cells from multiple angles and at difference cross-sectional slices. It also comes with a flow chart — sort of like a family tree — showing the cells’ lineages. This chart wasn’t something the scientists initially asked for, but I was driven by my curiosity to make it, and in the end it’s proven quite useful to the team. The project’s website, which I also built, allows the researchers to share the data, their models, and the visualizations.
What drew you to working on this project?
The imaging technology being used here is relatively new, so the researchers are seeing a lot of things no one has ever seen before. At the very early stages, mouse embryos are strikingly similar to those of other mammals, including humans, so the researchers are elucidating basic principles of embryology that can be applied to all organisms. Everyone is very excited about the project. The scientists get very animated when talking about the work, and they’re always eager to share what they’re learning. That has made it a really fun project to work on, and I’m always learning new things.
I’ve moved between jobs a lot in the past, but I’ve been with the SCC for over six years now, which is the longest I’ve stayed anywhere. That’s because there’s a lot of interesting work that I can contribute to. I also have a certain amount of freedom to explore what I want and connect with people from different disciplines.
I’ve also been able to work on so many different projects here at the SCC. In addition to our mouse embryo project, I’ve worked with other CCB scientists on gene expression, and I’ve even collaborated with CCA astrophysicists — I helped create data visualizations to show models of neutron stars and the evolution of galaxies. I think I’ve worked with all of the Flatiron Institute’s centers except the Center for Computational Mathematics, but I hope to collaborate with them in the future.
What is the next phase of your mouse embryo project?
I’m working on some software to help align and correct the rotation between different scientific images. When the scientists are imaging the living embryo, it moves and gradually rotates. So for each image, they have to figure out how far and in which directions the embryo rotated so that they can align the images to confidently know they’re looking at the same cells from image to image. This is currently done automatically by a computer algorithm. I’m developing a tool that will help the scientists look at the images and be able to confirm whether the algorithm is working correctly, and if not, make quick corrections by clicking and dragging the images until they align.
Why is data visualization important?
Visualizations empower scientists to truly understand their data and what it’s trying to tell them. Oftentimes, visualizations are employed toward the end of a project, right before a researcher might be ready to publish their work. But I think visualizations should be used more in the very early stages of research. They can be extremely useful in helping a researcher figure out what direction to explore next.