Saletan does a nice job of explaining sexually antagonistic selection and poses some interesting Bioethics 101 type questions about how the gay community might react to the scientific idea that homosexuality (at least in males) is a 'side effect'.

The thing is, this paper suggests that among genetic theories, sexually antagonistic selection is the best at predicting observed fertility levels in a sample population. But each of those theories proposes that some relative of the homosexual benefits from the gene while the homosexual does not (from strictly a fertility perspective).

The other big theory in competition to sexually antagonistic selection is overdominance. In that case, the male relatives of a homosexual are more fertile.(In this case, Homosexuality is caused when one copy of a new gene makes a man straight and more fertile, while two copies make a man gay. Most of the actually theories involve variations where there are actually multiples genes of this type interacting, but that's the basic version).

With the maternal effects theories, a gene the mother carries 'feminizes' the developing fetus, which increases the fertility of female offspring and decreases the fertility of the male offspring.

Even the kinship selection theory, which has fallen out of favor in the scientific community, proposes a general benefit to the fertility of the relatives of the homosexual man, due to his increased involvement in rearing their children.

In that light, the questions Saletan poses are not as premature as they seem. These questions about how scientific genetic explanations for homosexuality won't go away if later science refutes this particular theory in favor of another one. But I really don't see this idea of homosexuality as a side effect of an increased fertility elsewhere as a particular threat. I don't expect a theory like this, even if overwhelmingly supported by scientific evidence, to affect the way gay people perceive themselves. Geneticists explain almost all human behavior in terms of a benefit to survival or fertility for someone. Even if, like me, you're a scientists who understands and generally accepts these types of explanations, it doesn't normally change the way we perceive our behavior. That too is a genetic trait that increases our survival advantage. :)

Well, my perception from reading your post is that you know this stuff better than I do. In which case, here's something I couldn't decipher from the paper: Are maternal-effects theories a likely bridge between genetic and hormonal theories (of homosexuality or other traits)? The question is on my mind because we're going to have to sort ot the various studies suggesting causality from different directions. Lay readers might be asking whether the previous hormonal/brain-scan study and this genetic analysis somehow discredit each other.

I don't know exactly what you mean by "bridge", but it could be that I suppose. Maternal-effects in the context of the original post is an effect on the child in the pre-natal environment statistically linked to the genes of the mother. Hormonal theory, as you describe it, is simply any effect in the prenatal environment, environmental or genetic. The genetic mechanism that maternal effects might influence hte fetus may be hormonal, or may be something else entirely. The difference in hormonal environment of the wombs of gay or not gay fetuses may be influenced by genetics, or not. More than one of these mechanisms appears to be at work. They aren't necessarily linked, although they do tend to confound those who are attempting to find statistical evidence.

Basically, long story short; brain scans and genetic analyses in this case do not discredit each other. They may be linked or they may not, but even if not, it simply means there are multiple factors that influence homosexuality.

I like this line in the original post:

"Even if, like me, you're a scientists who understands and generally accepts these types of explanations, it doesn't normally change the way we perceive our behavior."

I was a geneticist too in a past life, and it is often shocking how much moral value some people who aren't that familiar with genetics place upon genes. To a geneticist, genes are cool, but they aren't magical and they have nothing to do with right and wrong. They are just macromolecules. We're all stardust anyway.

Wow, thanks for the complement. It's not my field; I'm more of an enzymologist, but I know enough genetics to follow all the diagrams in the paper, if not all the math.

To address your question, I simplified my descriptions of the maternal effects in order to make my broader point. Ciani et. al. mention two different types of maternal effects. One is referenced from the birth order study which makes an immune hypothesis that each successive male child increasingly sensitizes the mother's immune system to male antigens, resulting in a response that affects the brains of the youngest male fetuses. It's certainly likely that the degree to which a mother becomes immunized to these types of male antigens is genetic, at least partly. But the genetic role of a mechanism like this would be particularly hard to elucidate, because of the delicate interplay of environmental and genetic factors. Two genetically identical women (twin sisters) could each have a gene that makes their immune systems more suseptible to this type of effect, but if one had three sons and one had a son and two daughters, a geneticist wouldn't learn anything from that example. So in order to do the sort of statistical analysis that Ciani and coworkers did, you'd need a much, much larger sample size to learn anything of value. Given that the maternal immune hypothesis has yet to be firmly established, it'd be a long time before anyone would think about studying its genetics.

Maternal genomic imprinting is more complicated. Genomic imprinting is a specific epigenetic phenomena. Epigentic effects are those where some heritable trait is passed on to offspring, without a change in the actual DNA code. In normal Mendelian genetics, you get a copy of each gene from each parent and it doesn't matter where the copies came from. In genomic imprinting, it does. So maternal genomic imprinting is where a gene only causes a trait when it's inherited from the mother. So two individuals with the same genotype Aa (where A is a gene that contributes to homosexuality and a contributes to heterosexuality) could have different phenotypes if one inherited the A from his mother and another inherited the A from his father. The study Ciani and coworkers cite regarding this phenomena looked at families with two or more homosexual sons and looked for certain genetic regions that showed up in a statistically higher than normal proportion. One of these regions was always inheritened from the mothers, which suggested this type of effect.

So you see, my brief ealier description of these effects grossly oversimplified them. It's possible that the maternally imprinted genes result in the homosexual phenotype through some sort of homonal mechanism. The region they found contained some neurotransmitter interacting proteins, so it might be possible to someday link this phenomena to observable differences in the brain, be they caused by hormones or something else.

But it's also possible that the mechanism for production of phenotype in many other genetic hypotheses involves hormones. For example the 'androphilic gene' proposed by sexually antagonistic selection could work by affecting the balance of hormones in the fetal brain to alter the amygdala increase attraction to males. While the scientists who studied the amygdala believe that because they see the phenomena in lesbians as well as gay men that it isn't genetic, it is possible that a genetic effect in gay men produces the same morpological changes as a nongenetic effect in lesbians.

The weak link in all these sorts of studies is the biochemistry. We can study brains of gays and lesbians to look for morphological differences, but we don't really know what causes those differences. We can study genomes to determine things about the heritablity, but we don't know when and how those genes have their effect. Scientists believe the key events occur due to specific biochemical events that impact the neurology during fetal development, and obviously that's a very difficult thing to study directly.