The influence of viral within-host evolution on transmission chain reconstruction: a simulation study.

Abstract

The Human Immunodeficiency Virus (HIV) is a measurably evolving pathogen responsible for the Acquired Immune Deficiency Syndrome (AIDS). Since the within-host evolution timescale is faster than the between-host transmission one, phylogenetical information can be used to infer epidemiological parameters. One possible piece of information that was proposed to be embedded in the sequences sampled from the hosts of an epidemics is the transmission tree. The phylogeny reconstructed from the sequences is used as an approximation of the transmission tree. This assumes that the within-host evolution has a negligible effect on the transmission tree reconstruction. In order to explore this assumption, this work presents the implementation of a mechanistic model of within-host viral evolutionary dynamics. A preliminary analysis identified the tree triplets distance as a suitable metric for reconstructed transmission tree comparison and reconstruction performance inference. Subsequent analysis explore the role of sampling time and mutation rate on transmission tree reconstruction performance. Different sampling strategies along the epidemics were proposed and confronted. The results suggested heterochronous sampling have a better reconstruction performance in comparison to homochronous sampling at the end of the simulation. Nevertheless, the eterochronous sampling strategies seem to present analogous performance. The role of the mutation rate was explored by simulating epidemics with a higher and a lower mutation rate, respectively. A higher mutation rate seems to provide more evolutionary signal for the reconstruction, improving its performance. This work represents a step towards the comprehension of the role of various factors on transmission tree reconstruction, and provides a software for hypothesis testing in a within-host evolution framework.