Mutation model

[daikitag]

Short description of what you'd like added:
Document how users can use msprime.sim_mutations to use their own mutation model.

Describe briefly how you think stdpopsim could be better.
I think there are some people who are interested in simulating a tree sequence by using one of the demographic models offered in stdpopsim, and use their own mutation model to simulate mutations. I think the best way to implement this is to set zero mutation rate and use msprime.sim_mutations, as sim_mutaitons function has a lot of input parameters. However, when I set the mutation rate as zero, the following warning message comes out:

UserWarning: The demographic model has mutation rate 2.35e-08, but this simulation used the contig's mutation rate 0. Diversity levels may be different than expected for this species. For details see documentation at https://popsim-consortium.github.io/stdpopsim-docs/stable/tutorial.html

This makes us confused about the validity of setting mutation 0, as some people might think that mutation rate is influencing the output tree sequence structure. Instead, would it be possible for you to document that the mutation rate has no influence on msprime's ancestry simulation and users can specify their own mutation model by using msprime.sim_mutaitons afterwards?

[petrelharp]

This is a great idea; and sounds like a good topic for an (advanced) vignette.