HaploStats doesn't report correct low frequency haplotypes

[sjmack]

See my final comment in Issue #28.

<haplostats> is reporting converged haplotype frequency estimates differently from <emhaplofreq>, is not reporting haplotype counts, and is truncating frequencies that were returned in scientific E-notation to eliminate the exponent.

[alexlancaster]

see fuller description of issue here: #60 (comment) and proposed solution here: #60 (comment)

[alexlancaster]

@sjmack looking at this issue, it seems that the low frequency truncation applies to both emhaplofreq and haplostats and is actually separate to the lack of counts (which is haplostats-specific). I would open up another issue for the numbers of copies.

quick review of the code suggests it was disabled in the XML output pending a check of the numbers, there's a FIXME:

https://github.com/alexlancaster/pypop/blob/main/src/PyPop/Haplo.py#L1013

[alexlancaster]

@sjmack When you open up a new issue, I'd remove the comment in the code and re-run, and check to see if the copies are calculated correctly. if so, we can probably uncomment and add it. We should also unit tests for haplo-stats, so a self-contained .ini and .pop that I can use for that purpoes, would be really helpful.

[alexlancaster]

@sjmack I made a branch to test this fix. in your local git clone you can pull it and rebuild to test it on your data:

git pull
git checkout 38-problem-with-haplostats-implementation

if it works for you an passes all the unit tests I'll merge it with main.

[alexlancaster]

@sjmack check out the pull request #61 (see above for instructions forbuidling and give me an 👍 or 👎 )

as I mentioned on the pull request it would be good to be able get a unit test to go along with this, if possible.

[alexlancaster]

splitting off the count issue as a new issue in #65 and retitling this issue

[sjmack]

@alexlancaster I will try to get to this later this week. Sorry for the delay.

[alexlancaster]

@sjmack opened up a new issue #104 for preserving scientific notation when truncating.