Measles Dashboard - Now Live!

Genomics and wastewater - two of my favorite data sources

Excited to share with you today a project I’ve been puttering away on for a long time now and am finally getting around to sharing - a measles wastewater and genomic surveillance dashboard! I’m a big fan of wastewater-based epidemiology (WBE), and the US National Wastewater Surveillance System (NWSS) has been a great addition to the national surveillance infrastructure. Measles presents a great use for WBE because it is something that ideally we wouldn’t expect to be detecting at all, so any detections are meaningful signals of something happening in the community. Unfortunately detections have not been rare, with measles detected at 28 different wastewater treatment facilities in January alone. So lets walk through what’s available on the dashboard!

While there are several visuals available on the NWSS measles website, I wanted to be able to look at the data in a few different ways, so I pulled in the full dataset from the CDC to power all the different visuals for the dashboard. First up is a bar chart where you can choose whether you want to see the # of samples with detections, % of samples with detections, or unique sites. Test out hovering over each bar to get a few more details as well!

Next is the hexmap! This is a way to represent the US which makes all states equal-sized so it’s possible to actually see all those tiny states packed into the Northeast. I went back and forth a lot on how to make the states with detections jump out, and ended up going with a checkbox that allows you to toggle off and on a focused view the fades away states without detections and adds thicker outlines on the states with detections in the time period you select.

After the hexmap is the heatmap (I’m apparently into the map visuals!). By default it shows you the states that have had detections in the most recent month that has data, but you can uncheck that and choose whatever states you want to look at. You can also click on the cells or bar in the hexmap and heatmap to get more details about that particular set of data!

Genomic data

There are three different sections for the genomic data - charts, a multiple-sequence alignment viewer, and a phylogenetic tree viewer. The data for the top-level charts comes from Pathoplexus. There are two charts, one that shows sequences over time colored by genotype, and one showing the number of each genotype for the given time period and geographic location selected. The default view is all US states, and Jan 2021 forward - you can change either and that controls both charts.

Historically, the N450 region has been sequenced and used for genotyping. I added a multiple-sequence alignment viewer so you can explore that particular region, making adjustments like nucleotide or amino acid, and filtering down by location or date. This data comes directly from the Nextstrain measles page - they do a great job of curating sequences and making the outputs available for downstream users as well as on their site.

The second genomic visualization is also powered by the Nextstrain outputs - this time as an interactive phylogenetic tree using the Taxonium tool. The Nextstrain team rebuilds this tree over time as new sequences become available, and I’ve set up a Github Action to pull in the updated tree once a weekly. You can also explore the tree directly at Nextstrain where they have some additional filtering, coloring, and display options. It’s a tremendous resource, so please share that with anyone who could benefit!

It’s been a fun adventure working on the infrastructure and thinking through the different data visualization options. If you get a chance to poke around measles.genomicepi.com I’d love to hear feedback on what works well, what you think could be done better, and anything else you’d like to see!