Here are some epistatic patterns:
They're special (mostly) because they are all the possible binary two locus genotype-phenotype maps wherein one of the loci has no variance when the other locus is wildtype. Populations were simulated for 200 generations, in which time some patterns reached equilibrium, others went to extinction. The marginal variance at the B locus relative to the expected Hardy-Weinberg genotype frequencies at the second locus was plotted against the observed marginal variance at every generation for 200 replicate simulations in this graph, where alpha levels of points darken with population age:
The residual interaction variance plots look like this:
And the total variance deviations:
What are these boundaries? I have given up trying to understand, now I just look for pictures. Behold, interactions/10, a wood planer.
Typing and design by Gib Hemani