Eur. Phys. J. Special Topics 222, 1347-1358 (2013)
https://doi.org/10.1140/epjst/e2013-01930-0

Regular Article

Inferring epidemiological parameters from phylogenetic information for the HIV-1 epidemic among MSM

¹ Computational Science, University of Amsterdam, Science Park 904, 1098 XH Amsterdam, The Netherlands
² Department of Virology, Erasmus Medical Centre, Erasmus University, Rotterdam, The Netherlands
³ National Research University of Information Technologies, Mechanics and Optics (ITMO), Kronverkskiy 49, 197101 Saint Petersburg, Russia
⁴ Nanyang Technological University, 50 Nanyang Avenue, 639798 Singapore

^a e-mail: r.quax@uva.nl

Received: 3 June 2013
Revised: 10 July 2013
Published online: 13 September 2013

Abstract

The HIV-1 epidemic in Europe is primarily sustained by a dynamic topology of sexual interactions among MSM who have individual immune systems and behavior. This epidemiological process shapes the phylogeny of the virus population. Both fields of epidemic modeling and phylogenetics have a long history, however it remains difficult to use phylogenetic data to infer epidemiological parameters such as the structure of the sexual network and the per-act infectiousness. This is because phylogenetic data is necessarily incomplete and ambiguous. Here we show that the cluster-size distribution indeed contains information about epidemiological parameters using detailed numberical experiments. We simulate the HIV epidemic among MSM many times using the Monte Carlo method with all parameter values and their ranges taken from literature. For each simulation and the corresponding set of parameter values we calculate the likelihood of reproducing an observed cluster-size distribution. The result is an estimated likelihood distribution of all parameters from the phylogenetic data, in particular the structure of the sexual network, the per-act infectiousness, and the risk behavior reduction upon diagnosis. These likelihood distributions encode the knowledge provided by the observed cluster-size distrbution, which we quantify using information theory. Our work suggests that the growing body of genetic data of patients can be exploited to understand the underlying epidemiological process.