Incorporate inferred tree sequences to the software

[shz9]

The ultimate goal of this project is to align inferred tree sequences to the pedigree graph. Therefore, we need to create tools to infer tree sequences from genetic samples.

I talked to @kobrica and she will work on providing this functionality. This task is composed of 3 steps.

• Given genetic samples (e.g. genotype matrix), infer the corresponding tree sequences using tsinfer (tutorial).
• Convert the inferred tree sequences to Traversal objects.
• Given msprime simulation results, obtain the corresponding inferred tree sequences. This will involve 2 steps: First, obtain the genetic samples from the msprime simulation, and then run tsinfer on those samples, as in the tutorial above.

[andjelatodorovic]

I have completed all of the above tasks, and added some samples of genotype matrices in csv format in data folder.
This is easier than manually adding row by row to sample data.

The tree sequence to Traversal may need to be double-checked, I'm not quite sure the Traversal object has all of the properties. Please refer to commit 66a2702

[shz9]

Can you please modify the infer_from_msprime function so that it takes a simulation object (like the one we get from generate_msprime_simulations)? In other words, you can assume you already have the result of the simulation.

[andjelatodorovic]

@shz9 Sure! Should it take both simulation results and traversal object as parameters?

[shz9]

I think just the simulation results are needed.