Saturday, April 18, 2020

Machine learning helps scientists distinguish ancient human, dog poop

APRIL 17, 2020 / By Brooks Hays


Scientists have developed a machine learning algorithm to help archaeologists distinguish between ancient human and dog coprolites, fossilized poop. Photo by Jada Ko/Courtesy of the Anhui Provincial Institute of Cultural Relics and Archaeology
April 17 (UPI) -- Researchers have developed a new machine learning algorithm that can determine whether ancient excrement was deposited by a human or a dog.

Bones and artifacts are great, but ancient poop can offer archaeologists tremendous insights, too -- insights into dietary patterns, parasite evolution and more. The only problem is that it can be hard to identify the owner of really old feces.




Specifically, scientists have trouble differentiating between ancient human and dog feces. Dogs have been hanging out around humans for thousands of years. As a result, droppings from both are often found at archaeological dig sites. The droppings are frustratingly similar in size, shape and composition.

The difficulty of telling dog poop from human poop isn't insurmountable. Over the years, researchers have developed imperfect solutions to the problem.

RELATED Rare fossilized excrement found in La Brea Tar Pits

"It's not a complete black and white story, there was some work done before, especially by one of our co-author Karl Reinhard on identifying the 'host species' using pollen grains and parasites -- some are specific to some species -- but this still remained a challenge," Maxime Borry, researcher at the Max Planck Institute for the Science of Human History in Germany, told UPI in an email.

For a better way to identify ancient excrement, Borry and her colleagues at the Max Planck Institute teamed up with scientists from Harvard University and the University of Oklahoma to develop a machine learning algorithm.

The technology -- dubbed coproID, short for coprolite identification -- works by combining analysis of ancient host DNA with machine learning software trained to differentiate between the bacteria in the gut microbiomes of dogs and humans.

RELATED Researchers detect wildlife using DNA found in river water

"Our machine learning method was trained using gut microbiome composition of modern human, dog and soil samples," Borry said. "Once the model was giving good results on modern samples, we applied it to our paleofeces to predict their 'host species' -- human or dog."

Scientists hope their new technology will help researchers gain insights into the evolution of the human gut microbiome, including details related to the emergence of food intolerances and changes in human health.

The new technology -- described Friday in the journal PeerJ -- is not only useful, it also doesn't require much extra effort from scientists.

RELATED Earliest evidence of high-altitude living found in Ethiopia

"The key advantage of coproID is that it's using the same data that would anyway need to be generated to ask molecular evolution questions -- shotgun sequencing data -- so in a way, a coproID analysis in this kind of projects comes for free," Borry said.

Borry and her colleagues are currently working on a study of the evolution of the gut microbiome -- with the help of coproID, of course. Luckily for other archaeologists, the technology isn't proprietary.

"Scientists can already start using this technology if they have the data to work with: our method is freely and openly available online," Borry said. "This method is the first line of analysis for any further archaeological question: this helps verify the 'host species' of a sample before drawing any conclusion of the species microbiome with other techniques."

No comments:

Post a Comment