Bird DNA in water

We just finished analyzing ~100 stream eDNA samples for Nebraska DEQ. We ran these samples with 12S primers that amplify vertebrate DNA.

The goal of the project was to look for DNA that might be helpful in determining the sources of fecal coliform bacteria in the water. This part was pretty successful, but what was almost more interesting was the list of bird species whose DNA we found.

wood duck
rock dove
mourning dove
barn swallow
red winged blackbird
melodious blackbird
brown headed cowbird
gray catbird
house sparrow
swainson’s thrush

It’s really an amazing collection of birds that pretty well represents riparian bird assemblages. The chicken DNA is likely coming from agricultural operations, but that remains to be tested..

What’s fascinating is really how much bird DNA is in the water. First obvious question is how it got there. Defecation? And if bird DNA is in there, are their disease-causing organisms in there too? Also, could this tool be used for conservation purposes?

Seems likely.


Sequencing the Potomac River

The Potomac River starts small, at a place called Fairfax Stone. 300 miles later–as it passes by Washington DC and enters the Chesapeake Bay estuary, it’s very large.

Along the way the Potomac does more than pick up water. It picks up nutrients from the surrounding lands and the species that inhabit its waters and banks shift, too.

How much nutrient it picks up and how the species in its waters respond has never been worked out that well though. There have been surveys of rivers like the Potomac. People have looked for insects and fish at different spots. But, as the river gets big, it’s hard to use standard techniques to see what’s in the water.

Jonah Ventures, along with friends at the University of Maryland, tried to see if we could use environmental DNA to put together a biodiversity atlas of the river and whether we could put together a bio assessment index based on the relative abundance of the DNA of bacteria, phytoplankton, macroinvertebrates, and vertebrates along the river.


We just published our results up on the pre-print server Biorxiv (see here). Although the research is still undergoing peer review, we feel pretty confident that the results are pretty amazing (at least for a first shot).

The take home point of the paper is that the assemblages of the headwaters, main river, and estuary were distinct. Not only that, we found a good relationship between increases in phosphorus concentrations in the river and the relative abundance of different species. As the waters picked up phosphorus, a whole suite of bacteria, phytoplankton, and animals became more abundant as another suite dropped out.

By no means is the last word on the question. We still have to develop our techniques a bit more, but it’s hard not to see the potential here. Soon, we should be able to suck up a small amount of water from a stream or river and quantify enough of the organisms that are in the water to infer the status of the waters.

*Bonus fact not discussed to much in the paper. In some parts of the river, almost half of the vertebrate DNA was human DNA. We thought the most extreme parts would be down by DC. Not so. All the sites with a lot of human DNA were actually up in the headwaters, not down by DC.

**Bonus fact not in the paper. We found a lot of different species in the water that we didn’t discuss. The list of species we found was fascinating. Fish like trout, carp, eels and shad. Amphibians like red salamanders and spring peepers. Mammals like pigs, cattle, and fox. The coolest animal: mountain lion (!). At site 5, 2.5% of the mammal DNA was identified as Puma concolor. There are no cats with sequences close to that.







Learning more about diet

Some of our scientific ignorances are so vast we do not even realize it.

What animals eat is one of them.

At Jonah Ventures, we’ve analyzed the diets of over 25 different species and every time we do, we learn something entirely new. But, continuing to learn new things when we analyze the diets of species we’ve measured before, is the best marker of the depth of our ignorance.

Mary Miller at The Nature Conservancy’s Ordway Prairie in South Dakota has been collecting fecal samples for a few years now. The other day, we just ran ~2 more years of samples for her.


Just looking at the three taxa in the diet over the two years reveals insights that we really have not seen before. Here, samples from 2015 are in red and 2016 are in blue.

The most abundant species–a cool-season grass—shows how regularly the bison consume a given species in different years. The proportion of Trisetum (likely spicatum) in the diet follows the same pattern both years: peaking in spring (early May), declining to a minimum in late June, and then peaking again in the fall (mid September). The second most abundant taxa (Poa pratensis and similar species) shows similarities between the two years, but it was consumed a lot more in 2015 than 2016. Some years, it seems that the bison are consuming more of some species than in other years. What is driving this interannual variation just isn’t known. The third most abundant species (a Lotus—bird’s foot trefoil) demonstrates another aspect of bison diet: bison regularly switch the composition of their diet within a year. As spring turns to summer, the bison are switching from cool-season grasses for most of their protein to a legume.

In a previous paper, looking at multiple years of bison diet in Kansas, we saw the same patterns. Consistent shifts among taxa in the diet bison and interannual variation in the reliance on different taxa.

We really don’t have theory to explain the shifts that we’re seeing. Are the herbivores simply following phenological shifts in the plants? We also have yet to incorporate it into our management of herbivores. For example, no bison managers (that we know of) are currently working to promote legumes in their grasslands. In fact, most try to promote grasses.

Hopefully, as more people collect data like these we can produce more theory and then begin to incorporate into our management of wildlife.


Parasites in animals

A lot of things come out with fecals. Not only is their DNA from animals ate and the animals themselves, but also DNA from parasites that infect the animals. Common parasites in the digestive system include nematodes, trematodes, and cestodes. Traditional techniques for parasite assessment have been to look for eggs in the fecals. But this work requires experts (which can mean expensive, slow, inexact) and cannot always differentiate closely related species. Sequencing DNA in fecals also has the potential to identify parasites. For one client, we used 18S sequencing to look for parasites in their birds. To begin, voucher specimens sequenced well. When we looked at DNA in fecals, animals identified as having parasites visually also tested positive using DNA sequencing. One thing we did learn is that for 95% accuracy in testing for the presence of parasites, multiple replicates need to be run (using our standard approach of swabbing the fecal). Still, for those interested in whether their animals might be infected with parasites in the digestive system, this is a simple test to add.

Fish environmental DNA

The technique of assessing the relative abundance of different organisms is one of the strengths of next generation sequencing. For one recent project, we assessed the relative abundance of different fish species across a number of streams in Nebraska by amplifying a region of the mitochondria that is more or less specific to fish. When the percentage data are aggregated, you get a graph something like below. We still have some more work to do to see how well the relative abundance of different fish species is represented by the data below, but the technique identifies a number of the more prevalent taxa found in the streams. We should be introducing a new product for fish eDNA pretty soon as we work out the final details. Stay tuned.


Sampling organisms of Mississippi River

Jonah Ventures is currently working with Cameron Thrash at Louisiana State University to quantify the organisms of the Mississippi River using eDNA. The project is an amazing effort on the part of volunteers that have rowed the entire length of the Mississippi, collecting water samples as they go. Our contribution is to analyze the DNA in the samples for higher organisms such as phytoplankton, insects, and fish.

Something like this has never been done before. Our preliminary results have been encouraging. Analyzing the DNA in the samples, we’ve been able to reconstruct the phytoplankton community along 3000 km of river. Below is the abundance of just one taxa, Skeletonema marinoi (and/or related species), which is a diatom that apparently is more abundant in the larger portions of the river, especially below the confluence with the Missouri.


Skeletonema looks something like this:


Normally, if one wanted to determine its abundance, one would have to take a water sample, filter it, and then examine the contents under a light microscope, counting each shape that resembled something like the SEM above.

Using eDNA, we can do the same counting likely better (because we’re also counting all the green algae and cyanobacteria as we go) and a lot faster.

We’re working on publishing the phytoplankton data and should be acquiring data on insects and fish soon.





Phytoplankton–organisms such as cyanobacteria, green algae, and diatoms–are good indicators of water quality. The problem is that they are tiny and you need a microscope–and a lot of training–to identify them.

Good news is that all of the organisms have unique DNA and are pretty abundant in waters, which means we can sequence them.

We’re still in testing phase for phytoplankton, but are settling in on an approach that will rapidly allow us to quantify phytoplankton assemblages.

Using a plastid marker, we ran some tests on assorted samples we had collected. Some were from a reservoir in Colorado, a few were from a backyard koi point, and others were from rivers and ponds in Kansas.

The plastid marker we were using separated out the waters well.

CO reservoirs are ID’d by their Cyanobacteria. My koi pond by green algae and dinoflagellates. KS waters by their diatoms.


We’ve tested these out with other streams and feel they are working pretty well here, too.

We still have more testing to do, but we’ll likely be ready to offer phytoplankton testing soon.


Helping tribes with bison

Jonah Ventures is proud to announce that we will be assisting Sitting Bull College in research on bison diet and performance. The research, funded by the USDA, represents a collaboration among three Tribal Colleges: Little Big Horn (Montana, Crow Nation), Sinte Gleska (South Dakota, Lakota Nation), and Sitting Bull (North Dakota, Lakota/Dakota Nation). The goal of the research is to gain new knowledge of how to strengthen the health of bison and be able to care for them under new conditions. Jonah Ventures will assist in a number of aspects of the research, including analyses of bison diet and performance. Stay tuned for updates on this exciting project.

Vertebrate DNA in water

I remember the famous limnologist G. Evelyn Hutchinson was once mocked as supposedly believing that he could determine all there was to know about a lake simply by sticking one, perhaps two, fingers in a lake.

I am not sure how many fingers it would take, but with environmental DNA one can convince themselves that we are getting pretty close with just a few fingers worth of water.

Almost everything that touches a stream or lake should leave some DNA behind. We tested that a bit with some filtered stream samples from Nebraska taken by NE DEQ. Across a wide range of streams in eastern Nebraska, we amplified vertebrate DNA (here fishes and mammals) to see what animals were leave  DNA signatures behind.

Turns out a lot.

On the mammal side, a lot of human DNA, which might be from handling the samples, but also cattle, skunk, muskrat, beaver, deer and mice.

On the fish side, we found about 20 species of fish, including long nose gar, common carp, silver carp (!), fathead minnow, quillback, longnose sucker, and spotfin shiner.

Keys  to using eDNA for vertebrates is a combination of choosing the right primer and getting enough sequencing depth to capture as much of the assemblage as possible. These data were pulled from ~300,000 reads across ~25 sites. Doing it right will likely take at least an order of magnitude more sequencing depth.

Still, it’s encouraging to see so many species in our data.

We’re continuing to work on the issue, so stay tuned…



Cattle diet across US

One power of reconstructing diets with sequencing is to know what an animal had recently ate. Even more powerful is reconstructing diets for a large number of animals to infer how diets change over space or time.

In the global change world, it is an open question about how warming will affect the diets of animals. To test this, we worked with Texas A&M’s GANLab to sequence the fecals of cattle across the central US.

Geographic trends in individual OTUs were evident. For example, species like the cool-season grass Bromus were dominant in the northern grasslands.


When we put all the data together, it was clear that northern cattle relied more on grasses than southern cattle. Cattle in southern, warmer sites typically consumed a greater proportion of forbs and woody species in their diet.

The paper also showed that certain species were indicative of low-quality diets, which could benefit ranchers in improving the nutrition of their cattle.

The inference from this work is that warming favors consumption of non-grasses. Why this is was beyond the scope of this particular project, but even being able to reconstruct the diet of any species across such a large geographic gradient is a major advance in and of itself.