Environmental DNA (eDNA) has been applied to the surveillance of invasive crayfish populations with mixed results, with some researchers suggesting that crayfish may be less detectable than organisms like fish due to their exoskeleton and benthic habitat use. Dr. Eric Larson from the University of Illinois Urbana-Champaign reviews recent studies from his lab and collaborators on eDNA applications to invasive crayfish surveillance. These include laboratory tests of the role of the exoskeleton in affecting crayfish eDNA detectability (i.e., pre- and post-molt) and comparisons of crayfish eDNA detectability to other taxa from field studies. Results of these studies suggest that crayfish eDNA detectability is minimally affected by the exoskeleton. Rather, crayfish eDNA is likely derived from metabolic wastes and is most detectable at times of high activity (i.e., warmer water temperatures in late summer and early autumn).
(Natalia) Alright, hi everyone! Welcome to this Invasive Craigfish Collaborative webinar, the first webinar of 2024! Woo! Happy New Year! Um, I hope that wherever you are You’re safe and warm. The snowstorm and the rainstorm is getting pretty wild here in Northeast Illinois,
Um, but it just reminds me of how grateful I am that these seminars are virtual. Um, so, my name’s Natalia. I am with Illinois Indiana Sea Grant and am the main facilitator of the Invasive Crayfish Collaborative. And for those of you who don’t know what an ICC is, it’s a program that
Brings together a variety of experts and stakeholders to address the threat of invasive crayfish in the Great Lakes. Um, we create monthly newsletters highlighting crayfish literature and news. We host webinars like this one about new crayfish research and programs. Um, and work and collaborate with other organizations and just promote responsible crayfish practices.
Um, so before we begin, I, as usual, will quickly share a list of different links that you can follow if you’re interested in learning more about the ICC. So you can check out our main website, invasivecrayfish.org. Which acts as a one stop shop for invasive crayfish, uh, information and is always
Being updated with new information. Um, and if you’re interested in joining the ICC membership, you can follow the join link to find our subscription form. And you’ll subscribe to our monthly newsletters and will receive notifications about our upcoming webinars, meetings, and other events. Um, and if you’re interested in registering for future webinars
Or viewing past webinars, you can follow these last two links. Um, and then finally you can email me at szklaruk@illinois.edu if you have any questions or comments, um, and I believe that we are adding these links to the chat so you can easily access them.
Um, so one of the goals of the ICC is to make new information regarding invasive crayfish more accessible. Which in turn will make people more informed, enabling them to make better decisions about invasive crayfish management. So in today’s webinar, Dr. Eric R. Larson from the University of Illinois, Urbana Champaign will review recent
Studies on the applications of environmental DNA, also known as eDNA, to invasive crayfish surveillance. And that includes lab tests that investigate how exoskeletons affect eDNA detectability. And also field studies that compare crayfish eDNA detectability to other taxa. Um, and for those who don’t know Eric, he is an Associate Professor in
Natural Resources and Environmental Sciences at the University of Illinois, Urbana Champaign. He is a freshwater ecologist working at the Interface of Native Species Conservation and Invasive Species Prevention and Management. He holds a bachelor’s degree in fisheries resources from the University of Idaho, Moscow, and a PhD in aquatic and fishery sciences from the
University of Washington, Seattle. Thank you so much, Eric, for being here. Um, so to everyone joining us today, we encourage you to type in any questions you may have in the Q& A box or in the chat at any point during the presentation,
And then after the presentation, we will go through them together. Um, and also as a reminder, this webinar is being recorded, and it will eventually be posted on the ICC website with captions. Um, and alright, so with that, Eric, you can go ahead and share your screen. (Eric) Yeah, great.
Thanks for having me. And, um, I think that’s probably working. (Natalia) Yes, it is working. (Eric) And, um, yeah, nice to see, uh, so many friends in the audience, including folks who contributed to some of the studies here.
Um, you know, I, I kind of want to talk through a handful of our lab’s eDNA studies for crayfish that are, um, in most cases, uh, currently in review. Things that are kind of These are all in progress out of our lab work that relate to environmental DNA monitoring for especially invasive
Crayfish in the Great Lakes region. And, um, in doing that, I’m not going to spend too much time on background and environmental DNA. I think this, um, this set of kind of related approaches has boomed so much over the last few decades that most people are more or less familiar with
What we mean by eDNA, but this is DNA that’s been captured and identified from an environmental sample, something like water or soil or increasingly air. Using a variety of approaches that have been adapted out of uh, we’ve, uh, worked in microbiology or microbial ecology, but where we’re thinking about eDNA
For, for macrobiota, something like a crayfish in Lake Michigan, you know, this is generally DNA that we’ve collected and identified without collection of that, that target organism itself. Something that might be sloughed cells in the environment, intracellular DNA or extracellular DNA after those cells have broken down, but we
Don’t have a whole crayfish in hand. And that can focus on a single species, uh, or maybe sometimes, uh, levels in the hierarchy above or below a species. Some of our eDNA assays might target, target, uh, a genus like Corbicula that we might be interested in.
Uh, and that tends to use approaches like PCR, sometimes quantitative or real time PCR, things that look very similar to what a COVID test might look like if you do a nasal swab at a medical clinic. Uh, eDNA can also recover whole communities where we use universal
Primers to maybe hopefully identify all vertebrates or all invertebrates from a water or a soil sample at a site. And that applies a variety of kind of next generation sequencing approaches that in generating hundreds of thousands or millions of DNA reads from a sample
Requires a bit more bioinformatics muscle to interpret those results. And, you know, this EDNA approach really spilled over from the, the microbiologists to, to field biologists, to ecologists, to, to people working in fisheries or wildlife, uh, around 2008 or so with some applications to, uh, for example, uh, invasive bullfrogs in, in France.
And, you know, this is a, a, a web of science subject search for eDNA that I did yesterday for that time period 2008 to 2023, uh, about 3,500 papers use the term eDNA over that time interval, and by our categories, uh, from, from Web
Of Science, ISI, this is primarily in ecology, marine and freshwater biology, biodiversity conservation, fisheries, some of these are, um, These are not microbiology studies, but a lot of these are people looking to identify macrobiota in the field from DNA.
And here on kind of our chart of the number of publications on one of the two y axes over time, there’s been this enormous boom in eDNA research in the last 15 years from that kind of 2008 watershed moment where this spilled over to
Macrobiota to what might be a plateau at about 500 to 550 new eDNA papers a year. And so anyone on this call working in eDNA is kind of familiar that this is sort of an overwhelming avalanche of literature to stay on top of. New techniques from field sample collection to lab processes to
Statistical analyses, uh, I think of this as not dissimilar from when GIS really enabled this boom in species distribution or ecological niche models early in the 2000s, uh, kind of similar enormous growth in something that’s being applied in conservation, fisheries, wildlife, natural resources.
We’ll keep it a little bit smaller than all of eDNA and kind of talk specifically about does eDNA work for crayfish and why or why not? And our first crayfish eDNA paper really came out in about 2014 out of France,
The same group that had, um, initially, been looking for bullfrogs as an invasive species in Western France, uh, who were looking for invasive red swamp crayfish. And they had relatively ambivalent success. In a series of farm ponds that had been trapped for red swamp crayfish they
Failed to get detections at some sites where they knew crayfish occurred. And in this kind of first of the eDNA papers attributed, Some of this difficulty of applying eDNA to crayfish has one difference that the, um, exoskeleton of, um, uh, our, our arthropods that, that by being in an exoskeleton crayfish
May just be less detectable than something like a bullfrog or a silver carp or, or organisms that might slough more skin cells into the environment. And this is kind of continued as sort of the default explanation when an eDNA assay doesn’t work as well for crayfish as we might hope it does.
You don’t get detections places, you’re pretty sure something occurs, and you say, well, I, I think that’s the exoskeleton. Right? Which is sort of a hard thing to evaluate. So, as much as anything, and kind of taking maybe a half hour, maybe 40
Minutes today to kind of talk about eDNA for crayfish, I’m going to kind of try to stitch together a couple of studies that try to get at, at least tangentially, where does our crayfish eDNA come from, and how should that inform our study designs for crayfish.
So the first of these was a product from a master’s student, Sam Garcia, who defended for my lab this preceding fall. Focused on, is it the exoskeleton? And this specifically used sampling pre and post molt of marbled crayfish, Procambarus virginalis, to look at how eDNA detectability changed as crayfish held in aquariums molted.
I’m going to relate this to the only paper I’ll cover today, a study from a PhD student, Amanda Curtis, who graduated back in spring of 2022, that looked at do crayfish carcasses produce detectable eDNA. And I think this kind of interfaces with the exoskeleton question, um, in ways
About where crayfish eDNA comes from. And I’ll kind of conclude scaling up to a kind of multi year whole landscape study for Minnesota. This is work of Chris Rounds, a student of Gretchen Hansen at the University of Minnesota Twin Cities, looking at optimal timing for crayfish sampling across many lakes.
Multiple years at five sampling events through the year from, from North Temperate Lakes. Uh, so we’ll start, um, with Sam’s work. This was using marbled crayfish, uh, Procambarus virginalis, to ask two questions. Does eDNA detectability increase after a molt? And our thought here around this question of the exoskeleton is, if that exoskeleton
Is really important, when you take it off, maybe your eDNA detectability as a crayfish increases when you’re in that kind of, uh, uh, gooey, soft shell stage. We were also curious if crayfish molts affected the particle size of environmental DNA in the environment.
And the idea there is that the molt itself is this kind of violent act. It’s potentially resulted in sloughing of more cells, more intracellular DNA, and maybe that would tilt the particle size distribution of eDNA to larger particles relative to smaller particles. Before the molt.
And we did this with, uh, uh, marbled crayfish. Here we’ve got a, uh, uh, kind of rough looking marbled crayfish that had an unhappy time in its previous colony tank before we, we got our, our hands on it. We did this experiment with juvenile marbled crayfish.
This is a bit larger of an adult than what we actually used. The, the thought here, these clonal, parthenogenic, identical marbled crayfish are really nice lab animal to work with, and we had these under an Illinois Department of Natural Resources invasive species permit. The juveniles are fast growing. They’re molting frequently.
We were hoping this would give us a lot of power to resolve this effect of molting on eDNA detectability for crayfish. So we did this with marbled crayfish we had sourced from a neurobiology research lab in state here in Illinois.
We collected a number of juvenile marbled crayfish of about the same age class. We housed these in small two and a half gallon aquariums. They had a PVC half pipe for shelter. They were aerated and filtered with simple foam biofilters. They were kept under LED lights on a 12,12 light dark sample.
Uh, they did have a little bit of unglazed tile, uh, in the tanks for, um, substrate. So the photos here, uh, don’t quite show the tile. This is a larger crayfish than we used, but this gives a sense of what these
Two and a half gallon aquariums where we’re keeping marbled crayfish look like. And we took 250 milliliter surface water samples at day one and day three after our marbled crayfish went into the aquariums. We sampled as soon as we saw that they had molted, and this was a team
Project, multiple grad students and an undergrad assistant in the lab. We’re dropping in and checking on our crayfish through the day. Of course, they molt at midnight. We’re going to catch that with our first water sample at maybe 8 a. m. the next morning.
After that molt, we sample on days 1, 2, 3, 7, and 10, after we found our juvenile marble crayfish had molted. We excluded our crayfish that died during the experiment, that most often meant that they had died during a molt. Uh, and we excluded crayfish that extruded eggs.
And, um, the thought there was other work, uh, the United Kingdom has found that crayfish really increase their, their eDNA concentrations when they’re buried, when they’re ovigerous. Um, we, we maybe, we could have kept these in the analysis and had a different treatment for egg bearing.
This happened a lot, that invasiveness of our marbled crayfish. These small, juvenile, fast growing marbled crayfish, many of them produced eggs, and so this really left us with only seven crayfish to work with. So, sample size is small, but with the sampling pre and post molt on
These intervals, we ended up with 56 total sampling events across seven crayfish, and each of those sampling events, to get at the particle size distribution question, we did sequential filtration through nested pore sizes, and this adapts from other particle size distribution work with eDNA.
We ran our water samples through a 5 micron filter first. Through a 1 micron filter second, from that water that had passed through the 5 micron filter, and then through a 0.2 micron filter last, that the same water sample was passed through three filters.
The idea here is to let us kind of parse intracellular versus extracellular DNA across the filter pore sizes. We applied an existing QPCR assay for marbled crayfish, which looked pretty good on our, our QPCR machines, NICE r squared, and, and efficiency values.
Each of those filters from each sampling event was run as three plate replicates or three technical replicates under QPCR. So it’s only seven crayfish, but we advanced to having over 450 replicated detections or non detections from our our eDNA samples. We ultimately had to analyze this data as detectability rather than eDNA copy
Number because we were infrequently above the limit of quantification for our crayfish in these tanks. It’s a little two and a half gallon aquarium, but it was not that full of crayfish eDNA. So we modeled these zero or one non detection or detections.
And a mixed model where we had day and tank as a fixed effect, the filter pore size as a fixed effect, and that molt status, uh, molted or not as a fixed effect. We limited molt as something, any sample within 48 hours of the molt happening.
We’re using some literature to suggest that crayfish have mostly remineralized their exoskeleton by about 72 hours after the molt. So you get a couple of samples before the molt happens. You get three samples from the molt through 48 hours, and then anything after that we’re treating as a non molted sample.
And these had random effects for those individual crayfish, as we know, we’re making repeated observations on, on these seven, uh, marbled crayfish, which, again, are clones or, or parthenogenic, maybe, maybe we have one crayfish in reality. We did some model competition with AIC. We had five most supported models and performed model averaging
Of the model parameters. Tiny tables that I pulled out of a paper and review at freshwater crayfish. We’ll look at that as a model predicted or modeled relationship of our covariates here. This is days in the tank on the x axis. We included days in the tank as a fixed effect.
Because they’re, they’re recirculating and not flow through tanks, and we expected that eDNA would probably accumulate in the tanks during the experiment. We have our predicted eDNA detection probabilities on, on the y axis. We have our pore sizes coded as the 5 micron pores in green,
The, uh, 1 micron pores in blue, and the 0.2 micron pores in red. And we have, uh, pre and post molt comparisons. Pre molt as a solid line, those molt samples as the dashed line. We’re not showing confidence intervals on the plot, just for kind of simplification,
To not have too many lines here. But I would say that pore size had our strongest and our only significant effect on crayfish eDNA detectability. DNA of our crayfish was most detectable at that largest 5 micron pore size in green.
We did have this interaction between days in the tank and pore size, where our smaller pore sizes, 1 and 2 and 0.2 microns, increased in detection probability more steeply than 5 the longer the experiment had been run. And we think what you’re kind of seeing here, where detectability
Really increases for the finer pore sizes, is that that intracellular DNA is breaking down to extracellular DNA as the experiment progresses. Um, and I will say here, you know, days in the tank from 0 to 12, studying molting was somewhat challenging, we can’t predict when the crayfish are going to molt.
Some jumped the gun and molted early, some molted fairly late after being in the aquarium. Um, but it does appear from our work, you know, that marbled crayfish eDNA was most detectable at our largest pore size. This was a significant effect. That eDNA seems to be probably intracellular.
Our eDNA detectability decreased post molt. The dashed lines color matched to each pore size when simultaneously controlling for these other factors in our, our aquariums. This is a weak effect, it was not significant, but it would suggest that there is little role in that exoskeleton in explaining crayfish EDNA detectability.
A hard, mineralized exoskeleton versus shedding it at a molt to kind of a gooey, soft shell body, not only doesn’t increase your EDNA detectability, but maybe slightly decreases it. We also see here that our eDNA accumulated over the duration of the experiment.
Uh, also a weak, non significant effect, but a choice here that we made to use recirculating rather than flow through tanks for some of the constraints in our lab space. This could be realistic for burrowing crayfish, working with burrow water
With a slow turnover rate that maybe EDNA is going to accumulate in something without a lot of maybe stream flow. But this is also something that you might do a better job of in a lab experiment with something like a flow through aquarium. So this was an interesting result for us.
You know, removing the exoskeleton at the molt did not improve our EDNA detectability for crayfish. Potentially with more statistical power, more crayfish, we might have found stronger evidence that might be a significant effect. I think at the journal we’ll probably hear questions about statistical power.
I would say that across 150 filters, up to 450 plate replicates, we have power to detect something like the pore size effect on our crayfish detectability. If we ratcheted that power up with more crayfish replicates, I think this molt effect is still going to be weak relative to other factors affecting crayfish detectability.
Um, but we do think this result is interesting. It contradicts what we thought might happen when crayfish molted. And I think one reason that would happen is that our crayfish generally cease feeding and reduce their activity prior to a molt. And it, it suggests that this little effect of a molt, and potentially
A negative effect of a molt on eDNA detectability, it suggests that where crayfish eDNA is coming from is, is DNA excreted with metabolic waste of the crayfish. They need to be feeding and excreting to be detectable. And if they’re not doing that, they become less detectable.
One thing that I also think is nice about this study, uh, is we did well at a 5 micron pore size for, for this crayfish. This would be helpful in many field contexts where people are often using 1.0 or, .45, or .20 micron filters for, for eDNA.
Those filter sizes can clog quickly in the field if a lake is eutrophic, if a river is turbid with a lot of suspended sediment, we might be able to sample larger volumes of water with higher pore sizes, and it doesn’t appear that there’s a cost of that pore size in crayfish detectability,
At least from this Lab experiment. So I kind of want to connect this to an existing published paper really briefly because I think this interfaces with a study that Amanda Curtis led on crayfish carcasses that we published back in, uh, 2020 in the journal PRJ.
Uh, and so this was a question about Do crayfish carcasses produce detectable environmental DNA? And, you know, potentially, if so, for how long? How, how long would you worry about this affecting decisions around management action for crayfish? So this could inform, uh, invasive species removal projects.
Potentially, if you’ve applied a chemical, uh, to a red swamp crayfish population in a small pond, if carcasses produce detectable eDNA, those type of, you know, false positives from a non living organism could result in costly management actions when, when the management intervention has actually worked.
And this happens, for example, in invasive brook trout systems in the Rocky Mountains. Uh, federal agency biologists have used eDNA to monitor for the outcomes of, say, rotenone treatments to remove brook trout. Positive detections have resulted in discovery of brook trout that survived the eradication effort.
Um, that’s great, that sensitivity of eDNA being really appealing, but if you’re getting a false positive after that chemical treatment from a carcass, it’s going to cost you money and time. You’d like eDNA to be associated with living organisms, ideally. On the opposite side of that coin, this can also inform
Native crayfish conservation. The reason I was really thinking about this was we had done work with the U.S. Fish and Wildlife Service in California where an ESA listed crayfish had been introduced to new sites, spring fed ponds, um, that were isolated
From an invasive crayfish, uh, and, and the managers didn’t know if that introduction had succeeded or not. Of this ESA listed crayfish, they were hesitant to intensively sample the site to cause any harm to the introduced crayfish populations. We got eDNA detections there, and the question was, is it possible
That those are carcass derived? And I said, well, I don’t know, but we can try that as an experiment. So, that’s what had motivated this carcass addition project in autumn of 2018. So we used Red Swamp crayfish, Prochamberis clarkii. These were sourced from Ruben Keller and his lab’s removal
Projects in the Chicago River. I think I saw Rachel Rogers on the call, people very familiar with these crayfish they provided to us. They were frozen in a minus 20 freezer and transported to Champaign. And we put Three carcasses each in an enclosure.
Our enclosures in this stream kind of enclosure experiment were just crayfish traps that we had sealed the ends of to prevent fish, water snakes, anything from getting in and consuming the carcasses. And that’s, that’s a choice that we made to kind of try to prolong the
Carcasses persistence in the system longer than might be realistic if something were to consume that carcass. When we put three carcasses in each crayfish trap, and each carcass was sealed in an individual mesh bag that had a pore size of the mesh of about 300 microns.
Large enough for cells and certainly DNA to pass through, but we were really trying to limit the decay of these crayfish carcasses to microbial processes and not necessarily to processing by invertebrate consumers. We had 10 total crayfish enclosures using 30 total crayfish.
Five of these enclosures were places that we wanted to sample eDNA from. Five were installed just so that we could monitor crayfish carcass decay rates in the stream. So, what we imagined was going to happen was that crayfish eDNA would be
Detectable at the start of the experiment, and at some point it would cease being detectable, and we wanted to know how much of that crayfish carcass had to break down, how much decayed before these crayfish ceased being detectable. We wanted to be able to kind of regress detectability against biomass likely
Remaining at that time in the experiment. So we did this in the saline branch of the Vermilion River. This is in a University of Illinois protected area, so no one could come steal our beloved crayfish carcasses, a site closed to the public, on a USGS flow gauge.
And as you can tell from the photo, this is about a third order stream. It’s a weightable stream, but fairly wide, um, training, uh, primarily agricultural watershed, and we’ll talk about that in a second. So we again took 250 milliliter water samples on the day of the installation.
We put crayfish out in the cages. We went home, showered, changed clothes to deal with contamination, and came back within less than two hours to water sample. And then we took those samples again on day 3, 7, 14, 21, and 28 days later. A month long experiment.
We took these samples as three from 10 meters upstream of the upstream most enclosure. These were controls to confirm that there was no red swamp crayfish eDNA in our stream. We had no detections of red swamp crayfish from this site, but our site
Was about two stream miles downstream of a wastewater treatment plant for the city of Urbana, Illinois. A bar has a crawfish boil. It’s not implausible we could see eDNA as effluent from wastewater treatment. So each sampling event, 1 through 28 days, we would take three samples
Above any of our cages by 10 meters. We then took one sample directly below each enclosure, a 250 milliliter sample where the water was flowing out of the cage, and then we went about 20 meters downstream and we took four water samples across the stream channel below these enclosure cages.
Below that sample, we had five additional cages that, as I kind of previously said, they were out there to measure the decay rates of our crayfish carcasses over time. We removed one cage and measured its three crayfish on day 3, a second on day 7, a third on day 14.
Another on day 21, another on day 28. And this is why we had crayfish replicated in each enclosure to relate their mass at the date of the removal to the measurements before they went into the channel. And this is what that looked like.
Here’s a crayfish on day 3 that’s just been popped out of a freezer, a little frosty. At day 3 of this experiment, our red swamp crayfish are still pretty intact. By day 7, the, uh, chela have kind of come detached.
By day 14, we’ve got some crayfish goo, but there’s still claws there, and that’s what our crayfish really look like as we go forward to day 21 and 28. The, the claws in particular tended to persist over time. A lot of this is It’s exoskeleton.
It’s not really muscle or organ tissue remaining, and again, this is a really generous version of what crayfish decomposition might look like. Because we held out consumers by sealing these enclosures, fish cannot enter the crayfish traps, uh, and we also excluded invertebrate consumers, uh, by having the crayfish in mesh bags.
This is really limiting our crayfish decay to microbial decay. Just a little snapshot of what this stream looked like. In panel A, we have a hydrograph of our discharge over time. The experiment ran from a month from about mid September to mid October of 2018. Here, discharge is in liters per second.
I think that’s something a journal reviewer wanted to relate our sample volumes to, although CFS, or cubic meters per second, might be more intuitive. Uh, we did this experiment at base flow conditions, we had a tiny rain event in
Late September, we had a bigger rain event in October, but we lost no enclosures through that, that higher flood pulse. You know, we were primarily at base flow conditions. The bottom panel, uh, is a temperature logger that we put in the stream site during the experiment.
We started pretty warm, uh, temperatures wobbling up and down through the day around 23 C, declining into October to about 15 C after that, that rain event. And this is a, a basic stream, our, our, our pHs are around 8. 5. Conductivity is high on sedimentary rock in Illinois, but also from
Agricultural land use in the watershed. Turbidity was generally very low with the exceptions of our sampling at these flood events where we had some suspended sediment. Crayfish decayed, our top panel here is the percent loss in biomass of our, our decay crayfish that were removed at day 3, 7, 14, 21, 28.
This kind of reached a plateau where the crayfish had lost about 60% of their biomass. But, but held there, this kind of hard to break down carapace material sticking around in the bags up to a month after the crayfish have been put out in the stream.
Uh, the bottom panel is our actual eDNA crayfish. They’re, they’re, uh, weights at the start of the experiment versus the end of the experiment. Again, something like 60 to 70% loss of our crayfish, uh, after a month in the stream.
And so, you know, I, I want to Be clear that we feel fairly confident that this is a good eDNA assay and that environmental DNA was detectable at our stream site. So we filtered our samples through 1 micron cellulose nitrate filters.
We store in a lysis temperature, a lysis buffer at room temperature, CTAB. We let the cells lyse for a few weeks, we go into a freezer, we extract with a phenol chloroform isoamyl approach. Then our internal lab test is really resistant to PCR inhibitors. Maybe not universally resistant, but something where we’ve
Generally not seen PCR inhibition. Something where maybe your stream or lake’s water chemistry results in PCR not working very well. It’s something we haven’t really encountered with this method and in these stream sites. We used the Red Swamp crayfish assay from the group in France from 2014.
We reoptimize it for our QPCR machines. We had very good PCRR squared and efficiency values. We have a sensitive LOD, uh, a little bit before, um, some alternative methods to, to calculating quantification and, and detection limits. Uh, but in general, th this assay looked really good to us.
And, um, some other checks that, that we did on have we done things, well, here in the panel, we have, as the, the black, uh, box plots, our serial dilution of known concentrations of synthetic red swamp crayfish DNA. We do this on every well plate from a really high concentration, DNA copy
Numbers over a million per microliter, to really low concentrations, maybe one or ten estimated copies per microliter of red swamp crayfish DNA. We run that as a positive control serial dilution series on every plate to know if the assay is working. We use those positive control serial dilutions to estimate our
R squareds and our efficiencies and our limits of detection. And one thing we can do if you worry about PCR inhibition affecting detectability is you can spike in, in your field samples DNA and compare it against that DNA in ultra pure lab water and see if it amplifies later our CQ or
Amplification cycle on the y axis as evidence of inhibition in the sample. So in this case, in our gray boxes, we spiked in red swamp crayfish DNA into our lab samples, lab samples where it had not amplified, compared to our
Positive control in the plates, and we were looking for if this would amplify later than in our positive controls. And, and that did not happen. There’s no later amplification here. So we suspect that inhibition isn’t a factor for us. We get some amplifications from those spike ends that are a little earlier, that
We think are maybe just a little pipetting error, but we don’t see any delayed amplification of known DNA concentrations in our actual field samples. You’ll notice I’m kind of responding to some reviewer questions on, on this paper here that has been out for a little while.
Similarly, people said, are you sure you’re doing things right? Can you find DNA in the site in general? We were doing concurrent studies with corbicula clams at this site. We detected corbicula eDNA at a location that had lower densities for us than many sites we worked with.
We re ran these same samples for corbicula eDNA, and again, had no problems detecting eDNA out of these same water samples. So that’s a lot of preamble to say from this carcass experiment, we detected no red swamp crayfish eDNA from any of these samples in our field enclosure experiment.
Samples taken an hour and a half after the cages went into the stream directly downstream of crayfish carcasses, we had no detections. So all of that previous kind of method slide, are we sure we didn’t do something wrong? Is the assay any good? Are we inhibited? Can we find eDNA at that site?
All of that kind of sets up our belief that these crayfish carcasses, at least at the biomass of this, this fairly large stream, maybe a third order stream, were not detectable at any time they were in our stream enclosures.
And so I kind of present this here, paired with Sam Garcia’s work on our molting. It seems that crayfish eDNA is strongly associated with excreted metabolic wastes. You need live, actively foraging, excreting crayfish to detect them, and the carcasses in this case were not detectable in this stream.
And I, I think there’s a number pieces of good news there. You know, if the carcasses are less detectable than live crayfish, and that would be a different experiment to run, you know, in hindsight, this isn’t what we thought was going to happen.
You could do this in a mesocosm or a lab, contrast live versus dead crayfish of similar biomass. But it seems that if the crayfish carcasses are not very detectable by eDNA, that reduces a number of possible concerns you might have. For example, that crayfish DNA might be excreted by a fish
Predator someplace where the crayfish weren’t actually occurring. Uh, that itself would be a great experiment to run, right, to feed juvenile crayfish to a largemouth bass, move it to a new tank, and then sample for eDNA to see if predator excreted DNA could move crayfish DNA someplace.
There’s other things you might do differently than what we did. Did freezing the crayfish matter for eDNA detections? We wouldn’t think so. These crayfish were frozen in Chicago, transported to Champaign, they came out of the freezer to go into the stream. You know, maybe that’s affecting eDNA detectability. I wouldn’t think so.
Maybe we shouldn’t have enclosed the crayfish in these mesh bags to kind of track decay rates. I think if we did that, these carcasses would be gone much faster than in this experiment, and I suspect that it would point to less support for crayfish carcasses producing detectable eDNA.
So I’ll kind of wrap up here. We’ve gone from a little lab aquarium experiment, to a mesocosm enclosure experiment, to something very large scale across the state of Minnesota. And this was a big project with multiple PIs, Chan Lan Chun and Josh Dumke of
The Natural Resources Research Institute at the University of Minnesota Duluth, Gretchen Hansen at the University of Minnesota in the Twin Cities and my lab, who are kind of trying to resolve for a state like Minnesota. That may have 10, 000 plus lakes you’d like to monitor for invasive species with eDNA.
How do you do that? How do you design an eDNA sampling program, uh, that’s going to be effective? And, and one concern there is multispecies communities. Part of the appeal of eDNA is that you could monitor multispecies communities simultaneously. The same water sample could detect spiny water flea, or common carp,
Or zebra mussel, or rusty crayfish. When conventional sampling might need to apply really different methods for that set of four organisms. Some of the time and cost savings that people associate with the appeal of eDNA is that you can detect everything from the same approach up
To potentially hundreds of species that might vary a lot by the conventional methods required for their sampling. At issue with that is that eDNA may not be equally detectable to those species, especially over time. That as those species differ by life history or phenology of their activity,
The optimal timing for sampling is something you’re going to want to solve if you want to sample lakes across Minnesota for all of your priority invasive species. How much effort do you need at what time of the year? So that’s our question, and we had three graduate students work on this.
Chris Rounds at the University of Minnesota, Anna Tosh in Duluth, and Sam Garcia here at the University of Illinois. We chose four invaders. Spiny water flea, common carp, zebra mussel, and rusty crayfish. Because we expected that these species would differ in their eDNA detectability through the year.
We sampled 20 lakes over two years, some in 2021, some in 2022. We sampled those lakes five times through the year, from ice off to ice on. We took 10 water samples per lake and filtered those and ran three qPCR replicates for each of those field samples for each of the four species.
We did too much and that, uh, uh, our grad students here, uh, certainly took the brunt of, of the ambition of this project. And our bottom panel, we’ve got kind of these predictions from, uh, uh, Chris’s paper currently in review at Ecological Applications.
We expected from past work that spiny water flea was going to be most detectable in late summer. We expected common carp to be most detectable in the spring during spawning aggregations. We expected zebra mussel veligers in the water column to be detectable in mid
Summer and we thought from some of my past thoughts on rusty crayfish synchronous molting in populations from inactive to active forms of, uh, reproduction that maybe there’d be something like a little bimodal, uh, early summer, late summer peak, and rusty crayfish eDNA.
If you recall our, our, our molting results early in the paper, uh, maybe not so much, right. So, I’ll jump ahead. Chris did Bayesian hierarchical occupancy modeling. Uh, on this enormous data set, here’s his results. We have detection probabilities on a y axis.
We have Julian day, or calendar day of the year, on an x axis. These models include other things. Is the lake mixed or not? What’s the lake clarity? But we have strong effects of seasonality on, on these species, and we’re going to focus on those here.
Zebra mussel, in, in pink or purple, was by far our most detectable species, spiny water flea, in blue, was by far our least detectable, and common carp and rusty crayfish were intermediate. But the timing of our optimal sampling also varied by species.
As we had predicted, common CARP eDNA detectability was highest in the spring and declined through the year. Our other species had optimal sampling in summer to early autumn with just slight differences in kind of timing of that peak detectability of their eDNA.
But our, our rusty crayfish, to keep this a little crayfish centered, were most detectable in a kind of late summer to early autumn time window, consistent with warm lake water temperatures and high metabolic activity of these crayfish. Our crayfish were also more detectable than common carp in late summer.
The carp are more detectable in the spring in the same set of lakes, the crayfish are more detectable in the summer. So as much as anything, I might say that when we write eDNA papers about crayfish, we don’t necessarily have to say that crayfish are less detectable than fish.
I don’t think that’s true, and we don’t necessarily have to attribute that to exoskeletons. I, I think, um, I, I think our, our crayfish do pretty well. They’re, they’re not zebra mussels with a pelagic velliger that’s easy to pick up, but they’re more detectable even than, than common carp in late
Summer in north temperate lakes. Chris inverts this into the number of samples you need for a 95% probability of detecting each species when present. Zebra mussels, trivially easy to detect by eDNA. Uh, spiny water fleas, prohibitively hard to detect in these lakes, requiring something like 150 field samples, not plate replicates for detections,
Even at their most detectable. Um, this also highlights that our, our optimal sampling for multispecies communities is going to be more challenging than if you’re sampling for a single focal species. And I would especially connect that to people considering eDNA metabarcoding for whole communities.
Metabarcoding is less sensitive to species presence and especially low abundance in single species QPCR for a number of, of reasons in the metabarcoding process. So when doing metabarcoding for a whole community and getting false negatives where you don’t find something that you know is there, that, that may be as much
About timing of the sampling as it is about something in your lab methods. And, you know, we, we did tinker with some of these lab methods. Our Minnesota project stored samples in ethanol. It used a QIAGEN kit for, for our environmental DNA extractions.
You could probably improve on some of these results by using other eDNA capture, storage, or extraction methods. For spiny water flea, for example, we tested a subset of these lakes and samples for storage and extraction approaches. Ethanol stored samples extracted by a QIAGEN kit in blue performed
Worse across all three species than storage in a lysis buffer, CTAB, and extraction by phenyl chloroform isoamyl. That’s been true in eDNA studies, um, since the earliest applications to invasive species in the U.S., but in this case, spiny water flea was detectable in two of our five lakes.
By the CTAB PCI method, where it was undetectable by QIAGEN, you could probably get to better, more realistic field sampling effort to detect spiny water flea, but you’re still going to retain some of this heterogeneity and detectability by species. Even with CTAB PCI, zebra mussel was our most detectable species, rusty
Crayfish were second, and, and spiny water flea, we’re a distant third. And this is going to inform if you’re doing multi community monitoring with eDNA. Your hardest to detect species in its most detectable season is what’s going to determine your sampling design and, and effort.
So that kind of sprawled out of crayfish, but it’s kind of a summary of these three results. For crayfish eDNA, how much does the exoskeleton matter? Well, if you take it off for a molt, and, and you’re, you’re in kind of a soft bodied state, uh, that really doesn’t increase eDNA detectability
And seemingly decreases it, potentially due to that reduction in foraging and, and mobility of the crayfish. Do we think that eDNA is mostly being excreted with wastes? The CARCA study would seem to suggest so. Decent biomasses of crayfish in a third order stream and enclosures weren’t detectable when the crayfish were dead.
You could come up with a more clever way of, of testing that. We, uh, we did not expect this result, but knowing what we know, I think that’s, that’s something interesting to follow up on. Given that, I think it’s at least an encouraging result for, for rusty crayfish.
Some of our organisms are easier to detect in the winter. Some are easy to detect at, at ice off in the spring in these Minnesota lakes. But if you’re looking for, for rusty crayfish or something similar in the Great Lakes region, you’re going to do okay going out to eDNA
Sample when you would crayfish trap. July, August, September. That’s going to be your, your best time to sample for, for most of our crayfish. Uh, so I’m happy to, to take questions. This, this work, um, has been supported by a USDA hatch project to, to my lab and the big Minnesota project.
We, we owe, um, uh, a lot of thanks to funding from the state of, of Minnesota, uh, administered through the Minnesota Aquatic Invasive Species Research, uh, consortium. And, uh, with that, I’m, I’m happy to, to take questions for, for maybe 10 minutes. (Natalia) Great, thank you so much, Eric.
Um, we currently don’t have any questions in the Q& A box or in the chat. Um, but I actually have a few. Oh, nevermind. One just came in. Um, so we got somebody from the chat asking, I would argue that your carcass experiments support the notion that exoskeletons influence eDNA detectability
Via lack of slowing tissue because tissue is stuck within exoskeleton. DNA extractions involving insects, not eDNA, supports this as well. Thoughts on this? (Eric) I, you know, I think as the crayfish broke down, we still had some exoskeleton remaining, but a lot of that soft tissue, uh, is kind of
Disarticulated out of the, the carapace. And, um, I mean, it’s hard, it’s a bit hard to say, you know, if you put a fish in an exoskeleton and a crayfish out, is it the exoskeleton at the same biomass that, that affects our, our eDNA detections?
But I, I mean, I think from the molting experiment, you know, you, you molt and, uh, your eDNA detectability goes down instead of up when you’re in your soft bodied form. I don’t know that the hard exoskeleton is really what’s affecting our crayfish detections as much as excreted metabolic wastes. (Natalia) Great.
Thank you so much. Um, sorry if you already mentioned this, but how long were the Chicago River red swamp crayfish frozen for, um, before the carcass tests and before they were placed in the test stream? (Eric) Yeah, we’d have to check. Rachel might know. I mean, I think it was a few weeks.
We had tried to time close to a removal project from the Loyola group with, with Ruben’s lab and they, they had not been in a freezer for months or years. They’ve been collected that summer before our, our, um, our fall enclosure experiment. (Natalia) Okay. Okay.
Cause I was thinking, yeah, that it may be as possible that freezing samples for a certain time might reduce detectability, but. (Eric) Yeah, I mean, you know, if, if we knew what was going to happen with the power of hindsight, we may have done something in the lab rather
Than an enclosure and, and worked from, you know, here’s a crayfish, we can get its, uh, uh, detectability or eDNA concentrations in the aquarium. We may have killed it and then done the same in a fresh aquarium from the carcass and said, what’s the drop in eDNA that’s related to, to
Being alive for the same individual? Uh, no freezing and, and have a little more control in the lab. I, I think would be a, a really good thing to, to do. (Natalia) Yup, um, have you tested using the five micron filters in the field? If so, did you still get good detectability?
(Eric) I’ve not, but I, I think we should. And, um, I’ve reviewed as a journal reviewer papers that have gone up to fairly large pore sizes for, um, say, amphibians in South Texas in really high turbidity environments. And they’ve had really good eDNA detection success.
And I, I think sometimes we kind of default to, hey, use a tiny pore size. Uh, you want to, you want to capture as much DNA as you can, and it has this trade off of how much water you can filter if you are filtering, and there’s eDNA methods that don’t.
Um, but I, I think it would be worth experimenting with, ee-especially for people interested in eDNA for crayfish in turbid or eutrophic environments. Potentially burrowing crayfish, working with burrow water I, I think it could be really worthwhile. Hey here’s water siphoned from a burrow, let’s try a really large versus a really
Small pore size and, and see how we do. I, I, I think we could go up in pore size quite a bit for, for crayfish applications. (Natalia) Great, thank you. Um, next question asks, in your slide regarding the best timing for sampling, your graph goes May to October with zebra mussels peaking midway.
I’m curious if this would imply you would be less likely to detect zebra mussels in the winter? (Eric) Yeah, you know, we, we didn’t try. We, we chose, um, to sample, uh, ice off to ice on, on these Minnesota lakes. And, uh, I, I think, I think that was certainly a hardship for
Some of our, our grad students and, uh, uh, field technicians. We, we didn’t do any sampling through the ice itself. And, um, the way our, our zebra mussel detections in these Minnesota lakes really attenuates on, on the shoulder of the warm season, both in
Spring and fall, I think they would be pretty undetectable in the winter. But I, I am assisting, uh, some colleagues to, uh, sample through the ice for turtles in Minnesota at the end of this month. We’re going to go out with an ice auger.
And, uh, we, we will have a spring to fall to winter comparison for, uh, two turtle species. And that, that will be my first experiment with, uh, through ice eDNA sampling. And, uh I’m, uh excited and anxious about it as someone who grew up on the Gulf Coast of Texas. Yeah.
(Natalia) That’s so fascinating. (Natalia) Um, is there a way to synthetically produce activity or mobility in crayfish to promote detectability? (Eric) You know, I, I do think I, I think it could be really nice as a lab experiment to potentially with lab crayfish run a ramp of low to optimal
To too warm temperatures and look at eating a concentration and detection on, um, the temperature gradient. Something that I think Jacob Westhoff had, had been on the call earlier and Jacob’s done a lot of crayfish temperature work that, that I think, um, uh, you, you
Could kind of get at a temperature link to eDNA production for crayfish through, through lab experiments on, on, uh, uh, you know, uh, aquariums or mesocosms with, with manipulated temperatures. (Natalia) Great, thank you. Um, is there, do you think, a method for measuring where the source population
Is after finding a positive detection, especially, um, in, like, a lotic system? (Eric) Yeah, that, that’s a great question. You know, especially that eDNA transport distance question. And I, I think, um, I’ve always felt a little overwhelmed by that. Um, in part where we’ve worked in streams or rivers for things like corbicula
Clams, we, we tend to assume there’s just a lot of corbicula upstream, and we’re not trying to pinpoint, hey, here’s a source population early in an invasion, and it’s 700 meters upstream based on this eDNA concentration diluted to this river discharge. People are doing that often in things like anadromous salmon systems, where
If you’ve got kind of a, um, a post arrival of, of your fish within a narrow time window, and, and maybe there’s a weir or a fish counting station and you have estimates of salmon biomass entering kind of a controlled area. People have been able to relate environmental DNA with more accuracy
To relative abundance in a known area in stream and river systems in, in situations like that. You know, the, the transport piece, uh, eDNA may move 100 meters to a kilometer, but the transport distance of eDNA may be affected by stream size and discharge.
EDNA may move farther the higher discharge a lotic site is. And, um, I think I think it’s a little tricky to attribute the source for, for an EDA detection to, to upstream. Um, uh, people who are smarter than I am are, are working on it though. Yeah.
(Natalia) Sounds like a lot of variables to take into consideration for that. (Eric) Yeah I think there’s interesting ideas there about, um, spiking in known amounts of eDNA, uh, that that doesn’t occur in a stream as a, a tracer, uh, much like nutrient tracer and nutrient spiraling
Experiments of, of the 1980s or 1990s. And, um, eDNA is a really, uh, heterogeneous molecule. It’s intracellular, it’s extracellular, it does some funky things, but I, I think you could, you can inject novel eDNA into a stream from a source and kind of approach
It as nutrient spiraling, like nutrient spiraling for a nitrogen molecule, to get at some transport questions. And I think a number of people, uh, working on that transport question are, are gonna do really cool stuff with that, uh, analog of eDNA travel to nutrient spiraling work and kind of field experiments.
I, that’s something I expect to see in, in the next few years as something that’ll really inform those transport distances. (Natalia) Yeah, that’d be really exciting. Um, so a few more questions left before we, um, end the webinar, uh, but there’s one that asks.
Regarding potential tank experiments with crayfish carcasses, if not euthanizing through freezing, would you have to worry about other methods, in example, chemical, um, interfering with the eDNA sampling slash analysis process? (Eric) Yeah, totally. I, I, I don’t personally know if something like MS 222 or, uh, uh, clove oil would
Potentially work as a PCR inhibitor. Uh, so, so some of the chemicals that could, you know, inhibit, uh, our, our PCR reaction, people worry about acids, humic acids, tannins, really stained or brown water streams, PCR inhibition sometimes comes up. But it can also be things like, uh, calcium or, or magnesium.
Uh, you know. There, there, are, are probably more blunt mechanisms to kill a crayfish for that type of experiment than MS 222. Um, you know, we, we didn’t necessarily return to it. I think internally our, our labs curiosity on crayfish carcasses was satisfied, and
I think the hallway of my, uh, laboratory would prefer that we have fewer 28 day old crayfish carcasses in, uh, in my building. Uh, but I, I, I think there’s things to, um, come back to, uh, around, you know, living versus dead and temperature effects on, on EDNA detectability for, for crayfish.
Yeah, that, that I think would be really amenable to, to fairly simple lab experiments and often lab experiments would, would match well for like undergrad, senior capstone kind of projects. (Natalia) Great, thank you. Um, we’ll do one last question.
Um, has anyone used, uh, an eDNA assay or can eDNA play a role in detecting invasive crayfish within the organism and trade pathway? Um, so those coming from the pet trade or, um, from the live bait trade or accident accidental release by fish hatcheries?
(Eric) Yeah, I think that’s really interesting, um, you know, that, that, um, that possibility to screen the bait or, or pet or fish stocking trade through eDNA. I think, in our case, where we’ve, in, in Minnesota, for example, we did single species qPCR for four species, and it’s from the same water samples.
But, but each time we, we run a plate, we’re doing it separately for each species. As, as you approach four or five or six species of interest, it, it really becomes more time and cost effective to metabarcode rather than to do single species, um, qPCR.
So I, I, I think there’s promise there that you know, here is an aquarium shipment or, or maybe a, a bait supply pond at a supplier, take water samples, metabarcode, and compile a list of possible species of concern. Um, you know, I think it would be more of a metabarcoding
Than a single species question. I, I think it runs into, I think it will run into resistance of, say, bait suppliers or pet suppliers about screening for hundreds of species simultaneously. Um, what a detection potentially means, especially for organisms that might interface with things, um, like state or federally prohibited species, and,
And how, as people still worry about a risk of false positives, a bird deposited eDNA for silver carp into my bait pond and now I’m going to be closed and can’t sell to a bait shop. I, I, I think there’s, there’s both promise and, and things to, to
Resolve on, on how that would work. I, I think it would have to be metabarcoding, and I, I think it would, um, you know, it’s something that as it goes from a, a research question to a policy question needs a lot of thought about, um, confidence in, in
Implementation, uh, and, and, and how that, kind of relates to the the trades. Yeah. (Natalia) Yeah, those are very valid points. Definitely pros and cons to to doing that. Um, okay. So we are already a little past time. Thank you so much for everyone joining us today.
If you have any other questions regarding today’s talk, you can email me your questions and then I’ll make sure to forward it over to, to Eric and then, um, hopefully you guys can can discuss it then. Um, but thank you again, Eric, so much for giving us your time.
And presenting the very important work that you and your lab and your collaborators are conducting. Um, and thanks again to our attendees for joining. Um, I’ll see you guys at the next webinar. Thank you so much. (Eric) Yep, thanks.