phylodyn

phylodyn performs phylodynamic inference by sampling genealogies to estimate historical effective population sizes using Tajima's coalescent.

Key Features:

• Bayesian approach: Implements a Bayesian framework for inferring historical effective population sizes using Tajima's coalescent, which reduces the genealogical state space relative to Kingman's coalescent.
• Algorithmic innovation: Employs a directed acyclic graph (DAG) and a tailored Markov Chain Monte Carlo (MCMC) method to enable efficient and exact likelihood calculations for nonrecombining datasets.
• Performance comparison: Has been compared with BEAST using simulated and human data and demonstrated accurate inference of effective population sizes.

Scientific Applications:

• Evolutionary biology: Reconstruction of historical population size trajectories for evolutionary inference.
• Epidemiology: Analysis of past population dynamics relevant to infectious disease spread and outbreak history.
• Population genetics: Inference of demographic history and effective population size changes.
• BESTT algorithm implementation: Uses the Bayesian Estimation of Population Size Changes by Sampling Tajima's Trees (BESTT) algorithm as an alternative approach for demographic inference.

Methodology:

Uses Bayesian inference with Tajima's coalescent and the BESTT algorithm, applies a directed acyclic graph (DAG) for exact likelihood calculations, and samples genealogies with a tailored Markov Chain Monte Carlo (MCMC) sampler for nonrecombining datasets.