simGL simulates genotype likelihoods (GLs) from haplotypic genotype matrices, given per-sample coverage and sequencing error rates. It is designed to work seamlessly with msprime and tskit pipelines, but accepts any NumPy haplotype matrix.
pip install simGLOr via conda:
conda install -c conda-forge simGLOr from source:
git clone https://github.com/RacimoLab/simGL.git
cd simGL
pip install -e .importmsprimeimportnumpyasnpimportsimGL# 1. Simulate a tree sequence and extract the biallelic genotype matrixts=msprime.sim_ancestry(
samples=10, ploidy=2, sequence_length=100_000,
recombination_rate=1e-8, population_size=10_000, random_seed=1,
)
ts=msprime.sim_mutations(ts, rate=1e-4, random_seed=1)
gm_full=ts.genotype_matrix()
biallelic=gm_full.max(axis=1) ==1gm=gm_full[biallelic] # shape (n_sites, n_haplotypes)# 2. Get reference and alternative allelesvariants=list(ts.variants())
ref=np.array([v.alleles[0] forvinvariants])[biallelic]
alt=np.array([v.alleles[1] forvinvariants])[biallelic]
# 3. Simulate allele read countsarc=simGL.sim_allelereadcounts(
gm, mean_depth=10., std_depth=2., e=0.01,
ploidy=2, seed=42, ref=ref, alt=alt,
)
# arc shape: (n_sites, n_individuals, 4) — A, C, G, T read counts# 4. Compute genotype likelihoodsGL=simGL.allelereadcounts_to_GL(arc, e=0.01, ploidy=2)
# GL shape: (n_sites, n_individuals, 10) — all diploid ACGT genotypes# 5. Subset to biallelic genotypes and write a VCFRa=simGL.ref_alt_to_index(ref, alt)
GL_sub=simGL.subset_GL(GL, Ra, ploidy=2)
pos=np.array([int(v.site.position) forvinvariants])[biallelic] +1names= [f"ind{i}"foriinrange(ts.num_individuals)]
simGL.GL_to_vcf(GL_sub, arc, ref, alt, pos, names, "output.vcf")Full documentation — installation, user guide, API reference, and theory — is available at https://simgl.readthedocs.io.
If you use simGL in your work, please cite:
Coll Macià, M. & Gower, G. (2026). simGL: Simulate genotype likelihoods from haplotypic genotype matrices (v0.2.0). Zenodo. https://doi.org/10.5281/zenodo.19637642
Please also cite the methodological papers listed in the Citation page of the documentation.