Phylodynamic inference and model assessment with approximate bayesian computation: influenza as a case study

Area of Science:

• * Infectious disease epidemiology
• * Evolutionary virology
• * Computational biology

Background:

• * Understanding viral disease drivers requires integrating incidence, genetic, and antigenic data.
• * Phylodynamic models are crucial for simulating pathogen evolution and ecology.
• * Approximate Bayesian Computation (ABC) offers a simulation-based Bayesian approach for model fitting.

Purpose of the Study:

• * To apply Approximate Bayesian Computation (ABC) for fitting and assessing phylodynamic models of viral diseases.
• * To analyze spatial models of human influenza A virus subtype H3N2 phylodynamics.
• * To investigate the interplay between ecological and evolutionary processes in viral dynamics.

Main Methods:

• * Employed a simulation-based, Bayesian method: Approximate Bayesian Computation (ABC).
• * Analyzed two spatial models for influenza A(H3N2) phylodynamics: continuous waning immunity and epochal evolution.
• * Integrated long-term surveillance data from The Netherlands with hemagglutinin gene sequence data from Northern Europe.

Main Results:

• * ABC successfully estimated key phylodynamic parameters for influenza A(H3N2).
• * The epochal evolution model, within a spatial context, best reproduced observed incidence patterns and phylogeny.
• * Inconsistent incidence dynamics suggested a very high reproductive number, not aligning with empirical attack rates.

Conclusions:

• * Interactions between evolutionary and ecological processes impose quantitative constraints on viral phylodynamics.
• * Sequence and surveillance data can be synergistically utilized to understand viral dynamics.
• * ABC is a versatile method for integrating diverse data types with phylodynamic models, requiring careful calibration.