Overview of FFPopSim
FFPopSim consists of two parts enabling genotype-based or individual-based simulations. The class "haploid_lowd" can be used for simulations of a few sites (up to 20) and models the distribution of the population among all possible genotypes. "haploid_highd" allows to simulate longer genomes by modeling by tracking only the genomes present in the population.Both classes include a rich variety of utility functions, and can be extended by C++/Python subclassing seamlessly. Documentation and examples of C++ routines, Python scripts, and subclassing are included.
Genotype-based simulations: haploid_lowd
The binary hypercube is used to represent genotypes in "haploid_lowd".
Mutation is equivalent to a flow along edges of the cube, recombination takes two
genomes (vertices of the cube) and produces new genomes be mixing the parental genomes.
"haploid_lowd" models the whole genotype distribution efficiently, offering the following features:
- Tracks the frequencies of all possible genotypes
- Arbitrarily large populations
- Arbitrary fitness functions and genetic maps (including time-dependent)
- Asexual and sexual populations
- Low run-time complexity (compared to the naive O(8^L)):
- O(L 2^L) with single crossover recombination -> up to 20 loci.
- O(3^L) with general recombination -> up to 15 loci.
- Site-specific recombination rates, site- and direction-specific mutation rates
- Deterministic and stochastic simulations
Individual-based simulations: haploid_highd
Genomes are represented in "haploid_highd" as bitstrings, ensuring an
efficient memory usage. Mutation is a bit flip, and recombination creates
a new bitstring from two existing ones."haploid_highd" is a more traditional, individual-based simulation interface for populations with a larger number of sites. It uses clones instead of individuals to increase efficiency in case of low mutation and recombination rates; it is not, however, restricted to these conditions, and is currently used to simulate the highly variable Human Immunodeficieny Virus (HIV). It offers the following features:
- Long genomes (several millions of loci, depending on memory availability)
- Large populations (e.g. 100,000 individuals with 1,000,000 loci each)
- Tracks several phenotypic traits that codetermine fitness
- Arbitrary genotype-phenotype maps (including time-dependent)
- Asexual and sexual populations