Recombination and Gene Conversion

Crossing Over (Recombination)

Recombination is modeled using the -r flag, which takes the population recombination rate ρ = 4Nr as a parameter, where:

  • N is the current effective population size

  • r is the probability of crossover per base pair per generation

Recombination can occur between any of the discrete sites being modeled.

Example:

# Simulate with recombination rate rho = 2.4
./discoal 3 2 100 -t 2 -r 2.4

Gene Conversion

Gene conversion (non-crossover recombination events) is simulated using the -g flag. This models recombination without exchange of flanking markers.

Parameters:

  1. γ = 4Ng: The population gene conversion rate

    • g is the probability of initiating a gene conversion event per base pair

  2. Mean tract length: The average length of converted regions

    • Tract lengths follow a geometric distribution

Example:

# Gene conversion with rate 2.4 and mean tract length 10bp
./discoal 3 2 100 -t 2 -r 2.4 -g 2.4 10

Alternative Gene Conversion Model

The -gr flag allows specifying gene conversion as a ratio to crossover events:

# Gene conversion rate = 0.5 * recombination rate, tract length 50bp
./discoal 10 5 10000 -t 10 -r 20 -gr 0.5 50

In this model, the gene conversion initiation rate is calculated as rho × conversionToCrossoverRatio.

Combining Recombination Models

Both crossing over and gene conversion can be active simultaneously:

# Both crossovers and gene conversions
./discoal 20 10 50000 -t 50 -r 40 -g 20 100

This simulates: * Crossover rate: ρ = 40 * Gene conversion rate: γ = 20 * Mean conversion tract length: 100bp

Implementation Notes

  • Recombination breakpoints are tracked dynamically

  • The simulator uses efficient segment-based tracking of ancestral material

  • Recombination events only occur in regions with ancestral material

  • During selective sweeps, recombination continues but migration is suspended