Verity Hill: EEEV in the Northeastern US

yes so uh I am a postto uh with Nate guar at Yale um thanks for having me here today uh it's nice to talk about some Alba viruses again um so I'm going to present some work that uh we've been doing on easn Ain and citis um it's going to be a bit of a Whistle Stop tour today um the original draft of this paper was like 10,000 words long there's a lot of stuff in it so um if you're interested in any of the extra stuff uh it's all in this uh in this paper but um okay so quick introduction to Triple E because it's one of those uh weird little niche American and ketis is um so it lives mostly in Birds we don't know exactly which birds are relevant here but probably small uh song birds um it's transmitted between Birds by mosquitoes mostly kulac melanura but there's a few different ones um and then it spills over into humans and horses which are dead end hosts uh because they don't produce a high enough vmia to pass on the infection and it causes severe disease in both humans and horses with high fatality rates and um lifelong neurological injury in the survivors um so mostly it's up in the the northeast of the US um in this kind of regions this is uh New York Massachusetts and Connecticut um uh although it is present in uh for example Michigan as well Michigan has quite a few cases and kind of up across the um the East Coast at sporadic but really uh Massachusetts in particular uh you can see here there's this kind of focus uh in the the southeast of these three states and then in Upstate New York as well um so why do we care about it in 2019 there was a large outbreak of Triple E um when I say large it's uh not large in the scale of a lot of outbreaks um in the Northeast there were 19 human cases and 26 course cases um previously uh sorry no and then in across the whole country I think there were about 38 and a lot of those other ones were in Michigan so this was um this was a big year and obviously rais some concerns that these bigger outbreaks might be happening more commonly um so before we can ask if something changed in 2019 we have to understand what the normal Dynamics are right so to answer this question um we turn to the genomes because I am fundamentally a genomic epidemiologist um Triple E is a select agent um which makes it really hard to work with and even the RNA is a select agent so sequencing it is really tricky um thankfully the woodsworth center in New York State uh is able to work with it uh so they sequenced 80 new isets um 48 of which were from 2019 so you can see this like this big jump in the number of sequences there that's pretty much all the woodsworth um they're pretty much all from mosquito pools these sequences some from some exotic birds some horses uh things like that um and what's really good is that so you can see from this map here the stars are the newly sequenced isolates lots of those counties didn't have any sequences from them at all so it really expanded our data set um geographically as well as temporally and just like this sheer number uh of sequences so we used these to build a uh F Geographic analysis so this just means that we treated the location the sampling location of the viruses as uh a trait that evolves across the tree uh and what we found in this time resolve tree here that's colored by that inferred location uh is the Florida is the source of Triple E in the US so you can see that because this the backbone of this tree here is all in Florida so that means that it's spilling over from Florida to other parts of the US and that's what this map shows here these lines the thickness is the uh frequency that you see them in the full B in posterior um just thick is more right so um you can see most of it going from Florida up to uh New York and Massachusetts and Connecticut and also Virginia which is really interesting because we don't have a lot of data from Virginia so there's definitely some uh some unanswered questions there once it gets into the Northeast um these clusters are usually detected within like a year or two of introduction that's what this plot shows here um and then they they tend not to be super long so they're around for a couple of years um in the Northeast some do definitely stick around a little bit more there's a long Tale on that but fundamentally pretty quick um so we have 26 introductions into the Northeast in this uh time frame and you can see three kind of rough waves here so we've got like a group kind of haish a group kind of haish and a group kind of haish it's a bit it's a bit rough but we think these waves probably refer to some underlying process in the bird immunity um which I'll kind of come back to a little bit later um and most of these are in one state uh which you can tell by the colors uh the mosquitoes don't move very far and maybe once the birds have finished migrating they don't move very much I don't know really not an ecologist and I think a lot of people in this schol are so uh we can come back to that later if you're interested um there are six clusters which were sampled in multiple States um that's these six pulled out here you can see that it's not just that they're in like bordering counties like there is some genuine movement and five of the six start in Massachusetts and then go elsewhere so even though multi-state spread is rare within the region when it does happen Massachusetts is a regional source so next we want to ask what drove the outbreak in 2019 right so we we understand what the the normal Dynamics are um so first the one lots of people were worried about especially in this kind of postco era where we have a lot of variants um is adaptation uh so there should be some kind of philogenetic signal of adaptation even if we don't exact understand exactly what the phenotypic effects are so um this tree here is the um the time resolve philogyny of all of the Triple E sequences that we have so all of these ones at the end here are our 2019 sequences you can see they're all across the tree right there are no unique shared mutations to those 2019 sequences so this outbreak is not caused by no variant that's better adapted to infecting mammals so that's a really good like first thing that we're able to rule out um we then tested a heap of other stuff and got no signal in anything so you can see here it's pretty much all in the same regions that have detected it before those stars are in areas which have already had them or like you know right next to it uh sorry right next to it um there was no change in the case demographics in the humans there was no change in the underlying demograph Dynamics of the virus we tested whether humans were going outside more in 2019 they weren't we did some stuff with uh bird banding data we couldn't find any signal for that although I will say we weren't exactly sure which species to look at so there may be some signal there but it's hard to pass out so what else is left the mosquitoes so uh shockingly enough it was also a big infection year within the mosquitoes so the bars here are the infection rate and you can see that in Massachusetts and Connecticut where most of the cases were it was a big year for mosquitoes uh as compared to New York which didn't have so many cases it's a more normal year and when we uh look at the number of positive pools against number of cases they match pretty well that's not super shocking right if there's more infected mosquitoes around there's more cases um we started trying to look if there was anything different about 2019 specifically so these two plots here show values uh from 2003 to 2019 with the red line in 2019 for mosquito abundance infection rate Vector index which is infection rate times abundance um then we've got this Index P which is uh a modeled estimate of reproduction number based on climate factors like temperature humidity all of that kind of thing so for indexp we can actually see that it's pretty normal for 2019 right that red Line's right in the middle of the pack like it just shows that it's seasonal um um but what's interesting in 2019 is that you can see that this infection rate kicks up early and kicks up quickly so we actually have extremely high early Vector indices uh in both States uh in 2019 and these were the highest Vector index values in the entire 17-year data set so when there's a lot of infected mosquitoes there's a lot of human and horse cases but can we predict when there's going to be a lot of infected mosquitoes so we know that the climate fact factors at least in the Northeast don't have enough um information in them because we can't predict that introduction of Triple E into the Northeast from Florida um but what we tried to do was build a short risk a short-term risk model from the surveillance data in Connecticut and Massachusetts so Toby Co who is a really brilliant um he used to be a research assistant with us he's now a uh PhD student at Colorado State um built this model uh that took into account Vector index month of detection the year the month of the year Index P all of these things um to see if we could uh predict the cases you can see here it actually in Massachusetts where there's more data predicts it um pretty well and we've actually like even the big years to it it gets it pretty nicely and the main thing that came out of this was that was the month of first detection of Triple E then predicts a uh a worse season but the turnaround time required here to make that useful is really fast right so you have to be doing detailed mosquito surveillance with a a turnaround of like a week to get that positive uh result so we would like to understand the longer term Dynamics um but we actually can't yet do that we don't know what the deterministic processes are in the longterm Dynamics um and to understand that we need some more information on the ual human prevalence is this a rare disease with a high infection fatality rate or a more common disease with a low infection fatality rate because then obviously changes in the underlying transmission Cycles are going to have different impacts on those two scenarios the really the Holy Grail is really the bird data uh we desperately need more information on which birds are relevant what their immunity cycles look like what their immunity to Triple E looks like these sorts of things and finally what's happening in Florida because it almost certainly precedes large outbreak years in the Northeast but we haven't had sequencing data for Triple E from Florida since 2014 they're extremely busy as an aror virus Research Unit they've got much bigger fish to fry than Triple E but this is really one of our like our missing uh puzzle pieces so just quickly to thank uh all of these people the mosquito surveillance teams because they do extremely solid work year in year out um and this is Toby here who did um the modeling work in this and obviously uh Nate for supervising uh so s things sorry that was very Speedy if anyone has any questions I'd be very happy to talk about it