QC for recombination rate (AraTha)

[srgibbon17]

PR for new recombination rates: #1855 (and see original discussion in #1854)

Paper links: Original recombination rates for A. thaliana are from Salomé et al. 2012 and modified following the adjustment for selfing from Nordborg 2000. The selfing rate is from Platt et al. 2010.

QC'er requests: Peter suggested @silastittes might take this up?

[petrelharp]

I've taken the very useful context provided by @srgibbon17 here and moved it over to #1854. Since the role of the QC is to re-implement "from scratch" what we need here is an explanation of what information needs to be pulled from where (not explanation of the code, which the QC person won't look at). Here is the rationale, as I understand it:

1. The QC PR should branch off of AraTha recombination rate bug fix + selfing #1855. We could merge in AraTha recombination rate bug fix + selfing #1855 but I'd like to get things sorted in a PR to make sure we agree that it's right.
2. What needs to be QC'ed are two things: (1) the Salome2012 genetic map was estimated by F2 crosses and therefore doesn't account for selfing, so now has a "selfing correction factor" multiplied in; (2) the per-chromosome mean recombination rates are now derived as the average of the rates from the Salome2012 map.
3. The selfing correction factor is from Nordborg 2000 Genetics ("Recombination and selfing" section (bottom left of page 925))
4. The selfing rate is from Platt et al 2010 PLoS Genetics
5. the mean per-chromosome rate is calculated using msprime.RateMap.mean_rate

However: genetic maps aren't QC'ed (since that would require re-downloading the numbers, re-lifting-over, etc, I guess). So my suggestion is to (a) recalculate the mean per-chromosome rates using the description above, and (b) put these in some tests in tests/test_AraTha.py with an explanation of how the numbers that are in the tests were arrived at?

[silastittes]

Is the PR ready or still in progress?

[srgibbon17]

It's ready on my end! But, @petrelharp may still want to make a quick pass after my last few edits?

[petrelharp]

I still have some nitpicks on the PR but it's ready enough to do the QC on top of. (So TBC @silastittes you'll make a PR with #1855 as the base branch.)

[silastittes]

I still have some nitpicks on the PR but it's ready enough to do the QC on top of. (So TBC @silastittes you'll make a PR with #1855 as the base branch.)

To make sure I understand, my QC tests etc. will be commits to #1855 before it's merged @petrelharp?

[petrelharp]

No, your commits will be in a new PR, but the base branch for that PR will not be main but instead be the branch that corresponds to #1855. So when you're done we will (a) merge your PR to @srgibbon17 's and then (b) merge #1855 (which will now have your commits).