From an evolutionary standpoint, homosexuality is a trait that should not develop and persist in the face of natural selection. Yet it exists in most cultures, among men and women. Analyses have noted that homosexuality can run in families, leading researchers to hypothesize a genetic underpinning of sexual preference but no gene or group of genes for homosexuality have been found, despite numerous studies searching for a genetic connection. 

That may be because homosexuality is epigenetic, rather than genetic, according to a new model.  Epigenetic tags, a cell's epigenetic profile, tell genes whether to be on or off and those are impacted by the signals it has received during its lifetime. Epigenetics has a great deal of promise regarding understanding the human condition - understanding epigenetics means understanding how your environment and your own choices can influence your genetic code and even pass it on to your kids.  That promise also leads to dodgey deterministic ideas and cultural judgments, like if you vote for a political candidate you are the same as a parent who intentionally gives birth to a child with Down Syndrome. Or that a hamburger during pregnancy dooms your child to having bad grades in school. 

A paper in The Quarterly Review of Biology says that sex-specific epi-marks, which normally do not pass between generations and are thus "erased," can lead to homosexuality when they escape erasure and are transmitted from father to daughter or mother to son. 

They integrated evolutionary theory with molecular regulation of gene expression and androgen-dependent sexual development to produce a biological and mathematical model that delineates the role of epigenetics in homosexuality.  Epi-marks constitute an extra layer of information attached to our genes' backbones that regulates their expression. While genes hold the instructions, epi-marks direct how those instructions are carried out – when, where and how much a gene is expressed during development. Epi-marks are usually created each generation, but they contend that sometimes there is carryover between generations and that can contribute to similarity among relatives, resembling the effect of shared genes. 

Sex-specific epi-marks produced in early fetal development protect each sex from the substantial natural variation in testosterone that occurs during later fetal development. Sex-specific epi-marks stop girl fetuses from being masculinized when they experience atypically high testosterone, and vice versa for boy fetuses. Different epi-marks protect different sex-specific traits from being masculinized or feminized – some affect the genitals, others sexual identity, and yet others affect sexual partner preference. However, when these epi-marks are transmitted across generations from fathers to daughters or mothers to sons, they may cause reversed effects, such as the feminization of some traits in sons, such as sexual preference, and similarly a partial masculinization of daughters.

They say their study solves the evolutionary aspect of homosexuality, finding that "sexually antagonistic" epi-marks, which normally protect parents from natural variation in sex hormone levels during fetal development, sometimes carryover across generations and cause homosexuality in opposite-sex offspring. The mathematical modeling demonstrates that genes coding for these epi-marks can easily spread in the population because they always increase the fitness of the parent but only rarely escape erasure and reduce fitness in offspring. 

"Transmission of sexually antagonistic epi-marks between generations is the most plausible evolutionary mechanism of the phenomenon of human homosexuality," said co-author Sergey Gavrilets, Ph.D. in Physics and Mathematics and biomathematician at the University of Tennessee-Knoxville and the National Institute for Mathematical and Biological Synthesis (NIMBioS) associate director for scientific activities, in their statement

Paper: William R. Rice, Urban Friberg, and Sergey Gavrilets, 'Homosexuality as a Consequence of Epigenetically Canalized Sexual Development', The Quarterly Review of Biology Volume 87 Number 4  December 2012