Via Ars Technica, a report on a recent study involving Andrew Rambaut's laboratory: Flu viruses derive from a global selective sweep in the 1870s. Click through for the full report and links. Excerpt:
Phylogeny inference, or tree building, is used to deduce the history of entities related by common descent, whether they're genes, populations, or species. Ultimately, it can reveal the evolutionary relationships between these entities, even if we don't have access to the intermediate steps.
But it's a method that has to be used cautiously. "Erroneous conclusions can show strong statistical support," warns a new paper, published in Nature. Andrew Rambaut's lab, which wrote the paper, issued that warning because phylogenetic studies can have serious practical implications: "This is a serious problem given that such results are widely used to infer when, where, and how pandemic and panzootic viruses have emerged."
The "pandemic and panzootic" virus in question is the influenza A strain that swept the globe in 1918, killing millions. The new analysis by Rambaut's lab suggests that previous work on the 1918 flu was fundamentally flawed.
This problem is that viral genes mutate at different rates depending on which host species the virus is infecting. Researchers know that different species, and even different genes within a species, mutate at different rates; hence, during phylogenic analysis they usually use a "relaxed clock model" that allows for the mutation rate to vary over time and between organisms.
Rambaut himself used this relaxed clock in a 2006 paper he titled (probably without irony) "Relaxed Phylogenetics and Dating with Confidence." But this new paper says that an older one of his "very seriously underestimates the time to the most recent common ancestor," probably because the older one only looked at human flu viruses.
The flu viruses they have now examined anew include equine, porcine, avian, and human strains. This most recent phylogenomic analysis using a new host-specific local clock indicates that their most recent common ancestor dates very recently—to the latter half of the nineteenth century.
That's when avian influenza viruses underwent a global selective sweep, a process in which DNA sequences tag along with a beneficial mutation that's under positive selection. Even though these nearby sequences confer no benefit themselves, they become more prominent in a population along with the gene that is actually useful.