Scientist Stories: Jeremy Freeman, Using Computation to Understand the Brain

Thanks, everyone. I'm super excited to be here to be talking about the Data Coordination Platform and the data coordination effort. John already gave a fantastic description of this architecture and what all the different boxes do, so what I really wanted to do here is just say a little bit about how this project is sort of actually functioning, the people that are involved right now, how the collaboration is formed,, and then something about how we're at CZI trying to support it. So for this kind of project, a really critical part is leadership. And one thing we've been really excited by is that leaders from the community and from these institutes have really emerged naturally to sort of take on the charge of leading this project. I want to highlight those people here. Tim Tickle from the Broad, Brian O'Connor from UC Santa Cruz, and Laura Clark from EBI. Each is more or less the owner of a respective box. Brian gets bonus points because he picked his outfit to match the color of his box. We're working on getting green and purple outfits for the others. So it's been really amazing just to watch this group of 3 form and really take leadership over the different components of this project, as well as see teams form around each of them that are really working to actually build and implement all these components. One of the things at CZI that we're most excited about in this project is collaboration and bringing new kinds of expertise and different kinds of groups of expertise together to work on a problem of this complexity and a project of this complexity. I think in particular for this kind of work, there's 3 core things that it's really exciting to bring together. One is computational biology, which, you know, to us is really about framing the both scientific and computational nature of the problem. That's so much of what this workshop is going to be about. A second component is software engineering. So this is actually implementing algorithms and building architecture. And a third component is product management. So that's really about making sure the engineering effort is aligned to the needs of the computational community and biological community in general. So at CZI, as Corey mentioned, we've been building a technology team that is now very involved in helping to build this architecture, but we're doing it in collaboration. So we are working collaboratively with the groups that I was just talking about to sort of work together and bring all these different kinds of expertise together on and aligned to a big complicated project. To sort of evidence that these collaborations are happening, and we think it's really critical they happen in person, here is a real software engineer hanging out with a real computational biologist walking— the faces are concealed. That's John Marioni. And that's Bruce Martin, who's in the back. And, you know, I'm kidding, but really, actually, one of the coolest things that's happened over the last several months is a lot of face-to-face meetings. We've been hopping around different institutes, meeting in person, getting to know each other, really having fun developing and getting this project off the ground. And I really think, you know, collaboration, bringing people together, is critical for making this kind of project work. Along with collaboration, the other core theme from CZI is openness. So John already mentioned the value of having an open platform in terms of accessibility of data. To us, it's really important that the project itself be developed in the open, and to us that means a few critical things. One is that all of the code is open source all the way down, so not only the code that powers the platform, but the code for all of the analyses that are going to be implemented. Analyses have to be reproducible, they have to be transparent. As scientists, we have to understand what we're doing. So it's really critical to us that everything be open source. We want the development itself to happen in the open and be collaborative. So that means code is being written on GitHub. There are roadmaps that are publicly available. I'll give links in a second. And as far as the platform is concerned, we want it to be extensible, especially when it comes to analysis. So there's so many great ideas coming from this community, so many great things we've already heard about just this morning. And especially in those green boxes and red boxes, we want to make sure the platform is as easy as possible for people in the community, in the scientific community, to incorporate their analysis ideas and run their analyses at various stages of processing of a wide variety of data. In terms of, sort of, what's out there to look at, there's a GitHub repository you can check out at github.com/humancellatlas. There's already a whole bunch of repos there, some of which have prototypes, some of which have code. There's a wiki that has links to a bunch of different documents. If you want help knowing what's out there and sort of walking through it, The other thing I'd recommend is joining the Slack channel. So it's at humansellatlas.slack.com. If you need to join, you can visit that link there. We've tried to make this Slack channel as open as possible, so literally anyone can join. Hopefully we don't get spammed. And there are, within this Slack domain, there are a bunch of channels specific to each of the different components of the project. So just come in, start chatting with us, ask questions. We can send, you know, provide links to roadmaps and other documents that might be of interest. And there are, you know, hopefully lots of ways to get involved, both to contribute, to join the discussion, and hopefully to ultimately build tools for the platform. So with that, I want to give it back to Corey for one last thing to say.

So the last thing we want to say is that this whole project has been happening out of the goodness of people's hearts, like so much of the Human Cell Atlas. And people have already shown a lot of enthusiasm and and commitment to this project, even though they were, in theory, being paid to do something else. But we are now formally announcing that the Chan Zuckerberg Initiative is going to be funding the data platform for— the data coordination platform for the Human Cell Atlas with— and we're really thrilled to be working together with the leaders of the field at the Broad Institute of Harvard and MIT. The European Bioinformatics Institute, and the University of California, Santa Cruz. These are people who have been great contributors to genomics platforms in the past. They're great innovators and data— in data sharing and informatics, and there will probably be, you know, a variety of different people at all of these groups, but, you know, probably about— at the moment it looks like about 10 people at each site working together on this problem, and we're really thrilled to have them here. And we're thrilled to be part of the project.