 Have the slides come through, Sean. No, they're still filtering through the system. Well, I could just plop this through in an old fashioned and 19th century sort of way. I think I can make most of my points. And I apologize for not getting this right the first time. Sounds like a plan. Chalk talk without the chalk. OK. So you all know about shiny. And it's certainly very popular and very capable. I work primarily on matters of gene regulation in erythropoesis. And there's always new kinds of data. Most of the visualization that we see that takes advantage of all of the very fine JavaScript libraries, most of them are presentation graphics. And so, for instance, today we've seen some fine graphics. But I think the way that most of this work is much more exploratory. We spend most of our time exploring data, new kinds of data, looking for insight. And so what GVR does is it establishes an open and reciprocal connection between your R session and JavaScript in the browser. We do this using the same WebSocket library, HTTP UV that shiny is built upon. And the simple, the single, and the key differences that is that when you launch visualization in shiny, you lose your console. All manipulation then occurs within the viewing the browser. So I build upon the redoubtable IGV.js from Jim Robinson and his crew. And the IGV.js is the web bit-paced version of the IGV desktop that is probably known to many from the road. The scheme is that there's many kinds of tracks and they can be created through data manipulation in R, frequently made more tractable by just being covering a region of interest. So the simplest thing is if one has a simple bed-like structure with a chromosome name, start and stop, you create a data frame, you send that to the browser, and you see that track displayed. The next step up is rather than an annotation track, it is a quantitative track and think wig there where perhaps every chromosomal location has a different numerical value. And they too are either read-in in R or created on the fly. And again, one just creates a track of the right type and displays it. And so this sort of cumulative process of different sorts of tracks, we also support Jim Robinson's GWAS track and VCF. It's a very flexible system to think through your data learning what you can by visualizing it, which I think we all know from the basics of tried-and-true scatter plots. If you look at the data, you often see things which are not apparent if you just look at numbers or statistical summaries. Is there any hope on the slides? We've got the slides. We can't get your face off of the screen now. Not that we don't mind looking at you. You're very funny, too. Maybe up for discussion. The other thing that, OK, we've got the slide, but we're also running short on time, so you're going to give us the Blitz version. Yeah, one minute countdown, boss. One minute. So if you show the final slide, and so this is a good example from a current problem. So we have CTE-CF chip seek in the vicinity of Gata 2. You can see the pileup. It was just a commander to render that data. Then the narrow peak is shown in red. And then motif matches in brown below that. And histone barks at the bottom. And my claim is that this is all very straightforward to do with data bioconductor data types and that the aggregation of these different kinds of data can be helpful. I will say I wish I had known about this one before our talk because I used it instead of regular IGV. This is terrifically powerful. It seems like it can't possibly be any harder to use than the web sharing. So thank you for a wonderful talk. I don't think we have a ton of time for questions, but if people want to ask them online, I'd highly encourage you to check out the slides. This is a tremendously powerful tool. And next up, we've got our 10 o'clock. We'll try to make these slides available to everyone, too. So apologies for the technical.