## Inference of effective population sizes and migration rates using MASCOT

Nicola Felix Müller
Taming the Beast 2018

• Birth-Death.
• Coalescent.

### How the coalescent process generates trees

N=8
• Probability that 2 lineages share a common ancestor in the previous generation $=\frac{1}{N(=8)}$

### Changes in population sizes are reflected in the shape of trees

The smaller a population, the more likely individuals are to coalesce.

### Phylodynamics: HCV in East-Asia

Phylodynamics allows to infer past population dynamics.

### Integrating over every possible migration history drastically speeds up analysis

Müller et al., Bioinformatics, Accepted
Derivations: Müller et al., MBE, 2017

### Speed up allows to consider more complex scenarios: H3N2

Müller et al., Bioinformatics, Accepted

### Problem: The number of parameters grows quadratically with the number of locations

Use additional data source to inform parameters. For example:

1. Population size and density of host populations.
2. Movement data between different locations.

Use predictors to inform effective population sizes and migration rates by using a GLM:

### 2014 Sierra Leone Ebola outbreak

1. Migration rate predictors: Distances between locations and origin/destination effects, such as population size at origin or destination .
2. Ne predictors: Weekly case data, travel times to cities, population density, etc.

### 2014 Sierra Leone Ebola outbreak

Müller et al., BioRxiv, 2018

### Tutorial: MASCOT

https://taming-the-beast.github.io/