Virus-Microbe Interaction Ecology

“The title of this talk could be the title of my whole career,” says Melissa Duhaime of the University of Michigan. Virus-microbe ecology examines the interactions between viruses, hosts, and the environment. Duhaime and colleagues have begun researching viral ecology in the context of colorectal cancer treatment, which is presented here for the first time.

“Most of what I know about colon cancer I learned in the last 48 hours,” Duhaime warns, but she goes on to explain that it is the second-leading cause of cancer related deaths in the US. In one year, the equivalent of the population of Philadelphia will be diagnosed with colon cancer. There is a need for cheap, non-invasive screening, and the progression of colorectal cancer is marked by a fairly consistent series of bacterial community perturbations. Viruses are key to microbial community composition, stability, and function, and altered human virome is associated with several diseases.

Examining viral diversity in three different groups – pre-cancerous polyps, cancerous, and healthy individuals – found no difference in alpha or beta diversity. Stool samples were analyzed using random forest models instead. Colorectal cancer-associated communities were found to follow predictable patterns, and microbial and viral association with colon cancer was found to be driven by a few key organisms. But are the most influential viruses present simply because they target the most abundant, influential microbes? No – relative abunances of influential viruses and microbes had no correlation.

But there could still be several reasons for this: lysogenic infections, non 1:1 relationship between viruses and possible hosts, difficulties in relative abundance estimates, or that influential viruses are not infecting the most influential bacteria. Predicting “who infects whom” is a “trivial task to perform, but a non-trivial task to evaluate”. So Duhaime and colleagues used a random forest model again to classify virus-host interactions. Do the most influential viruses in CRC etiology have a broad range of potential hosts? Possibly, but questions remain. The results show that that colorectal cancer progression is associated with changes in the microbiome and virome, and that certain virus OTUs are associated with these changes. Duhaime is most excited about the fact that her work provides a framework to establish predator-prey interaction networks, community ecology, and ecosystem function.

-Jacob Grace