@peercommunityin.bsky.social is one solution to this problem. We should all support it. #peercommunityin

Here's a link to the website of this course as hosted by the University of Utah: content.csbs.utah.edu/~rogers/tch/...

Maybe so. The estimated rate of M>N gene flow is lower than that in the other direction, but it would surely contribute.

Troff came before TeX, at least in my life. But yes, it was wonderful.

Thanks. I'll look at your preprint.

Very interesting. 2 questions: 1. How large are samples from A and B? 2. Can the result be expressed in terms of a molecular clock instead of a drift clock? A drift clock is usually needed only when one doesn't know which allele is ancestral.

Over multiple generations, you need to account for changes in p_i q_i. If the initial p is near 1/2, then pq changes much less than p does, so you can ignore changes in p_i q_i over short intervals.

In daughter population i, one generation of drift adds a random increment to allele frequency with mean 0 and variance p_i q_i/2n. These increments are independent, so the between-daughter variance would be twice this value.

If I understand you correctly, 1/2n, where n is sample size. But why do we need 2 samples? The Fst of each sample relative to the pop is 1/2n.

He wasn’t even adjusting anything he was walking calmly when they opened fire on him