 Arden Shaw from the School of Biological Sciences, whose three-minute thesis title is Neural Networks to classify stars can also be used to measure proteomes. Can you hear me? Testing, testing. Great. Inside you there's a world of tiny machines called proteins. Usually we think of proteins as something that you eat, but in biology proteins are information. It's how your body communicates. And because proteins do so many things, when something goes wrong in your body, you're almost guaranteed to at some point find a protein struggling to do its job. But a lot of treatment ends up being focused on how do you compensate for a protein that's struggling. But a big problem in the field is we often don't know which protein is struggling. So we're trying to find ways to look at just the proteins that are different between a healthy state and a disease state. Well, there's a great technique to do that that was invented right here at Carnegie Mellon. It's called DGAY. It's a way to separate and compare all proteins at the same time. It's kind of like that game I played as a child, guess who? Where you start with a bunch of faces and you separate them based on different attributes, like whether they have glasses or a jacket. Well, we can separate proteins based on chemical traits. And you end up with a map that looks kind of like this. And each spot in that map represents a different kind of protein. And the brighter that spot, the more of that protein is there. So what we are able to do with this technique is run two samples at the same time. What that means is we can put a healthy state and label it all in green. And we can take a diseased state and label it all in red. And what that tells us is when we see a bright red spot, there's a lot more of that protein, but it's only in the diseased state. So most of the proteins end up combining to make yellow, and that's great news for us, because the differences pop out. These differences are like a needle in the haystack where we can focus treatment. Well, this created a new problem. We now have to compare really precisely the brightness of thousands of different spots. And a key insight is, well, we know who's good at comparing the brightness of many spots. It's astronomers. So we can steal some of their techniques, lovingly with proper attribution, and use the same techniques that were designed originally to measure stars and galaxies and teach that program how to measure differences in the constellation of proteins instead. And I love a few things about this. One is that it makes science more open, because the original software to analyze these maps, it's like a proprietary black box. But this is open source, and we know exactly where information came from. Another thing I love is that science can seem so technical and difficult, and it is, but it's also creative, and it's playful, and I want people to know that. And lastly, this is why science communication matters. When we can learn from different fields, when we can teach different fields what we do, we can ask beautiful questions, questions like, what happens if we look at proteins the same way that we look at stars? Thank you.