Towards a phylogenetic measure to quantify HIV incidence

TL;DR

This paper introduces a phylogenetic approach using approximate Bayesian computation and a novel tree statistic to estimate and differentiate epidemiological parameters of HIV, aiding in understanding epidemic dynamics from genetic data.

Contribution

It develops a new phylogenetic measure based on coalescent theory to infer HIV epidemic parameters solely from genetic data, improving epidemic assessment.

Findings

The tree statistic differentiates epidemiological parameters effectively.

Phylogenetic data can estimate epidemic sizes and diagnosis rates.

Method enhances understanding of undiagnosed HIV populations.

Abstract

One of the cornerstones in combating the HIV pandemic is being able to assess the current state and evolution of local HIV epidemics. This remains a complex problem, as many HIV infected individuals remain unaware of their infection status, leading to parts of HIV epidemics being undiagnosed and under-reported. To that end, we firstly present a method to learn epidemiological parameters from phylogenetic trees, using approximate Bayesian computation (ABC). The epidemiological parameters learned as a result of applying ABC are subsequently used in epidemiological models that aim to simulate a specific epidemic. Secondly, we continue by describing the development of a tree statistic, rooted in coalescent theory, which we use to relate epidemiological parameters to a phylogenetic tree, by using the simulated epidemics. We show that the presented tree statistic enables differentiation of epidemiological parameters, while only relying on phylogenetic trees, thus enabling the construction of new methods to ascertain the epidemiological state of an HIV epidemic. By using genetic data to infer epidemic sizes, we expect to enhance understanding of the portions of the infected population in which diagnosis rates are low.