A single tooth, part of a jaw, a piece of a skull: sometimes, that is all scientists have to tell the story of animals that went extinct millions of years ago. But new methods developed by scientists at the American Museum of Natural History and Stony Brook University are making it easier to accurately place these animals in the tree of life with partial fossils and DNA data. The novel approach, published in the journal Systematic Biology and tested on ancient bats, could be applied across a wide field of research.
“In mammals, and particularly in bats, teeth are a very useful tool for identifying species,” said Paul Velazco, a postdoctoral researcher at the American Museum of Natural History and an author on the paper. “Bats have very fragile bones that are not well preserved by fossilization. But teeth are sturdy and species-specific. So when we want to know about where an extinct species fits in the bat family tree, we often look at the characters on their teeth.”
Building evolutionary trees was traditionally done by detailing the morphology, or anatomy, of living and extinct species. This method requires close examination of the shape and structure of fossils as well of the bones, skin, and tissue of modern animals. But in recent years, scientists have found that they can build larger trees faster by sequencing the DNA of many species.
“By and large people have switched over to molecular-based trees,” said Nancy Simmons, a curator in the Museum’s Department of Mammalogy and an author on the paper. “However, if you want to give calibration points for when evolutionary events occurred within a group, or figure out how extinct species are related to living animals, you have to work with morphology.”
The problem is that while methods to estimate molecular trees have built-in statistical calibrations that account for an effect called “saturation”—when genetic sequences undergo so many changes that information about evolutionary relationships is lost—methods traditionally used to estimate trees based on morphology do not correct for these high rates of change. In addition, because bones and teeth are actively shaped by natural selection, some subset of features ends up being similar in different species because of a similar function, not because the different species evolved from a common ancestor.
“Without stepping back and examining the statistical profile of anatomical features, researchers have a difficult time distinguishing what traits were inherited as opposed to features that evolved independently and ended up being similar because of natural selection,” said Liliana Dávalos, lead author on the new paper, who is an assistant professor at Stony Brook University and a research associate at the Museum. “For example, bats that eat nectar may have very similar skulls and teeth even if they evolved independently from species with different dental anatomy and diets.”
The researchers developed two new approaches to combine molecular and morphological data to help place three fossils from 12-million-year-old bats. The fossil species are part of the New World leaf-nosed bats, the most ecologically diverse family of living mammals.
Evolutionary trees built with the new methods confirmed previous ideas about the positions of two of the fossil bats, which group with living carnivorous bats. However, the third fossil species, a nectar-feeding bat of the genus Palynephyllum, was placed in quite a different place in the tree of life than previously assumed, found to belong to a lineage of nectar-feeding bats that originated and spread throughout South America. This finding confirms the deep roots of nectar-feeding specializations on that continent.
“We can now put these firm calibration points in place unambiguously with the combination of these two types of data,” Dávalos said. “This is very exciting information for researchers working on bats, but these methods are going to be much farther-reaching.”
Read the paper's abstract here.
Learn more about Dr. Simmons's research in video.
Funding for this study was provided by the National Science Foundation, grant numbers DEB 0949759 and DEB 0949859.