Seminars

October 20 2016

5:00 pm TLSB 1100

EcoEvoPub Series

Graduate Student Presentations

Summary

Mairin Balisi
Department of Ecology and Evolutionary Biology, UCLA

"Hypercarnivory and extinction risk in North American fossil dogs"

Hypercarnivory (a diet consisting of over 70% meat) and bone-crushing are metabolically costly specializations, and their appearance in a lineage is invariably irreversible: an example of a "macroevolutionary ratchet". While modern ecosystems are relatively depauperate of hypercarnivores and bone-crushers, these specializations have repeatedly arisen in the fossil record, permitting examination of a) how hypercarnivory may affect extinction risk and b) how quickly an empty hypercarnivore niche is filled. North American fossil dogs (Mammalia: Carnivora: Canidae) comprise over 100 species spanning a wide range of ecomorphologies, including iterative occurrences of hypercarnivory. Here, we reconstruct the rise of hypercarnivory in dogs, examining a set of 12 ecomorphological indices over nine time slices from the origin of Canidae (40 million years ago) to the height of canid species richness (25 species at peak richness; 15 million years ago). Hotspots of elevated extinction risk correspond to areas of the dog morphospace occupied by large but not small hypercarnivores, matching the prediction of hypercarnivory representing an evolutionary “dead end” but showing size to exert a confounding effect. Hypercarnivory is slow to arise in dogs: With non-dog carnivorans initially occupying the carnivore and large-hypercarnivore space, dogs first saturated the omnivore and small-carnivore space and remained there well after large hypercarnivorous non-dogs became extinct. This significant lag in the movement of dogs into the large-hypercarnivore space suggests that the turnover resulted from passive replacement and ecological release rather than active displacement. Little ecomorphological overlap occurs among dogs and non-dogs, contrary to hypotheses that dogs declined taxonomically from competition with non-dog clades. We explore the "cat gap" and the potential for competition between dogs and bear-dogs.

Katie Gostic
Department of Ecology and Evolutionary Biology, UCLA

"Birth year predicts immunity against emerging influenza viruses with pandemic potential"

Novel pathogens that emerge from animal reservoirs are commonly assumed to face an immunological blank slate in their new host populations. Influenza A viruses (IAVs) are a prime example, where sixteen distinct subtypes of influenza’s hemagglutinin (HA) antigen are known to circulate in animal reservoirs, but only three of these subtypes (H1, H2 and H3) have circulated in humans within the past hundred years. A fundamental tenet of influenza epidemiology has always been that a pandemic caused by one of these novel subtypes would put the entire human population at risk. However, a growing body of research shows that cross-immunity is possible between subtypes established in humans (e.g. H1, H3) and subtypes emerging from avian reservoirs (e.g. H5, H7). Observed human cases of H5N1 and H7N9, two emerging avian subtypes of great concern, show unusual and previously unexplained age distributions. These unusual age distributions of infection might be signatures of age-specific differences in pre-existing immunity.
We compiled and analyzed data from all known human cases of H5N1 and H7N9 to test whether differences in exposure history across birth years can explain observed H5N1 and H7N9 age distributions. Model selection showed overwhelming support for the HA imprinting hypothesis, where a child’s first influenza A infection confers lifelong, partial protection against HA subtypes in the same phylogenetic group. We estimated protective HA imprinting reduces the risk of severe infection with H5N1 or H7N9 by 75%. Thus, contrary to prevailing scientific opinion at this time, we show that pre-existing immunity strongly and predictably shapes age distributions of severe infection with novel influenza A viruses. These results provide new scope to forecast age distributions of severe infection in future pandemics, and to predict probabilities that specific emerging subtypes would be able to spread through current immune landscapes.


 

 

 

 

 

 



 

 

 

 

 

 

 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



 



this is idtest: