New paper: Biased phylodynamic inferences from clusters

My new paper with Simon Frost and Fei Xiang is out in Virus Evolution showing how using phylogenetic clusters for phylodynamic investigation is problematic. Importantly, false clusters can masquerade as growing epidemics despite being identified from a randomly mixing, constant-sized population. We also show how other implicit biases can have an impact on parameter estimation and model selection, in part due to left censoring in the first coalescence event.

Our findings have important ramifications for targeting intervention strategies using clusters identified from large databases, adding to the findings of Art Poon recently published in Virus Evolution here. They demonstrate that going forwards, we should be cautious in our interpretations, to ensure that resources are targeted in the populations where they will have most impact.

July 3rd , 2017 by Bethany Dearlove