Simulate genotypes for unit tests

I’m trying to simulate some genotype calls for unit tests using a simple model from msprime. At @jeromekelleher’s suggestion, I tried this:

# Using 1.x code, not yet released
!pip install git+https://github.com/tskit-dev/msprime.git#egg=msprime

import msprime
model = msprime.Demography.island_model(2, migration_rate=1e-3, Ne=10**4)
samples = model.sample(10, 10) # 10 haploids from each pop
ts = msprime.simulate(
     samples=samples,
     demography=model, # what should this be?
     mutation_rate=1e-8,
     recombination_rate=1e-8
)
list(ts.variants())
[]

What is demography supposed to be? I saw a “FIXME” for documenting it in simulate but I wasn’t sure what to set it to if not the Demography instance that comes back from island_model.

Also, is this why I’m not getting any variants back out?

I am ultimately trying to simulate hardy weinberg equilibrium. I’d like to have some fraction of the variants be in or near perfect equilibrium and the remainder way out of it. If you get a chance, could I get an assist @jeromekelleher?

[eric-czech]

+1 @eric-czech when we figure out the right toolset for (test) data generation it could be valuable to expose them as strategies in hypothesis (this could be a separate test module within this repo). wdyt?

Yea, I don’t see why not

[jeromekelleher]

I think we can - closing.