REGENS: an open source Python package for simulating realistic autosomal genotypes

REcombinatory Genome ENumeration of Subpopulations (REGENS) is an open source Python package that simulates autosomal genotypes by concatenating real individuals’ genomic segments in a way that preserves their linkage disequilibrium (LD), which is defined as statistical associations between alleles at different loci (Slatkin, 2008). Recombining segments in a way that preserves LD simulates autosomes that closely resemble those of the real input population (Shi, 2018) because real autosomal genotypes can be accurately modeled as genomic segments from a finite pool of heritable association structures (LD haplotypes) (Druet, 2009). REGENS can also simulate mono-allelic and epistatic single nucleotide variant (SNV) effects of any order without perturbing the simulated LD pattern. The SNVs involved in an effect can contribute additively, dominantly, recessively, only if heterozygous, or only if homozygous. All simulated effects contribute to the value of either a binary or continuous biological trait (phenotype) with a specified mean value and a specified amount of random noise.