Aede

Local Ancestry Simulation

Install

From github

git clone https://github.com/BioShock38/aede.git

cd aede

python setup.py install --user

Tutorial

You need an array of distances in morgan also called a genetic map. So for each SNP you have the distance in morgan in gen_map.

The most simple way to run a simulation, with default parameters

import aede.simulator as simu

loc_anc = simu.LocalAncestrySimulator(gen_map, simu.default_pipeline_hapmix)

Then we can build admixed haplotypes with parental haplotypes

h_adm, clusters = loc_anc.run(pop_1=H_ceu, pop_2=H_yri)

We obtain h_adm a numpy array of admixed haplotypes and clusters is an array indicating from which population is the SNP selected from.

You can wrap the functions, to build one function.

import aede.simulator as simu
from functools import partial

def h_adm_simu(lambd, H_1, H_2, genmap, mu=0.8, sigma2=0.01):
    jumps_fun = partial(simu.jumps_builder, lambd=lambd)
    cluster_fun = partial(simu.cluster_builder, a_beta=mu, b_beta=sigma2)

    loc_anc = simu.LocalAncestrySimulator(genmap, simu.pipeline_hapmix)
    h_adm, clusters = loc_anc.run(pop_1=H_1, pop_2=H_2,
                                  jumps_builder=jumps_fun,
                                  cluster_builder=cluster_fun)
    return h_adm, clusters"

Full Example

import numpy as np
from functools import partial

H_ceu = np.load("H_ceu.npy")
H_yri = np.load("H_yri.npy")

gen_map = np.load("gen_map.npy")

import aede.simulator as simu

# Custom lambda parameter for exponential law
jumps_10 = partial(jumps_builder, lambd=10)

loc_anc = simu.LocalAncestrySimulator(gen_map, simu.pipeline_hapmix)
h_adm, clusters = loc_anc.run(pop_1=H_ceu, pop_2=H_yri,
                              jumps_builder=jumps_10,
                              cluster_builder=simu.cluster_default)