I'm not sure what you mean by "non-random mating"; that certainty is not an accepted element of evolutionary theory. Indeed, sexual reproduction is the exception to how organisms reproduce (the bulk of life on earth reproduces asexually).
WOW David Attenborough lied to me about all those animals that evolve exagerated traits because they prefer certain traits in mates.
Weirdly it is still listed in wikipedia
http://en.wikipedia.org/wiki/Mate_choice
Mate choice, or intersexual selection, is an evolutionary process in which selection of a mate depends on attractiveness of its traits.
Sorry I must have missed the memo when it was overturned by science. I know Darwin included this selection process under natural selection but it generally got shifted out because the evolution it creates is not neccessarily the fitess of survival driven.
Epigenetic changes usually only last a couple of cellular generations. In metazoans [animals] most (perhaps all) epigenetic changes are "reset" in the embryo - i.e. they are not inherited parent->child.
WOW again can you cite evidence for those statements.
See again here is wikipedia
http://en.wikipedia.org/wiki/Transgenerational_epigenetics
Inherited epigenetic effects on phenotypes have been documented in bacteria, protists, fungi, plants, and animals.
Gene conversion is a form of mutation, and would be accounted for in the "mutation" stuff I mentioned earlier.
You view Biased gene conversion as a form of gene mutation???????
For me mutation is generally defined as a failure to store genetic information faithfully while genetic recombination is exactly that.
I certainly don't equate those two things but I guess it is not my area so it must be me.