Monday, 19 October 2015

It's just an epigenetic phase

There's an arresting moment in Michael Sandel's Harvard lecture series on Justice when he asks his audience who is an only child or the firstborn. Virtually everyone in the room puts up their hand, with the implication that younger sons and daughters are grinding away in some State School because they weren't smart enough to make the Ivy League. One implication of that is: parents lavish care and attention on Dau.I and run out of steam or interest or money for Dau.II, Dau.III and Son.IV.  The idea that some hormonal change makes second and subsequent children duller, if not exactly dullards, is hard to credit. On the other hand, you could note that rich well-connected families are small, while large families come from poorer neighbourhoods which have crappier schools. I think the status of this is that the observation is true but a widely accepted explanation is still awaited. If you a) haven't seen the Sandel Justice lectures b) have teenage children; then they are recommended.  We sat through a chunk of them before Dau.II left home as leaven from serial binge episodes of House, Desperate Housewives, Masterchef and The Wire.

It appears that birth order may affect whether you will grow up gay.  Seems that with each older brother blokes become a third more likely to love blokes. That doesn't mean that all fourth boys are gay; it means that if, say, 5% of #1s are gay then 6.6% of #2s are and 8.8% of #3s. If that's true, and we're still citing only that 2006 study, we're still adrift about the chemical / psychological /genetic mechanisms by which the phenomenon is achieved. Until the observation is incontrovertible, it is a waste of time to talk about, for one widely circulated example, the maternal immune system getting increasingly fed up with all the extra testosterone down there.  Now my pal El Asturiano has flagged a link in the most recent Nature which suggests that gay is probably not genetic but might be epigenetic. He also came up with the neat title to this post.

Epigenetics is a rather hot topic in genetics these past 5 or 10 years. Most famously in the investigations of the Dutch Hongerwinter of 1944-1945.  After the failure of Operation Market Garden [bloboprevious] when capturing Arnhem became A Bridge Too Far, occupied Netherlands suffered a terrible famine because the Germans diverted all available food to their core territory and the Allies were not in a position to make up the calorie deficit. A study in 2008 reported that children who were in utero during those desperately hard times grew up smaller than their siblings who were in process when their mother was comparatively better fed both before and after 1944. In particular they noted that the gene for insulin-like growth factor II (IGF2) was methylated in such a way as to damp down the effectiveness of its growth-promoting protein.  These epigenetic changes don't effect the underlying DNA but they definitely alter the way genes work . . . and the changes can persist into subsequent generations. You can imagine why such epigenetic control might be a good idea: if times are tough and belts are tight you need to be small and hardy to survive but the system needs to be flexible because cyclical good times favour fatter and taller offspring.  John Greally [see below] has written an informative editorial about the current status of epigenetic studies.

The genetics of gay has always been a bit of a conundrum because it is hard to work out how such a trait could be 'adaptive'. Gay people are surely less likely to have offspring than their heterosexual siblings because we are still ultimately dependent on one sperm and one egg to start the next generation and straight people can achieve this union with rather less trouble [5 minutes in the car-park outside a disco is all that is strictly necessary]. Everything-is-driven-by-selection people are therefore reduced to more subtle arguments about women benefitting if they have a gay brother on hand to help with the household chores and raising lots of nieces and nephews. A Nature report finding epigenetic tags to certain genes to be more common among homosexual men is the sort of thing that keeps the journal afloat as the general science journal of record.  Nature is fully commercial and much of the content is firmly behind their paywall, so material that is freely available may have slipped through the editorial process as too-sexy-to-turn-down. And even a cursory reading of the report should set your skepdar a-whirling because the key finding is based on only 37 hetero-homo identical twin pairs AND was only correctly predictive in 2/3rds of the cases. They report "UCLA computational geneticist Tuck Ngun [the lead author] will present the work on 8 October at the American Society of Human Genetics meeting in Baltimore, Maryland".  It's coincidental only in The Blob (because I read Natures out of sync) that just two weeks ago I was writing about a psychological study published by Nature that also allowed sexy to outweigh grossly under-powered to get something into print which should not have been published at all. And it would be good if the editors at Nature would reflect on their own failings when they take the high moral ground to criticise systemic failings in other areas of science and politics.

Well it seems that there was some robust criticism when these sketchy findings were presented to several hundred active researchers at ASHG15 rather than an editorial board that wanted a bit of spicy copy. Among them was John Greally who gave a really stonking talk to Irish Binfos two years ago in Galway.  He was at ASHG15 and blogs about the tearing asunder of the study, and also about the review process that should have stopped the publishing of such a weak preliminary study much earlier.  There's a great pop-sci summary of the methodological flaws in The Atlantic. It's really refreshing to have  a skeptical, clearly written critique of science and the flaws of the scientific process in a glossy magazine. There is another seam of comment and opinion on Metafilter. Now it may turn out that this little dataset is the first feather in the wind of our understanding the developmental under-pinnings of homosexual attraction; but on its own, as it stands, not yet replicated, it means . . . nothing.

1 comment:

  1. Interesting study review. First read begs the question why was the study on 'gay' men, rather than gay men and women? Isn't that like studying eye colour using one sex? Do scientists expect genetic traits to be different given a male/female subject. Or was it just that women aren't 'gay' enough?