Simulation of selected genealogies

Area of Science:

• Population Genetics
• Evolutionary Biology
• Computational Biology

Background:

• Gene genealogies are fundamental to understanding population genetics.
• Simulating genealogies under selection is computationally challenging.
• Existing models often simplify or ignore selection's complex effects.

Purpose of the Study:

• To develop and present algorithms for generating genealogies conditional on sample configuration under selection.
• To utilize the ancestral selection graph for accessible simulation of non-neutral genealogies.
• To estimate ancestral times and quantify selection's impact on key genealogical metrics.

Main Methods:

• Developed algorithms for generating genealogies in haploid and diploid models with selection.
• Employed enhanced integro-recursions using the ancestral selection graph (ASG).
• Utilized a Monte Carlo simulation scheme for estimating ancestral times under selection.

Main Results:

• Selection significantly alters the expected depth of conditional ancestral trees.
• Branch lengths are demonstrated to be ineffective for detecting selection.
• Quantified the effects of selection on the expected time to the most recent common ancestor.

Conclusions:

• The ancestral selection graph provides an accessible method for simulating genealogies under selection.
• Selection's influence on genealogical depth is dependent on mutation-selection balance.
• New methods are needed to reliably detect selection using genealogical data.