FFPopSim: An efficient forward simulation package for the evolution of large populations

TL;DR

FFPopSim introduces a fast, scalable simulation algorithm for large populations with many loci, enabling efficient modeling of complex evolutionary dynamics in population genetics.

Contribution

The paper presents a novel FFT-based algorithm that reduces simulation complexity from 8^L to 3^L, allowing efficient multi-locus population simulations.

Findings

Simulation runtime scales as 3^L, significantly faster than previous methods.

Supports arbitrary fitness functions and genetic maps.

Implemented as C++ classes with a Python interface.

Abstract

The analysis of the evolutionary dynamics of a population with many polymorphic loci is challenging since a large number of possible genotypes needs to be tracked. In the absence of analytical solutions, forward computer simulations are an important tool in multi-locus population genetics. The run time of standard algorithms to simulate sexual populations increases as 8^L with the number L of loci, or with the square of the population size N. We have developed algorithms that allow to simulate large populations with a run-time that scales as 3^L. The algorithm is based on an analog of the Fast-Fourier Transform (FFT) and allows for arbitrary fitness functions (i.e. any epistasis) and genetic maps. The algorithm is implemented as a collection of C++ classes and a Python interface.