A gene for gay?
October 10, 2015 11:35 AM Subscribe
A controversial talk by Tuck Ngun at the ongoing American Society of Human Genetics 2015 meeting in Baltimore presented evidence of epigenetic mechanisms associated with homosexuality in discordant male twins (i.e., one gay, the other straight). The conference organizers and news outlets quickly trumpeted that scientists had discovered epigenetic markers capable of predicting the sexual orientation of a male; however, the reaction of scientists at the meeting was less enthusiastic. Ed Yong at the Atlantic wrote a particularly thorough takedown. Criticisms centered around the small sample size (37 pairs of twins), the fact that the samples were taken from saliva (whereas you'd expect epigenetic variants influencing sexuality to occur in the brain), and the fact that the predictive model they developed was not terribly predictive (67% accuracy).
As an attendee at the talk, it was interesting to watch the story evolve into a discussion about the role of journalism in science, especially when the findings in question have not yet been subject to peer review.
Tuck has handled the criticisms admirably and stands by his work.
As an attendee at the talk, it was interesting to watch the story evolve into a discussion about the role of journalism in science, especially when the findings in question have not yet been subject to peer review.
Tuck has handled the criticisms admirably and stands by his work.
More snark on Leonid Kruglyak's twitter stream. This tweet and the ones following.
posted by quaking fajita at 11:45 AM on October 10, 2015
posted by quaking fajita at 11:45 AM on October 10, 2015
I am not a geneticist and I am confused. Is the backlash exclusively over the lack of rigor/shoddy methodology of the study (i.e., bad science) or is there something else going on?
posted by echocollate at 11:57 AM on October 10, 2015 [2 favorites]
posted by echocollate at 11:57 AM on October 10, 2015 [2 favorites]
I've been seeing a bunch of cranky commentary on this on my Twitter this weekend--thanks for bringing it here! I agree, that study size is ridiculously underpowered for what he's trying to do.
Incidentally, there's another really good, thorough discussion of the flaws in the work and in epigenetics reporting more generally here. I find it really interesting that this is getting so much more public pushback--and faster pushback!--than I remember being the case for the Rice paper, also on epigenetics and homosexuality, that came out a few years ago and was also massively publicized. I recall quite a bit of criticism of that but it seemed much quieter than the quick, decisive public-science pushback on this thing.
Echocollate: the backlash is actually in part about the implications of being able to detect a mechanistic marker of sexual orientation. Those are big claims with a lot of cultural weight, and this study just doesn't have the power to make that statement. It's also about general problems with reporting of epigenetics in the popular media and about what constitutes an "acceptable" epigenetic study on behavior.
My labmates actually work on an epigenetics-and-behavior system, although *I* work on something different (we have two systems in my lab and I'm working with the other one). I'm kind of surprised we haven't had a chance to talk this over at lab meeting, since it dovetails very neatly with the interests of several of our lab members. I might bring it up next week and ask the opinions of some of my more-into-epigenetic-marking-than-me colleagues.
posted by sciatrix at 12:03 PM on October 10, 2015 [11 favorites]
Incidentally, there's another really good, thorough discussion of the flaws in the work and in epigenetics reporting more generally here. I find it really interesting that this is getting so much more public pushback--and faster pushback!--than I remember being the case for the Rice paper, also on epigenetics and homosexuality, that came out a few years ago and was also massively publicized. I recall quite a bit of criticism of that but it seemed much quieter than the quick, decisive public-science pushback on this thing.
Echocollate: the backlash is actually in part about the implications of being able to detect a mechanistic marker of sexual orientation. Those are big claims with a lot of cultural weight, and this study just doesn't have the power to make that statement. It's also about general problems with reporting of epigenetics in the popular media and about what constitutes an "acceptable" epigenetic study on behavior.
My labmates actually work on an epigenetics-and-behavior system, although *I* work on something different (we have two systems in my lab and I'm working with the other one). I'm kind of surprised we haven't had a chance to talk this over at lab meeting, since it dovetails very neatly with the interests of several of our lab members. I might bring it up next week and ask the opinions of some of my more-into-epigenetic-marking-than-me colleagues.
posted by sciatrix at 12:03 PM on October 10, 2015 [11 favorites]
True. We only had so much space in the abstract. But to address the point more directly, there is almost no way of getting to the best tissue (the brain) for the population we were interested in. I would have been much happier with the ‘right’ tissue type.
That's a fair point. You have a project, say, where gene expression values are measured for glioblastoma, except that very few get permission to take samples of healthy brain tissue. So when we're comparing gene expression of healthy and cancerous tissues, the sample sizes are so wildly unbalanced that you can't really say much that's useful. And yet no one is calling for NIH's funding to get revoked, etc. I'm not sure that this reality is Ngun's fault — sometimes you do the experiment with the data you can collect, prepare and measure it with the best available practices, and you hope that future iterations offer improved data sources or collection methods, if there is potential for something interesting.
posted by a lungful of dragon at 12:13 PM on October 10, 2015 [6 favorites]
That's a fair point. You have a project, say, where gene expression values are measured for glioblastoma, except that very few get permission to take samples of healthy brain tissue. So when we're comparing gene expression of healthy and cancerous tissues, the sample sizes are so wildly unbalanced that you can't really say much that's useful. And yet no one is calling for NIH's funding to get revoked, etc. I'm not sure that this reality is Ngun's fault — sometimes you do the experiment with the data you can collect, prepare and measure it with the best available practices, and you hope that future iterations offer improved data sources or collection methods, if there is potential for something interesting.
posted by a lungful of dragon at 12:13 PM on October 10, 2015 [6 favorites]
Previously re: the Rice paper. God, that is some painful framing about the origins of homosexuality in the title of that FPP, which is understandable since the written-for-a-general-audience piece isn't much better. In a lot of ways, they're really similar papers--this new one is actually a bit better because he went looking for actual epigenetic data, whereas the Rice paper is basically a combination journal review and mathematical model. Models are useful and all, but they don't actually tell you anything about what happens, just whether it's plausible that a given theoretical concept might work the way you think. (Which is totally useful! But "Scientists prove that homosexuality is caused by epigenetic markings" it is not.)
I'm going to be curious to read how that old FPP went, too. I wasn't here then, and I kind of want to know how the discussion went differently. One big thing I think might be different is that Carl Zimmer, who is a very eminent popular science writer, has been very active about promoting criticisms of this new Ngun presentation on Twitter. I wonder if that's having the effect of dampening the enthusiastic "scientists said we know where homosexuality comes from!" rush that popular media tends to engage in for stories like this.
posted by sciatrix at 12:15 PM on October 10, 2015
I'm going to be curious to read how that old FPP went, too. I wasn't here then, and I kind of want to know how the discussion went differently. One big thing I think might be different is that Carl Zimmer, who is a very eminent popular science writer, has been very active about promoting criticisms of this new Ngun presentation on Twitter. I wonder if that's having the effect of dampening the enthusiastic "scientists said we know where homosexuality comes from!" rush that popular media tends to engage in for stories like this.
posted by sciatrix at 12:15 PM on October 10, 2015
FTFA:
> whereas you'd expect epigenetic variants influencing sexuality to occur in the brain
I'd actually expect science to analyze data without a preference for what they hope to see. No one expected tooth cavities to play a role in heart disease, and yet the data proves they do.
This complaint is without merit.
the fact that the predictive model they developed was not terribly predictive (67% accuracy)
Dead stupid complaint. ANY statistically significant reliability in the predictive model is potentially important. FFS, that's actually a pretty high number, in fact. If only one-in-six homosexuals were affected by this epigenetic marker (and given the huge diversity of human sexuality, that is certainly possible), it would still have meaning.
There are serious issues with the study - most of which are openly admitted by the study authors, like its small size and the earliness of this investigation into the topic at all. That doesn't mean the science is bad; on the contrary, it means the science warrants further study. They realize their results might not be repeatable. That's what "small sample size" means.
The authors aren't claiming this will make their patented Gay-O-Meter device work. Critics seem to be falling over each other to prove the non-existent claims are overstated.
posted by IAmBroom at 12:31 PM on October 10, 2015 [15 favorites]
> whereas you'd expect epigenetic variants influencing sexuality to occur in the brain
I'd actually expect science to analyze data without a preference for what they hope to see. No one expected tooth cavities to play a role in heart disease, and yet the data proves they do.
This complaint is without merit.
the fact that the predictive model they developed was not terribly predictive (67% accuracy)
Dead stupid complaint. ANY statistically significant reliability in the predictive model is potentially important. FFS, that's actually a pretty high number, in fact. If only one-in-six homosexuals were affected by this epigenetic marker (and given the huge diversity of human sexuality, that is certainly possible), it would still have meaning.
There are serious issues with the study - most of which are openly admitted by the study authors, like its small size and the earliness of this investigation into the topic at all. That doesn't mean the science is bad; on the contrary, it means the science warrants further study. They realize their results might not be repeatable. That's what "small sample size" means.
The authors aren't claiming this will make their patented Gay-O-Meter device work. Critics seem to be falling over each other to prove the non-existent claims are overstated.
posted by IAmBroom at 12:31 PM on October 10, 2015 [15 favorites]
Why do scientists bother doing a study that they know at the outset will be pointless? If you know that your sample size is too small for your results to be meaningful, why bother?
As the Atlantic takedown points out: The sample size here, and use of multiple methods, and the number of genetic markers mean that even if one of the models had been 100% predictive it still would have been statistically meaningless. Once you start looking for patterns then you are statistically more likely to find a pattern than not. So it was doomed to failure before they even had the data. Why bother?
They just wanted a publication.
posted by mary8nne at 12:32 PM on October 10, 2015
As the Atlantic takedown points out: The sample size here, and use of multiple methods, and the number of genetic markers mean that even if one of the models had been 100% predictive it still would have been statistically meaningless. Once you start looking for patterns then you are statistically more likely to find a pattern than not. So it was doomed to failure before they even had the data. Why bother?
They just wanted a publication.
posted by mary8nne at 12:32 PM on October 10, 2015
ANY statistically significant reliability in the predictive model is potentially important.
Sure, except they didn't find any statistically significant differences.
posted by quaking fajita at 12:35 PM on October 10, 2015 [1 favorite]
Sure, except they didn't find any statistically significant differences.
posted by quaking fajita at 12:35 PM on October 10, 2015 [1 favorite]
The authors aren't claiming this will make their patented Gay-O-Meter device work.
The news writers certainly think they are, which is why they are in the news instead of at the bottom of a footnote in a journal read only by members of their profession.
posted by Bringer Tom at 12:37 PM on October 10, 2015 [1 favorite]
The news writers certainly think they are, which is why they are in the news instead of at the bottom of a footnote in a journal read only by members of their profession.
posted by Bringer Tom at 12:37 PM on October 10, 2015 [1 favorite]
Why bother?
They just wanted a publication.
Well, okay, except this fellow didn't actually get a publication out of this. This is a conference talk, which hasn't yet undergone peer review and is a considerably less formal venue. It's also not entirely uncommon for somewhat preliminary or otherwise incomplete data to be presented at conferences.
....I must admit, I totally do feel kind of sorry for this guy and am impressed by how well he's handling this. It looks like he's got a pretty solid idea of what this data can and can't do, and he's not actually the one blowing it too much out of proportion. Would I be in his place, though? Not if you paid me--there's a reason I don't work on humans.
posted by sciatrix at 12:40 PM on October 10, 2015 [14 favorites]
They just wanted a publication.
Well, okay, except this fellow didn't actually get a publication out of this. This is a conference talk, which hasn't yet undergone peer review and is a considerably less formal venue. It's also not entirely uncommon for somewhat preliminary or otherwise incomplete data to be presented at conferences.
....I must admit, I totally do feel kind of sorry for this guy and am impressed by how well he's handling this. It looks like he's got a pretty solid idea of what this data can and can't do, and he's not actually the one blowing it too much out of proportion. Would I be in his place, though? Not if you paid me--there's a reason I don't work on humans.
posted by sciatrix at 12:40 PM on October 10, 2015 [14 favorites]
From one of the links that sciatrix posted:
What is not described in the abstract is that the cells used were from saliva samples, which includes a variable mixture of buccal epithelium with a majority of leukocytes. The presence of microbial DNA also has the potential to cross-hybridise to human probes on a microarray, so the screening approach used could be criticised for several potential technical flaws. Also not described is the marginal, uncorrected significance for this underpowered study. These only came to light when the presentation happened.
In other words, because the human mouth is a microbe party, other DNA could be making it into the array that they're analyzing?
As a non-scientist, reading all of this made me go "Well, gee, that's an awfully small sample size for an analysis of the complex stew that is human sexuality."
The news writers certainly think they are, which is why they are in the news instead of at the bottom of a footnote in a journal read only by members of their profession.
Yeah, in terms of the initial media pickup this is reminiscent of Simon LeVay's A difference in hypothalamic structure between heterosexual and homosexual men (pdf here), which got breathless coverage of the "biological basis for homosexuality found" or some such when it was published.
posted by mandolin conspiracy at 12:41 PM on October 10, 2015 [1 favorite]
What is not described in the abstract is that the cells used were from saliva samples, which includes a variable mixture of buccal epithelium with a majority of leukocytes. The presence of microbial DNA also has the potential to cross-hybridise to human probes on a microarray, so the screening approach used could be criticised for several potential technical flaws. Also not described is the marginal, uncorrected significance for this underpowered study. These only came to light when the presentation happened.
In other words, because the human mouth is a microbe party, other DNA could be making it into the array that they're analyzing?
As a non-scientist, reading all of this made me go "Well, gee, that's an awfully small sample size for an analysis of the complex stew that is human sexuality."
The news writers certainly think they are, which is why they are in the news instead of at the bottom of a footnote in a journal read only by members of their profession.
Yeah, in terms of the initial media pickup this is reminiscent of Simon LeVay's A difference in hypothalamic structure between heterosexual and homosexual men (pdf here), which got breathless coverage of the "biological basis for homosexuality found" or some such when it was published.
posted by mandolin conspiracy at 12:41 PM on October 10, 2015 [1 favorite]
In other words, because the human mouth is a microbe party, other DNA could be making it into the array that they're analyzing?
Ngun: "We used a sequencing approach (RRBS) not arrays. We did find microbial DNA sequences which we filtered out. Some could have gotten through I suppose."
posted by a lungful of dragon at 12:48 PM on October 10, 2015 [3 favorites]
Ngun: "We used a sequencing approach (RRBS) not arrays. We did find microbial DNA sequences which we filtered out. Some could have gotten through I suppose."
posted by a lungful of dragon at 12:48 PM on October 10, 2015 [3 favorites]
It looks like he's got a pretty solid idea of what this data can and can't do
I would be more sympathetic towards the authors if it weren't for the conflation of training and test sets when they were doing model selection. (apparently, I am not at the meeting and am just working off of what people are saying) That's like machine learning rule #1
posted by quaking fajita at 12:54 PM on October 10, 2015
I would be more sympathetic towards the authors if it weren't for the conflation of training and test sets when they were doing model selection. (apparently, I am not at the meeting and am just working off of what people are saying) That's like machine learning rule #1
posted by quaking fajita at 12:54 PM on October 10, 2015
I don't feel sorry for them. The reason for doing this kind of study when you know the results will be useless is that you want attention, and they got it.
As I have written with regard to other nature/nurture arguments, the whole idea of twin studies should be discredited for at least a century. The fact is there is only one possible interesting result from a twin study, and the fact that you are doing the study predisposes you to hope for it. Anyone who voluntarily puts their name on a twin study should be hounded by the ghosts of every single person who was forcibly sterilized, institutionalized, or denied educational opportunities because of that evil hack Cyril Burt shouting SHAME into their ears until they go mad.
posted by Bringer Tom at 1:02 PM on October 10, 2015 [1 favorite]
As I have written with regard to other nature/nurture arguments, the whole idea of twin studies should be discredited for at least a century. The fact is there is only one possible interesting result from a twin study, and the fact that you are doing the study predisposes you to hope for it. Anyone who voluntarily puts their name on a twin study should be hounded by the ghosts of every single person who was forcibly sterilized, institutionalized, or denied educational opportunities because of that evil hack Cyril Burt shouting SHAME into their ears until they go mad.
posted by Bringer Tom at 1:02 PM on October 10, 2015 [1 favorite]
It looks like he's got a pretty solid idea of what this data can and can't do
Well, yeah thats what's so shocking, he says that he knew, going into the experiment that it Wouldn't produce statistically significant results. That's what I find so sisyphean. He knew that the methodology was totally flawed, but went ahead anyway.
posted by mary8nne at 1:04 PM on October 10, 2015
Well, yeah thats what's so shocking, he says that he knew, going into the experiment that it Wouldn't produce statistically significant results. That's what I find so sisyphean. He knew that the methodology was totally flawed, but went ahead anyway.
posted by mary8nne at 1:04 PM on October 10, 2015
One of the basic logics of scientific investigation is exploring variation, so in that sense it's reasonable to try to figure out why some people are 'straight' and others are 'gay.' However, given that we don't really know why (or maybe more to the point, how) people have sexual orientation itself in the first place, or more radically, if that's an externally valid construct anyway, it seems premature to look for the explanation of the 'gay' orientation. On the other hand, one could also say that that's exactly why investigation is warranted, but it seems very important to do that investigation without inadvertently reifying or assuming the validity of the thing that's being investigated.
posted by clockzero at 1:24 PM on October 10, 2015 [4 favorites]
posted by clockzero at 1:24 PM on October 10, 2015 [4 favorites]
I'd actually expect science to analyze data without a preference for what they hope to see. No one expected tooth cavities to play a role in heart disease, and yet the data proves they do.
Lemme guess, you aren't a scientist? Because lemme tell you, its never done that way. You are always testing hypotheses, selecting among a variety of models, etc. There is a lot of subjectivity in the actual doing of science.
posted by MisantropicPainforest at 2:16 PM on October 10, 2015 [1 favorite]
Lemme guess, you aren't a scientist? Because lemme tell you, its never done that way. You are always testing hypotheses, selecting among a variety of models, etc. There is a lot of subjectivity in the actual doing of science.
posted by MisantropicPainforest at 2:16 PM on October 10, 2015 [1 favorite]
They realize their results might not be repeatable. That's what "small sample size" means.
Ugh, no, that's not what that means. It means the tests will have low power.
posted by MisantropicPainforest at 2:18 PM on October 10, 2015 [4 favorites]
Ugh, no, that's not what that means. It means the tests will have low power.
posted by MisantropicPainforest at 2:18 PM on October 10, 2015 [4 favorites]
Let's posit a world in which homosexuality is considered acceptable only because it is considered to be an unfortunate birth defect. If a way is found to prevent those births the world might well choose to do so, rationalizing "Why would you let a child be born with such a horrible disease?"
In our world, it may well be that widespread social acceptance of gay people just barely wins the race against the technology to eliminate it. Interesting times.
posted by LastOfHisKind at 3:10 PM on October 10, 2015 [2 favorites]
In our world, it may well be that widespread social acceptance of gay people just barely wins the race against the technology to eliminate it. Interesting times.
posted by LastOfHisKind at 3:10 PM on October 10, 2015 [2 favorites]
It means the tests will have low power.
It's been awhile since I was awash in this math, but the back of my notepad suggests that with their methodology they have about a twenty percent chance of getting the result they got even if the distributions are all completely random. There are several ways to approach it none of which are just like the casino games I cut my statistics teeth on, but that looks like a pretty solid estimate by several approaches.
Now in the absence of other studies or other factors to sharpen the results, that makes the results completely useless. In fact, it means that it should have been clear before they wasted the time, money, and energy to do the study that there was no possibility of a meaningful result. It is difficult to overstate how useless this result is. It is at best a two-sigma result. Two sigma anomalies happen all the time. If you have ever driven up to a busy intersection with turn arrows in all directions and found it green for you, and you drew the conclusion that stop lights tend to be green, you're these guys.
This is why I have a rather uncharitable view of the researchers. If they didn't realize how useless their study was then they were very stupid, and if they did then they knew the kind of attention they were looking for and that makes them very evil.
posted by Bringer Tom at 3:24 PM on October 10, 2015 [2 favorites]
It's been awhile since I was awash in this math, but the back of my notepad suggests that with their methodology they have about a twenty percent chance of getting the result they got even if the distributions are all completely random. There are several ways to approach it none of which are just like the casino games I cut my statistics teeth on, but that looks like a pretty solid estimate by several approaches.
Now in the absence of other studies or other factors to sharpen the results, that makes the results completely useless. In fact, it means that it should have been clear before they wasted the time, money, and energy to do the study that there was no possibility of a meaningful result. It is difficult to overstate how useless this result is. It is at best a two-sigma result. Two sigma anomalies happen all the time. If you have ever driven up to a busy intersection with turn arrows in all directions and found it green for you, and you drew the conclusion that stop lights tend to be green, you're these guys.
This is why I have a rather uncharitable view of the researchers. If they didn't realize how useless their study was then they were very stupid, and if they did then they knew the kind of attention they were looking for and that makes them very evil.
posted by Bringer Tom at 3:24 PM on October 10, 2015 [2 favorites]
There are plenty of discussions on the mermaids tale about this stuff.
Man, are twin studies ever suspect.
posted by clvrmnky at 3:36 PM on October 10, 2015
Man, are twin studies ever suspect.
posted by clvrmnky at 3:36 PM on October 10, 2015
Criticisms centered around the small sample size (37 pairs of twins), the fact that the samples were taken from saliva (whereas you'd expect epigenetic variants influencing sexuality to occur in the brain), and the fact that the predictive model they developed was not terribly predictive (67% accuracy).
The very least they should do is p-hack to get 69%.
posted by srboisvert at 4:25 PM on October 10, 2015 [1 favorite]
The very least they should do is p-hack to get 69%.
posted by srboisvert at 4:25 PM on October 10, 2015 [1 favorite]
Now in the absence of other studies or other factors to sharpen the results, that makes the results completely useless.
I mean, it's useless for the sake of making grand pronouncements about finding "genes for gays" or "methylation sites for gays," or whatever, sure. But bad reporting aside, this is clearly a pilot study -- the point of that type of study isn't to be definitive, it's to gather some preliminary data that could go into a grant that could actually fund the study that would be large enough to confirm whether there is any effect. If there were really training/test set contamination, that's a serious problem (the author denies that so it might have just been a misimpression from the talk), but assuming there wasn't, p = 0.03 on a small study isn't worthless for the purposes of suggesting further investigation.
(BTW, p = 0.03 is a test of the entire classifier predicting better than random, not a test of the individual associations; multiple-testing isn't relevant in that case. It would be relevant if there was training/test contamination and the classifier was retrained a bunch of times.)
posted by en forme de poire at 5:33 PM on October 10, 2015 [3 favorites]
I mean, it's useless for the sake of making grand pronouncements about finding "genes for gays" or "methylation sites for gays," or whatever, sure. But bad reporting aside, this is clearly a pilot study -- the point of that type of study isn't to be definitive, it's to gather some preliminary data that could go into a grant that could actually fund the study that would be large enough to confirm whether there is any effect. If there were really training/test set contamination, that's a serious problem (the author denies that so it might have just been a misimpression from the talk), but assuming there wasn't, p = 0.03 on a small study isn't worthless for the purposes of suggesting further investigation.
(BTW, p = 0.03 is a test of the entire classifier predicting better than random, not a test of the individual associations; multiple-testing isn't relevant in that case. It would be relevant if there was training/test contamination and the classifier was retrained a bunch of times.)
posted by en forme de poire at 5:33 PM on October 10, 2015 [3 favorites]
the backlash is actually in part about the implications of being able to detect a mechanistic marker of sexual orientation. Those are big claims with a lot of cultural weight, and this study just doesn't have the power to make that statement
At the same time, though, the author actually doesn't (seem to?) claim that these are causal, mechanistic associations, merely that they have some plausibility from previous work -- it seem to me that the point of bringing them up is mainly to show that the classifier isn't based on things that look like complete nonsense.
I understand being weary of epigenetics reporting (believe me, as someone who works on the microbiome, I am familiar with having my field be oversold and distorted by mainstream news outlets and alt-med hucksters!), but at the same time, the reality seems to me to be that without pilot studies, especially in the funding environment in which we live, you quash the possibility of even having any bigger, more definitive studies. I certainly agree that it is possible to speculate irresponsibly from limited data, but it really doesn't sound like that's what was done here.
posted by en forme de poire at 5:55 PM on October 10, 2015 [5 favorites]
At the same time, though, the author actually doesn't (seem to?) claim that these are causal, mechanistic associations, merely that they have some plausibility from previous work -- it seem to me that the point of bringing them up is mainly to show that the classifier isn't based on things that look like complete nonsense.
I understand being weary of epigenetics reporting (believe me, as someone who works on the microbiome, I am familiar with having my field be oversold and distorted by mainstream news outlets and alt-med hucksters!), but at the same time, the reality seems to me to be that without pilot studies, especially in the funding environment in which we live, you quash the possibility of even having any bigger, more definitive studies. I certainly agree that it is possible to speculate irresponsibly from limited data, but it really doesn't sound like that's what was done here.
posted by en forme de poire at 5:55 PM on October 10, 2015 [5 favorites]
Yeah, I honestly think a lot of the furor around this talk is based on the distortions and overselling of PR outlets and mainstream news outlets... which has been happening in epigenetics for a long time, and is in no way unique to this particular topic. (I'm in sexual selection and I work with people who study monogamy, so I'm just as used to seeing media grab studies and go nuts--which might be coloring my reaction.) And then serious scientists react to the distortions and get pissy, even if they're not necessarily statements that Ngun himself is making.
There's a huge problem with mainstream science reporting more generally and I suspect that's a big part of this public discussion. Especially when anything touching on sexual behavior and humans (or even sometimes sexual behavior not in humans!) is concerned, I see a lot of incredibly rotten overblown claims about what a given study can tell us about humans... which naturally pisses off scientists, particularly those who bother to read the actual papers that come out.
Assuming he hasn't biased the training of his predictive algorithms--which, I agree with en forme de poire there, it seems to be hard to tell without seeing the talk--this is totally a level of data that makes sense to appear at a talk. That's part of the point of talks at academic conferences--get feedback from other people on ways to improve your work! Figure out how interesting people think your work is! Pitch new ideas you've been kicking around! Network and see who else might be interested in what you want to do! As far as I can tell, Ngun is currently a postdoc in a stage of fairly intense career flux, and that means it's critical for him to get himself out to these conferences and try to get people excited about what he's doing. If it hasn't been possible for him to get the amount of data he wanted before now--which seems pretty likely to me given his statements--I can see why he opted to go ahead and present the results of his pilot anyway.
I gotta say, the commentary over whether the talk "should have been accepted" at an international meeting (per the "discussion about the role of journalism" link) also has me a little weirded out. My first ever talk at an international meeting was as an undergrad. I'd lucked out and gotten an interesting result on my undergrad thesis and it had helped my PI get some grant money, so she brought me along to an Evolution meeting right before I left her lab. My experience with conference vetting is that the organizers only really see abstracts ahead of time and don't really vet them in much detail before accepting the talk--hell, I routinely see people write up the abstract for data that hasn't yet actually been achieved and then scramble to get it in before the talk goes up. Is the vetting in human genetics just that much more aggressive, or am I missing something?
posted by sciatrix at 6:33 PM on October 10, 2015 [10 favorites]
There's a huge problem with mainstream science reporting more generally and I suspect that's a big part of this public discussion. Especially when anything touching on sexual behavior and humans (or even sometimes sexual behavior not in humans!) is concerned, I see a lot of incredibly rotten overblown claims about what a given study can tell us about humans... which naturally pisses off scientists, particularly those who bother to read the actual papers that come out.
Assuming he hasn't biased the training of his predictive algorithms--which, I agree with en forme de poire there, it seems to be hard to tell without seeing the talk--this is totally a level of data that makes sense to appear at a talk. That's part of the point of talks at academic conferences--get feedback from other people on ways to improve your work! Figure out how interesting people think your work is! Pitch new ideas you've been kicking around! Network and see who else might be interested in what you want to do! As far as I can tell, Ngun is currently a postdoc in a stage of fairly intense career flux, and that means it's critical for him to get himself out to these conferences and try to get people excited about what he's doing. If it hasn't been possible for him to get the amount of data he wanted before now--which seems pretty likely to me given his statements--I can see why he opted to go ahead and present the results of his pilot anyway.
I gotta say, the commentary over whether the talk "should have been accepted" at an international meeting (per the "discussion about the role of journalism" link) also has me a little weirded out. My first ever talk at an international meeting was as an undergrad. I'd lucked out and gotten an interesting result on my undergrad thesis and it had helped my PI get some grant money, so she brought me along to an Evolution meeting right before I left her lab. My experience with conference vetting is that the organizers only really see abstracts ahead of time and don't really vet them in much detail before accepting the talk--hell, I routinely see people write up the abstract for data that hasn't yet actually been achieved and then scramble to get it in before the talk goes up. Is the vetting in human genetics just that much more aggressive, or am I missing something?
posted by sciatrix at 6:33 PM on October 10, 2015 [10 favorites]
But when what you’ve got are the makings of a fatally weak study, of the kind well known to cause problems in a field, it really is an option—perhaps the best option—to not do it at all.
This is terrible advice and I am disappointed that some professional scientists would espouse this kind of flawed logic. To me this says that more scientists should study philosophy of science.
posted by polymodus at 6:40 PM on October 10, 2015 [3 favorites]
This is terrible advice and I am disappointed that some professional scientists would espouse this kind of flawed logic. To me this says that more scientists should study philosophy of science.
posted by polymodus at 6:40 PM on October 10, 2015 [3 favorites]
When you adopt the methods of discredited charlatans you need to be prepared to answer the accusation that you have adopted the methods of discredited charlatans and to show how you are not doing what the discredited charalatans did.
This study did not take those precautions.
It is also a certainty that if you do a study on certain topics such as homosexuality or climate change that you will attract the attention of journalists, who will jump to hasty conclusions. This study was very poorly designed to do anything scientifically useful but it was very well designed to attract the attention of journalists. I suspect it worked much better in that regard than the researchers expected which is why some here feel a little sorry for them.
Don't feel sorry for them. They are getting exactly what they asked for, and they will live with it for the rest of their careers and lives, as they should.
posted by Bringer Tom at 6:50 PM on October 10, 2015
This study did not take those precautions.
It is also a certainty that if you do a study on certain topics such as homosexuality or climate change that you will attract the attention of journalists, who will jump to hasty conclusions. This study was very poorly designed to do anything scientifically useful but it was very well designed to attract the attention of journalists. I suspect it worked much better in that regard than the researchers expected which is why some here feel a little sorry for them.
Don't feel sorry for them. They are getting exactly what they asked for, and they will live with it for the rest of their careers and lives, as they should.
posted by Bringer Tom at 6:50 PM on October 10, 2015
The research consortium my lab is a part of put out a series of high-profile papers in Nature a couple years back. I see a lot of similarities in the degree of visceral hatred in the response to our work, where the criticism about the science became tangential to going after the people involved — to the extent where, at one point, my PI got called a racist on a bioinformatics form without any reason or provocation. I have to admit that I'm getting really uncomfortable with how people who are really just jealous (for lack of any better word) of the level of media coverage are using Twitter and other social forums to become cranky personalities who level personal attacks on scientists. Not what I joined the field to be a part of. Ugly stuff. Gross. Just my two cents.
posted by a lungful of dragon at 7:31 PM on October 10, 2015 [5 favorites]
posted by a lungful of dragon at 7:31 PM on October 10, 2015 [5 favorites]
When you adopt the methods of discredited charlatans you need to be prepared to answer the accusation that you have adopted the methods of discredited charlatans and to show how you are not doing what the discredited charalatans did.
Please be more specific about what you mean here, in particular about the ways in which you think the study used "the methods of discredited charlatans."
posted by en forme de poire at 7:49 PM on October 10, 2015 [2 favorites]
Please be more specific about what you mean here, in particular about the ways in which you think the study used "the methods of discredited charlatans."
posted by en forme de poire at 7:49 PM on October 10, 2015 [2 favorites]
The fact is there is only one possible interesting result from a twin study, and the fact that you are doing the study predisposes you to hope for it.
And if this is what you are referring to, what exactly do you suppose this "one possible interesting result" to be?
posted by en forme de poire at 7:56 PM on October 10, 2015 [1 favorite]
And if this is what you are referring to, what exactly do you suppose this "one possible interesting result" to be?
posted by en forme de poire at 7:56 PM on October 10, 2015 [1 favorite]
And if this is what you are referring to, what exactly do you suppose this "one possible interesting result" to be?
You expect correlation, obviously. There would be no point in doing the twin study if you didn't expect correlation. If you don't get correlation it is essentially a negative result -- uninteresting and boring. You haven't discovered anything. But if you get correlation, you have discovered something. You have explained something. You have moved something from the realm of the mysterious to the known.
It is a fundamental problem of twin studies that they will be plagued by small sample size and uncontrolled cofactors. You cannot get away from that, especially when you are looking into messy complicated epiphenomena like intelligence or sexual orientation. It's very obvious without even looking at the results that this study is a complete waste of time. There was never any possibility of a result with real scientific value. But suggest that you might have a pointer however tenuous to the Gay GeneTM, and here we are talking about it on Metafilter.
That was the only possible useful result. Don't kid yourself that these guys had anything else in mind. And when you go into a study wanting a result, it becomes a lot more likely that you will get it.
posted by Bringer Tom at 6:30 AM on October 11, 2015
You expect correlation, obviously. There would be no point in doing the twin study if you didn't expect correlation. If you don't get correlation it is essentially a negative result -- uninteresting and boring. You haven't discovered anything. But if you get correlation, you have discovered something. You have explained something. You have moved something from the realm of the mysterious to the known.
It is a fundamental problem of twin studies that they will be plagued by small sample size and uncontrolled cofactors. You cannot get away from that, especially when you are looking into messy complicated epiphenomena like intelligence or sexual orientation. It's very obvious without even looking at the results that this study is a complete waste of time. There was never any possibility of a result with real scientific value. But suggest that you might have a pointer however tenuous to the Gay GeneTM, and here we are talking about it on Metafilter.
That was the only possible useful result. Don't kid yourself that these guys had anything else in mind. And when you go into a study wanting a result, it becomes a lot more likely that you will get it.
posted by Bringer Tom at 6:30 AM on October 11, 2015
Also, if you do not know who Cyril Burt is and what he did, you should not be allowed to do a twin study at all.
posted by Bringer Tom at 6:32 AM on October 11, 2015
posted by Bringer Tom at 6:32 AM on October 11, 2015
I'm glad this has generated so much great discussion. Having been in the science game for a bit, I've come to be open to, but also skeptical of, anything presented at a conference. The way scientific conferences work is that you have to submit an abstract 3-6 months ahead of time. So you have to decide: do I submit work I've already published, or do I submit my exciting and novel, but potentially too-preliminary work? If you go with the former approach, you get dinged for not presenting anything original; why will people want to hear you talk about a paper when they could (and probably already have) read it? If you go with the later approach, then, honestly, a lot of what you're doing is guessing at the results you will have 3-6 months from now, which may or may not be accurate, because the time scale of experimental science is terribly unpredictable. In that case, there's huge pressure to oversell any preliminary results you do have (which could get you media coverage and lead to more funding), because conference attendees are also not very excited about listening to a talk where the presenter lays out the experimental design but doesn't have any results to show (a notable exception at ASHG this year was the fantastic talk by Chris Mason on the design of their N-of-1 study of the twins discordant for space travel, complete with video of Andy Feinberg trying to pipet into a Nanopore sequencer in the Vomit Comet). I really don't see any solution to this dilemma on the scientific side, so the thing that really needs to change is science reporting. We need to figure out how to effectively discourage or discredit science writing that parrots or embellishes a press release without substantial due diligence.
I think Tuck did the best with what he had, and I think he tried to be honest about what his results did and did not say, but once he gave that talk, and once the PR people sent out the press release, it was in the hands of Twitter and the media. It's hard to say anything concrete about their methods, since a 10 minute talk is not sufficient to present any great level of detail, but from what was presented there did seem to be several things that were problematic. They have also set themselves up for what will likely be a horrendously difficult peer review, because, given all the negative attention this has received already, nobody wants to be the editor or reviewer that lets a study like this get published unless the methods are rock-solid and the results are correctly interpreted. In fact, this study may never get published in its current form; it may just have been an attempt to generate additional funding for a larger study.
posted by infinitemonkey at 7:31 AM on October 11, 2015 [4 favorites]
I think Tuck did the best with what he had, and I think he tried to be honest about what his results did and did not say, but once he gave that talk, and once the PR people sent out the press release, it was in the hands of Twitter and the media. It's hard to say anything concrete about their methods, since a 10 minute talk is not sufficient to present any great level of detail, but from what was presented there did seem to be several things that were problematic. They have also set themselves up for what will likely be a horrendously difficult peer review, because, given all the negative attention this has received already, nobody wants to be the editor or reviewer that lets a study like this get published unless the methods are rock-solid and the results are correctly interpreted. In fact, this study may never get published in its current form; it may just have been an attempt to generate additional funding for a larger study.
posted by infinitemonkey at 7:31 AM on October 11, 2015 [4 favorites]
In fact, this study may never get published in its current form; it may just have been an attempt to generate additional funding for a larger study.
I suspect thisis spot-on. Ngun had to know he was courting the whirlwind but he probably didn't realize it would come so soon or so hard.
In the years after Cyril Burt's fraud was revealed someone did a comprehensive survey and came to the conclusion that as of that time in the 1980's, every twin study ever done had either been fraudulent or so flawed as to be useless. Every. Single. One. Now given that history, if you come to me proposing to do a twin study, my first question is going to be how are you going to assure me that yours will somehow be different? Why would you even want to go to that poisoned well?
And the answer to the second question will be the elephant in the living room: There are powerful forces which hope for these unpleasant social problems to have some physically observable and maybe treatable explanation. There is a lot of money ready to be offered to someone who thinks they can produce the gay gene or the IQ gene or the mass murderer gene. There are people eagerly waiting to start designing social programs and other "fixes" as soon as the thinnest bit of evidence appears, just as so many people were denied basic human rights in the name of Cyril Burt's "proof" that IQ was heritable.
If I sound a bit grumpy about this it's not because I have anything against science or scientists. Twin studies have a history that makes cold fusion look like next week's practical engineering project. Ngun should have been aware of that.
posted by Bringer Tom at 8:24 AM on October 11, 2015
I suspect thisis spot-on. Ngun had to know he was courting the whirlwind but he probably didn't realize it would come so soon or so hard.
In the years after Cyril Burt's fraud was revealed someone did a comprehensive survey and came to the conclusion that as of that time in the 1980's, every twin study ever done had either been fraudulent or so flawed as to be useless. Every. Single. One. Now given that history, if you come to me proposing to do a twin study, my first question is going to be how are you going to assure me that yours will somehow be different? Why would you even want to go to that poisoned well?
And the answer to the second question will be the elephant in the living room: There are powerful forces which hope for these unpleasant social problems to have some physically observable and maybe treatable explanation. There is a lot of money ready to be offered to someone who thinks they can produce the gay gene or the IQ gene or the mass murderer gene. There are people eagerly waiting to start designing social programs and other "fixes" as soon as the thinnest bit of evidence appears, just as so many people were denied basic human rights in the name of Cyril Burt's "proof" that IQ was heritable.
If I sound a bit grumpy about this it's not because I have anything against science or scientists. Twin studies have a history that makes cold fusion look like next week's practical engineering project. Ngun should have been aware of that.
posted by Bringer Tom at 8:24 AM on October 11, 2015
I am absolutely terrified of an actual test for sexual orientation being developed, given the continued retrograde nature of acceptance of people like me.
posted by feckless fecal fear mongering at 9:30 AM on October 11, 2015 [2 favorites]
posted by feckless fecal fear mongering at 9:30 AM on October 11, 2015 [2 favorites]
"Butbutbut this test says you're straight. Why are you nevertheless fucking same-sex people? There must be something horrifically wrong with you!" That's were the real "fun" will begin.
posted by ZeroAmbition at 10:43 AM on October 11, 2015 [1 favorite]
posted by ZeroAmbition at 10:43 AM on October 11, 2015 [1 favorite]
The fact is there is only one possible interesting result from a twin study, and the fact that you are doing the study predisposes you to hope for it.
And if this is what you are referring to, what exactly do you suppose this "one possible interesting result" to be?
You expect correlation, obviously. There would be no point in doing the twin study if you didn't expect correlation. If you don't get correlation it is essentially a negative result -- uninteresting and boring. You haven't discovered anything. But if you get correlation, you have discovered something. You have explained something. You have moved something from the realm of the mysterious to the known.
I think there's perhaps a misunderstanding somewhere here. The fact that twins are virtually genetically identical at birth means that they are, potentially, an excellent population to sample for testing hypotheses about epigenetic phenomena, doesn't it? The idea that choosing the object of an empirical study in this intentional way should constitute bias seems incoherent.
posted by clockzero at 1:22 PM on October 11, 2015 [1 favorite]
And if this is what you are referring to, what exactly do you suppose this "one possible interesting result" to be?
You expect correlation, obviously. There would be no point in doing the twin study if you didn't expect correlation. If you don't get correlation it is essentially a negative result -- uninteresting and boring. You haven't discovered anything. But if you get correlation, you have discovered something. You have explained something. You have moved something from the realm of the mysterious to the known.
I think there's perhaps a misunderstanding somewhere here. The fact that twins are virtually genetically identical at birth means that they are, potentially, an excellent population to sample for testing hypotheses about epigenetic phenomena, doesn't it? The idea that choosing the object of an empirical study in this intentional way should constitute bias seems incoherent.
posted by clockzero at 1:22 PM on October 11, 2015 [1 favorite]
You expect correlation, obviously. There would be no point in doing the twin study if you didn't expect correlation.
This is equally true for every observational study, regardless of the design. Correlation of what to what, exactly? Correlation of genotype to behavior?
someone did a comprehensive survey and came to the conclusion that as of that time in the 1980's, every twin study ever done had either been fraudulent or so flawed as to be useless. Every. Single. One. Now given that history, if you come to me proposing to do a twin study, my first question is going to be how are you going to assure me that yours will somehow be different? Why would you even want to go to that poisoned well?
Do you have a citation for this? (And why should a critique of methods in common use for twin studies in the 80s prevent you from doing a twin study with more modern methods in 2015?) A particular type of study is not useless just because there can be confounding variables or because the data can be interpreted in multiple ways. And of course, not every twin study is designed to study the heritability of traits. We also shouldn't abandon them just because once a bigot used them to argue something bigoted: if we're going that route we would need to ban half of statistics.
posted by en forme de poire at 1:26 PM on October 11, 2015 [2 favorites]
This is equally true for every observational study, regardless of the design. Correlation of what to what, exactly? Correlation of genotype to behavior?
someone did a comprehensive survey and came to the conclusion that as of that time in the 1980's, every twin study ever done had either been fraudulent or so flawed as to be useless. Every. Single. One. Now given that history, if you come to me proposing to do a twin study, my first question is going to be how are you going to assure me that yours will somehow be different? Why would you even want to go to that poisoned well?
Do you have a citation for this? (And why should a critique of methods in common use for twin studies in the 80s prevent you from doing a twin study with more modern methods in 2015?) A particular type of study is not useless just because there can be confounding variables or because the data can be interpreted in multiple ways. And of course, not every twin study is designed to study the heritability of traits. We also shouldn't abandon them just because once a bigot used them to argue something bigoted: if we're going that route we would need to ban half of statistics.
posted by en forme de poire at 1:26 PM on October 11, 2015 [2 favorites]
OK, I see now that you are indeed referring to genotype-to-behavior correlations: "But suggest that you might have a pointer however tenuous to the Gay GeneTM, and here we are talking about it on Metafilter."
However, as clockzero has pointed out already, that is in fact exactly the opposite of what this study is designed to investigate. The study is looking at twins who have the same genotype but are discordant for sexual orientation, to determine if they differ at certain non-inherited DNA modification sites. The study is not attempting to either look for or disprove heritability.
posted by en forme de poire at 1:48 PM on October 11, 2015 [3 favorites]
However, as clockzero has pointed out already, that is in fact exactly the opposite of what this study is designed to investigate. The study is looking at twins who have the same genotype but are discordant for sexual orientation, to determine if they differ at certain non-inherited DNA modification sites. The study is not attempting to either look for or disprove heritability.
posted by en forme de poire at 1:48 PM on October 11, 2015 [3 favorites]
Also, finally, as a gay scientist I reject the idea that we should censor research into the origin of human sexual orientation. In the past, ignorance about the etiology of homosexuality has not helped gay rights: instead it has mainly given bigots plausible deniability to peddle abusive, harmful "therapies" and to deny that sexual orientation even exists. The effect of cutting off this research, particularly in such an early stage (and here I think it is important to note that we still don't even have plausible models of how human sexuality is biologically determined or influenced, let alone a sensitive, specific, accessible pre-natal test!) is to rob us of our own natural history.
posted by en forme de poire at 2:25 PM on October 11, 2015 [2 favorites]
posted by en forme de poire at 2:25 PM on October 11, 2015 [2 favorites]
Do you have a citation for this?
Sorry, this was back in the 1980's before everything was electronic and I ran across it in an actual university library back when I had access to an actual university. It was part of a small flood of studies that came out after the Cyril Burt fiasco was unearthed though.
The fact that twins are virtually genetically identical at birth means that they are, potentially, an excellent population to sample for testing hypotheses about epigenetic phenomena, doesn't it?
You'd think so, but the abject failure of twin studies to ever actually accomplish anything useful seems to indicate otherwise. The problem is that twins begin to diverge as soon as they are exposed to their environment and humans are virtual chaos engines, our lives a sea of butterfly effects. Approaching it from a software and engineering perspective, asking how I would build something like a human being, I find the idea that any such behavior is hard-wired or mediated genetically to be frankly ridiculous. The brain is far too complex for the genome to form anything but the most approximate blueprint for it. But as I mentioned above, there are powerful forces which want simple explanations to these unpleasant phenomena.
The study is not attempting to either look for or disprove heritability.
No, it is fishing for a different evil, a physical test for what many people would consider a factor of psychological fitness. If you don't think such a test would be misused please give me a little while to complete my time machine so I can take you on a tour of the 1930's.
I reject the idea that we should censor research into the origin of human sexual orientation.
My problem isn't that they were researching sexual orientation. (My first comment in this thread was a snarky one suggesting that we might go looking next for the markers for other paraphilia, such as latex fetishes, since those obviously have a basis too. Must have ruffled a mod's feathers because it succumbed to Late Era Metafilter Comment Evaporation Syndrome.) My problem is that they are barking up a tree that has never had a fox in it and is very likely to fall on somebody's head. I'm all for funding your study of imprinting mechanisms and maybe finally figuring out exactly how things like sexual tension are coded in the nervous system. It seems kind of crazy that with all our medical capability we don't even know where the fundamental mechanisms of orgasm arise. But the real money isn't behind understanding sexuality, it's about controlling it.
posted by Bringer Tom at 2:45 PM on October 11, 2015
Sorry, this was back in the 1980's before everything was electronic and I ran across it in an actual university library back when I had access to an actual university. It was part of a small flood of studies that came out after the Cyril Burt fiasco was unearthed though.
The fact that twins are virtually genetically identical at birth means that they are, potentially, an excellent population to sample for testing hypotheses about epigenetic phenomena, doesn't it?
You'd think so, but the abject failure of twin studies to ever actually accomplish anything useful seems to indicate otherwise. The problem is that twins begin to diverge as soon as they are exposed to their environment and humans are virtual chaos engines, our lives a sea of butterfly effects. Approaching it from a software and engineering perspective, asking how I would build something like a human being, I find the idea that any such behavior is hard-wired or mediated genetically to be frankly ridiculous. The brain is far too complex for the genome to form anything but the most approximate blueprint for it. But as I mentioned above, there are powerful forces which want simple explanations to these unpleasant phenomena.
The study is not attempting to either look for or disprove heritability.
No, it is fishing for a different evil, a physical test for what many people would consider a factor of psychological fitness. If you don't think such a test would be misused please give me a little while to complete my time machine so I can take you on a tour of the 1930's.
I reject the idea that we should censor research into the origin of human sexual orientation.
My problem isn't that they were researching sexual orientation. (My first comment in this thread was a snarky one suggesting that we might go looking next for the markers for other paraphilia, such as latex fetishes, since those obviously have a basis too. Must have ruffled a mod's feathers because it succumbed to Late Era Metafilter Comment Evaporation Syndrome.) My problem is that they are barking up a tree that has never had a fox in it and is very likely to fall on somebody's head. I'm all for funding your study of imprinting mechanisms and maybe finally figuring out exactly how things like sexual tension are coded in the nervous system. It seems kind of crazy that with all our medical capability we don't even know where the fundamental mechanisms of orgasm arise. But the real money isn't behind understanding sexuality, it's about controlling it.
posted by Bringer Tom at 2:45 PM on October 11, 2015
You'd think so, but the abject failure of twin studies to ever actually accomplish anything useful seems to indicate otherwise.
Unless you have both a citation for this claim and an explanation of why the paradigm is inherently flawed, you can't reasonably expect other people to dismiss the comparative logic of the study design out of hand.
The problem is that twins begin to diverge as soon as they are exposed to their environment
That's not a problem for this kind of analysis, though. That's exactly why it can be effective.
and humans are virtual chaos engines, our lives a sea of butterfly effects.
This is hand-waving.
Approaching it from a software and engineering perspective, asking how I would build something like a human being, I find the idea that any such behavior is hard-wired or mediated genetically to be frankly ridiculous.
This sentiment is itself completely ridiculous, for a number of reasons, but to be brief it also amounts to arbitrarily anti-positivist hand-waving, which is supremely unconvincing.
The brain is far too complex for the genome to form anything but the most approximate blueprint for it.
This is gibberish.
But as I mentioned above, there are powerful forces which want simple explanations to these unpleasant phenomena.
The fact that something was mentioned doesn't constitute proof or even evidence of that thing. Unless there's some concrete and specific evidence of bias here, and none has been presented so far, it's important for these frivolous complaints to be correctly identified as sophistic nonsense.
Please don't take any of that personally, though.
posted by clockzero at 3:07 PM on October 11, 2015 [1 favorite]
Unless you have both a citation for this claim and an explanation of why the paradigm is inherently flawed, you can't reasonably expect other people to dismiss the comparative logic of the study design out of hand.
The problem is that twins begin to diverge as soon as they are exposed to their environment
That's not a problem for this kind of analysis, though. That's exactly why it can be effective.
and humans are virtual chaos engines, our lives a sea of butterfly effects.
This is hand-waving.
Approaching it from a software and engineering perspective, asking how I would build something like a human being, I find the idea that any such behavior is hard-wired or mediated genetically to be frankly ridiculous.
This sentiment is itself completely ridiculous, for a number of reasons, but to be brief it also amounts to arbitrarily anti-positivist hand-waving, which is supremely unconvincing.
The brain is far too complex for the genome to form anything but the most approximate blueprint for it.
This is gibberish.
But as I mentioned above, there are powerful forces which want simple explanations to these unpleasant phenomena.
The fact that something was mentioned doesn't constitute proof or even evidence of that thing. Unless there's some concrete and specific evidence of bias here, and none has been presented so far, it's important for these frivolous complaints to be correctly identified as sophistic nonsense.
Please don't take any of that personally, though.
posted by clockzero at 3:07 PM on October 11, 2015 [1 favorite]
The problem is that twins begin to diverge as soon as they are exposed to their environment and humans are virtual chaos engines, our lives a sea of butterfly effects.
Isn't this exactly the point of twin studies? And the treatment of a negative result as uninteresting is a problem - it should be interesting - common to every kind of research. It would seem the more particular concern for twin studies is the combination of necessarily small samples with statistical techniques that can force a favored "result."
posted by atoxyl at 3:14 PM on October 11, 2015 [2 favorites]
Isn't this exactly the point of twin studies? And the treatment of a negative result as uninteresting is a problem - it should be interesting - common to every kind of research. It would seem the more particular concern for twin studies is the combination of necessarily small samples with statistical techniques that can force a favored "result."
posted by atoxyl at 3:14 PM on October 11, 2015 [2 favorites]
(My first comment in this thread was a snarky one suggesting that we might go looking next for the markers for other paraphilia, such as latex fetishes, since those obviously have a basis too. Must have ruffled a mod's feathers because it succumbed to Late Era Metafilter Comment Evaporation Syndrome.)
It certainly ruffled my feathers, as a queer academic; I am incredibly unhappy with analogizing a sexual orientation to a kink and found your comment offensive on those grounds. If you'd rather, I can elaborate further on that, but hey! It got deleted for a reason.
I am with en forme de poire on the point of researching the etiology of sexual orientation, too. I think it's an interesting topic, and I don't really think that it is or should be off limits because we're scared of the political implications of what we find. (Hell, I don't think that the findings are even that politically dangerous. More on that in a second.) It's worth noting that Tuck Ngun is himself gay and makes absolutely no bones about that.
I'm all for funding your study of imprinting mechanisms and maybe finally figuring out exactly how things like sexual tension are coded in the nervous system
This comment betrays a total lack of understanding of how any of this works. I don't mean that to be insulting, but I do mean to say that what it tells me is that you are completely adrift in this field of biology and not entirely sure what you're talking about. (That's fine; this is a specialist field.) Imprinting is a particularly form of epigenetics that mostly serves to modify the expression of certain alleles of a gene based on the parent those alleles were inherited from; it's a special case and not at all the sort of epigenetics which is being discussed here.
Neither of those things are remotely coded in the nervous system or "hard wired." The nervous system is not a set of static circuitry; it changes in response to experience and in response to stimuli. It's not completely immutable, but it is not static enough to be reliably used in the way you're talking about either. This shit is much too complicated to give anyone a hope of controlling at this stage of the game, and maybe ever--there's too many caveats and interactions and, I suspect, multiple mechanisms that can lead to homosexuality. What we call "homosexuality" is almost certainly a catch-all of multiple subtly different phenotypes, and I'd be very surprised to find a single cause even for all cases of gay men.
With respect to controlling this, or selectively aborting gay fetuses or whatever--I find this logistically almost impossible. If that makes you guys feel better. For one thing, if I'm right about there being effectively multiple genotypic/epigenetic/developmental paths to wind up at "gay dude", good luck screening for all of them, people! Just look at the success the GWAS people have had over the past ten years trying to find genetic markers for heart disease.
For another thing, controlling epigenetic development via biochemical means is more or less impossible to do right now, and certainly not on the very specific scale you'd need for this--you'd need to be able to figure out what methylation patterns you needed to see in particular places and cell types in the brain, and you'd need to alter specific genes' methylation patterns on those cells. When people want to do epigenetics studies in vivo, they basically inject subjects with a drug that fucks up methylation throughout the entire brain, and that causes its own problems and is REALLY sketchy as a methodology. (Basically, it's like trying to figure out what effect flipping a particular switch in a complicated machine has on system outputs by wading in and smacking the whole machine with a hammer.)
Frankly, it would be easier and cheaper to just clone a straight person, and rather more ethical.
posted by sciatrix at 3:14 PM on October 11, 2015 [4 favorites]
It certainly ruffled my feathers, as a queer academic; I am incredibly unhappy with analogizing a sexual orientation to a kink and found your comment offensive on those grounds. If you'd rather, I can elaborate further on that, but hey! It got deleted for a reason.
I am with en forme de poire on the point of researching the etiology of sexual orientation, too. I think it's an interesting topic, and I don't really think that it is or should be off limits because we're scared of the political implications of what we find. (Hell, I don't think that the findings are even that politically dangerous. More on that in a second.) It's worth noting that Tuck Ngun is himself gay and makes absolutely no bones about that.
I'm all for funding your study of imprinting mechanisms and maybe finally figuring out exactly how things like sexual tension are coded in the nervous system
This comment betrays a total lack of understanding of how any of this works. I don't mean that to be insulting, but I do mean to say that what it tells me is that you are completely adrift in this field of biology and not entirely sure what you're talking about. (That's fine; this is a specialist field.) Imprinting is a particularly form of epigenetics that mostly serves to modify the expression of certain alleles of a gene based on the parent those alleles were inherited from; it's a special case and not at all the sort of epigenetics which is being discussed here.
Neither of those things are remotely coded in the nervous system or "hard wired." The nervous system is not a set of static circuitry; it changes in response to experience and in response to stimuli. It's not completely immutable, but it is not static enough to be reliably used in the way you're talking about either. This shit is much too complicated to give anyone a hope of controlling at this stage of the game, and maybe ever--there's too many caveats and interactions and, I suspect, multiple mechanisms that can lead to homosexuality. What we call "homosexuality" is almost certainly a catch-all of multiple subtly different phenotypes, and I'd be very surprised to find a single cause even for all cases of gay men.
With respect to controlling this, or selectively aborting gay fetuses or whatever--I find this logistically almost impossible. If that makes you guys feel better. For one thing, if I'm right about there being effectively multiple genotypic/epigenetic/developmental paths to wind up at "gay dude", good luck screening for all of them, people! Just look at the success the GWAS people have had over the past ten years trying to find genetic markers for heart disease.
For another thing, controlling epigenetic development via biochemical means is more or less impossible to do right now, and certainly not on the very specific scale you'd need for this--you'd need to be able to figure out what methylation patterns you needed to see in particular places and cell types in the brain, and you'd need to alter specific genes' methylation patterns on those cells. When people want to do epigenetics studies in vivo, they basically inject subjects with a drug that fucks up methylation throughout the entire brain, and that causes its own problems and is REALLY sketchy as a methodology. (Basically, it's like trying to figure out what effect flipping a particular switch in a complicated machine has on system outputs by wading in and smacking the whole machine with a hammer.)
Frankly, it would be easier and cheaper to just clone a straight person, and rather more ethical.
posted by sciatrix at 3:14 PM on October 11, 2015 [4 favorites]
Approaching it from a software and engineering perspective, asking how I would build something like a human being, I find the idea that any such behavior is hard-wired or mediated genetically to be frankly ridiculous.
You cannot extrapolate meaningfully to biology from experience in software or electrical engineering. Biological systems are not organized like intelligently designed software. This statement also indicates a lack of awareness of existing biological research into sexual behavior (for instance, work on animal models shows that sexual behavior can be affected using genetic tools), as well as how gene regulation works (gene regulation is dynamic, integrating information from the environment, and both affects and is affected by epigenetic markers).
posted by en forme de poire at 4:00 PM on October 11, 2015 [4 favorites]
You cannot extrapolate meaningfully to biology from experience in software or electrical engineering. Biological systems are not organized like intelligently designed software. This statement also indicates a lack of awareness of existing biological research into sexual behavior (for instance, work on animal models shows that sexual behavior can be affected using genetic tools), as well as how gene regulation works (gene regulation is dynamic, integrating information from the environment, and both affects and is affected by epigenetic markers).
posted by en forme de poire at 4:00 PM on October 11, 2015 [4 favorites]
Also, all of that is completely unrelated to how orgasm works. It's as if we're talking about the steering system on car models and you suddenly dropped in a comment about how the engine combustion works.
posted by sciatrix at 4:05 PM on October 11, 2015 [2 favorites]
posted by sciatrix at 4:05 PM on October 11, 2015 [2 favorites]
Metafilter: You cannot extrapolate meaningfully to biology from experience in software or electrical engineering.
posted by en forme de poire at 4:28 PM on October 11, 2015 [2 favorites]
posted by en forme de poire at 4:28 PM on October 11, 2015 [2 favorites]
Well that stirred up a little hornet's nest.
Sorry to have offended you Sciatrix, but please understand that from my perspective, it appears that sexual orientation and kink are very deeply intertwined and are very likely exactly the same thing, a set of images and sensations we imprint on which become the permanent triggers for sexual response. I don't believe those triggers are encoded in the genome because there's no room for them; the genome is very, very small compared to the information storage density of brain structures and the great plasticity and inconsistency of these image sets would seem to argue that they are acquired from the environment, as we know for certain happens with many other species.
As for knowing how things work and applying engineering observations to biology, living things are machines that perform functions according to the laws of physics. Take a very, very small fact, that orgasm is propelled by rhythmic contractions. Those contractions have to originate somewhere and we don't know where. Sexual tension has to be encoded somehow; it is obviously a thing we can experience that exists, so does it represent the level of a hormone or chemical (unlikely) or electrical activity in a ganglion somewhere (more likely)? We don't know. Our ignorance in these areas is vast.
Considering this ignorance, I think it is rather premature to be devoting much effort at all against questions like "why does this particular image or situation cause me to become sexually excited." That question encodes a sea of ignorance about how the brain, nervous system, and body work. The fact is we have only the vaguest notions of how humans perceive images and situations at all, much less how those might get selectively processed as part of the sexual arousal process.
The reason I brought up how orgasm works is that how orgasm works should be easy and we hardly know anything about it. How perceptual cues trigger and affect sexual response is hard because it involves a lot of other stuff we hardly understand at all about how the brain works.
And this brings us to the fundamental problem with the study in the OP, which is that it is too goddamn simplistic. It is as sensible as looking for epigenetic markers for an interest in rocket science or skill with languages or skill at cooking. It is nonsense but it is nonsense likely to get enthusiastic funding by certain folks with both deep pockets and prejudice. And because that kind of nonsense has been used to justify much evil in the past, it is our duty to call it out when it appears today.
posted by Bringer Tom at 4:35 PM on October 11, 2015
Sorry to have offended you Sciatrix, but please understand that from my perspective, it appears that sexual orientation and kink are very deeply intertwined and are very likely exactly the same thing, a set of images and sensations we imprint on which become the permanent triggers for sexual response. I don't believe those triggers are encoded in the genome because there's no room for them; the genome is very, very small compared to the information storage density of brain structures and the great plasticity and inconsistency of these image sets would seem to argue that they are acquired from the environment, as we know for certain happens with many other species.
As for knowing how things work and applying engineering observations to biology, living things are machines that perform functions according to the laws of physics. Take a very, very small fact, that orgasm is propelled by rhythmic contractions. Those contractions have to originate somewhere and we don't know where. Sexual tension has to be encoded somehow; it is obviously a thing we can experience that exists, so does it represent the level of a hormone or chemical (unlikely) or electrical activity in a ganglion somewhere (more likely)? We don't know. Our ignorance in these areas is vast.
Considering this ignorance, I think it is rather premature to be devoting much effort at all against questions like "why does this particular image or situation cause me to become sexually excited." That question encodes a sea of ignorance about how the brain, nervous system, and body work. The fact is we have only the vaguest notions of how humans perceive images and situations at all, much less how those might get selectively processed as part of the sexual arousal process.
The reason I brought up how orgasm works is that how orgasm works should be easy and we hardly know anything about it. How perceptual cues trigger and affect sexual response is hard because it involves a lot of other stuff we hardly understand at all about how the brain works.
And this brings us to the fundamental problem with the study in the OP, which is that it is too goddamn simplistic. It is as sensible as looking for epigenetic markers for an interest in rocket science or skill with languages or skill at cooking. It is nonsense but it is nonsense likely to get enthusiastic funding by certain folks with both deep pockets and prejudice. And because that kind of nonsense has been used to justify much evil in the past, it is our duty to call it out when it appears today.
posted by Bringer Tom at 4:35 PM on October 11, 2015
it appears that sexual orientation and kink are very deeply intertwined and are very likely exactly the same thing
uhhhhhhhhh no.
posted by feckless fecal fear mongering at 4:41 PM on October 11, 2015 [3 favorites]
uhhhhhhhhh no.
posted by feckless fecal fear mongering at 4:41 PM on October 11, 2015 [3 favorites]
uhhhhhhhhh no.
Please elaborate, I'm quite interested in this.
posted by Bringer Tom at 4:44 PM on October 11, 2015
Please elaborate, I'm quite interested in this.
posted by Bringer Tom at 4:44 PM on October 11, 2015
Imprinting is a particularly form of epigenetics that mostly serves to modify the expression of certain alleles of a gene based on the parent those alleles were inherited from; it's a special case and not at all the sort of epigenetics which is being discussed here.
I hate to be That Dick, but do you have a cite for this? Because I'd be amazed to find there is any epigenetic component to imprinting at all, just as there is hardly any epigenetic component to the general formation of memories. I would have thought that imprinting occurs when experiential recordings, which I think everyone agrees are encoded by process growth and synapse formation, happen in areas of the brain that are genetically hardwired for specific extra functionality. But I'm very interested if there is actual research showing something else.
posted by Bringer Tom at 4:47 PM on October 11, 2015
I hate to be That Dick, but do you have a cite for this? Because I'd be amazed to find there is any epigenetic component to imprinting at all, just as there is hardly any epigenetic component to the general formation of memories. I would have thought that imprinting occurs when experiential recordings, which I think everyone agrees are encoded by process growth and synapse formation, happen in areas of the brain that are genetically hardwired for specific extra functionality. But I'm very interested if there is actual research showing something else.
posted by Bringer Tom at 4:47 PM on October 11, 2015
Sexual orientation seems to be largely biologically driven; continuation of the species and whatnot. Sexual expression (which is where kinks would be) is likely not. Kinks run all over the map from spanking to clowns to tickling to whatever, whereas orientation comes in one of a very few flavours (comparatively speaking), and the overwhelming majority is heterosexual--that is, that whole passing on the DNA thing.
Plus, homosexual behaviour is seen in animals other than humans, whereas kinks don't really seem to be AFAIK.
In any case, you're the one making the assertions here, you need to defend them--and you're not doing a great job so far.
posted by feckless fecal fear mongering at 4:50 PM on October 11, 2015 [1 favorite]
Plus, homosexual behaviour is seen in animals other than humans, whereas kinks don't really seem to be AFAIK.
In any case, you're the one making the assertions here, you need to defend them--and you're not doing a great job so far.
posted by feckless fecal fear mongering at 4:50 PM on October 11, 2015 [1 favorite]
Very briefly, since I'm on a phone: two totally different meanings of imprinting. Google genomic imprinting for an overview of the kind I mean. It also, and I mean this kindly, does not speak well for your familiarity with epigenetics, since genomic imprinting is probably the most famous variety of epigenetics. Prader Wili and Angelman syndrome is also a good thing to look into.
posted by sciatrix at 4:51 PM on October 11, 2015 [4 favorites]
posted by sciatrix at 4:51 PM on October 11, 2015 [4 favorites]
Bringer Tom, I kinda feel like, given what the actual-as-in-they-work-in-the-field experts are saying here, you're more or less in Not Even Wrong territory. And I think you're perhaps not realizing how deeply insulting it can be to compare sexual orientation to a kink, given that's one of the many clubs us gender/sexual minority folks have been beaten over the head with many, many times. e.g. the frequent conflation of homosexuality with pedophilia.
posted by feckless fecal fear mongering at 4:56 PM on October 11, 2015 [2 favorites]
posted by feckless fecal fear mongering at 4:56 PM on October 11, 2015 [2 favorites]
Sexual orientation is an expression. It just happens to be an expression that is particularly important with regard to continuation of the species, but evolution doesn't care if a solution is the best, only that it works well enough.
Kink may be uniquely human because it involves technology to a high degree, but things that resemble it aren't unknown in nature and especially in domestic animals. Anting in crows is probably the closest thing that doesn't involve human technology. Anting may or may not have a sexual component because who can tell with birds that don't have external genitalia, but it certainly does seem to be deliberate elicitation of bliss from pain.
And we are all making assertions here; the fact is nobody really knows squat about any of this. How we got here is a dude did a highly simplistic study very similar to studies which have been used to support bogus claims in the past which formed convenient justifications for people to do evil things. I don't think I know for sure how all this shit works, but I don't think you do either. What I do think is that when I hear the words "twin study" I want to put my hand on my wallet and make sure I'm sitting at the rear of the restaurant.
posted by Bringer Tom at 4:58 PM on October 11, 2015
Kink may be uniquely human because it involves technology to a high degree, but things that resemble it aren't unknown in nature and especially in domestic animals. Anting in crows is probably the closest thing that doesn't involve human technology. Anting may or may not have a sexual component because who can tell with birds that don't have external genitalia, but it certainly does seem to be deliberate elicitation of bliss from pain.
And we are all making assertions here; the fact is nobody really knows squat about any of this. How we got here is a dude did a highly simplistic study very similar to studies which have been used to support bogus claims in the past which formed convenient justifications for people to do evil things. I don't think I know for sure how all this shit works, but I don't think you do either. What I do think is that when I hear the words "twin study" I want to put my hand on my wallet and make sure I'm sitting at the rear of the restaurant.
posted by Bringer Tom at 4:58 PM on October 11, 2015
How we got here is a dude did a highly simplistic study very similar to studies which have been used to support bogus claims in the past which formed convenient justifications for people to do evil things.
Yeah but like, the dude in question is actually and openly gay so the likelihood of that happening is low, and the actual experts who are commenting in this thread are explaining in great detail where and why and how you are wrong and you're kinda doubling down. You may not know how it works, I may not know how it works (except, y'know, for being a queer dude so there's a leetle life experience to draw on there), and there are people right here in this thread who do have some inkling how it works so mayyyyyyyyybe listen to them?
And I submit that comparing sexual orientation to kink, particularly when you are expressing concern about evil people using science for evil purposes, is something that it would be a really fantastic idea to avoid doing anywhere, not just MeFi.
posted by feckless fecal fear mongering at 5:04 PM on October 11, 2015 [1 favorite]
Yeah but like, the dude in question is actually and openly gay so the likelihood of that happening is low, and the actual experts who are commenting in this thread are explaining in great detail where and why and how you are wrong and you're kinda doubling down. You may not know how it works, I may not know how it works (except, y'know, for being a queer dude so there's a leetle life experience to draw on there), and there are people right here in this thread who do have some inkling how it works so mayyyyyyyyybe listen to them?
And I submit that comparing sexual orientation to kink, particularly when you are expressing concern about evil people using science for evil purposes, is something that it would be a really fantastic idea to avoid doing anywhere, not just MeFi.
posted by feckless fecal fear mongering at 5:04 PM on October 11, 2015 [1 favorite]
Also, my bad for assuming you meant the genomic imprinting vs behavioral imprinting. That said, behavioral imprinting is also a really weird place to jump for other reasons. I think that's what you were getting at, though?
posted by sciatrix at 5:10 PM on October 11, 2015
posted by sciatrix at 5:10 PM on October 11, 2015
Well that stirred up a little hornet's nest.
I don't know what your intentions are, but you're starting to sound like a troll.
Our ignorance in these areas is vast.
Well, your ignorance certainly is. Maybe you should listen instead of talking when you have little relevant expertise.
posted by clockzero at 5:21 PM on October 11, 2015 [5 favorites]
I don't know what your intentions are, but you're starting to sound like a troll.
Our ignorance in these areas is vast.
Well, your ignorance certainly is. Maybe you should listen instead of talking when you have little relevant expertise.
posted by clockzero at 5:21 PM on October 11, 2015 [5 favorites]
it appears that sexual orientation and kink are very deeply intertwined and are very likely exactly the same thing
This is certainly how a lot of straight and straight-privileged people have wanted to treat me, as though my sexuality were a fetish activity for them.
I wish someone would do a study on whether there's an epigenetic marker for homophobia instead of investigating whether being gay is a choice.
posted by bile and syntax at 5:27 PM on October 11, 2015 [3 favorites]
This is certainly how a lot of straight and straight-privileged people have wanted to treat me, as though my sexuality were a fetish activity for them.
I wish someone would do a study on whether there's an epigenetic marker for homophobia instead of investigating whether being gay is a choice.
posted by bile and syntax at 5:27 PM on October 11, 2015 [3 favorites]
I wish someone would do a study on whether there's an epigenetic marker for homophobia instead of investigating whether being gay is a choice.
Multiple studies have shown that, in men at least, homophobia is highly correlated with sexual arousal when shown naked men. So... that epigenetic marker for homophobia would be the same one this researcher was looking for, probably.
posted by feckless fecal fear mongering at 5:30 PM on October 11, 2015
Multiple studies have shown that, in men at least, homophobia is highly correlated with sexual arousal when shown naked men. So... that epigenetic marker for homophobia would be the same one this researcher was looking for, probably.
posted by feckless fecal fear mongering at 5:30 PM on October 11, 2015
And we are all making assertions here; the fact is nobody really knows squat about any of this.
But we do actually know a few things about sexual behavior -- it's not as if nobody has ever studied anything relating to it. For instance, a "kink" can be experimentally induced in rats just using a simple associative learning paradigm, yet nobody that I'm aware of has ever demonstrated something similar for sexual orientation. SSRI treatment also appears to attenuate some paraphilias without any discernible effect on sexual orientation: if they were controlled in similar ways this is not what we would expect. So there are indeed already indications that these phenomena aren't explained by the same processes.
posted by en forme de poire at 5:30 PM on October 11, 2015 [4 favorites]
But we do actually know a few things about sexual behavior -- it's not as if nobody has ever studied anything relating to it. For instance, a "kink" can be experimentally induced in rats just using a simple associative learning paradigm, yet nobody that I'm aware of has ever demonstrated something similar for sexual orientation. SSRI treatment also appears to attenuate some paraphilias without any discernible effect on sexual orientation: if they were controlled in similar ways this is not what we would expect. So there are indeed already indications that these phenomena aren't explained by the same processes.
posted by en forme de poire at 5:30 PM on October 11, 2015 [4 favorites]
And we are all making assertions here; the fact is nobody really knows squat about any of this...I don't think I know for sure how all this shit works, but I don't think you do either.
"A towel, The Hitchhiker's Guide to the Galaxy says, is about the most massively useful thing an interstellar hitchhiker can have...wrap it round your head to ward off noxious fumes or avoid the gaze of the Ravenous Bugblatter Beast of Traal (such a mind-boggingly stupid animal, it assumes that if you can't see it, it can't see you)"
posted by clockzero at 5:44 PM on October 11, 2015 [2 favorites]
"A towel, The Hitchhiker's Guide to the Galaxy says, is about the most massively useful thing an interstellar hitchhiker can have...wrap it round your head to ward off noxious fumes or avoid the gaze of the Ravenous Bugblatter Beast of Traal (such a mind-boggingly stupid animal, it assumes that if you can't see it, it can't see you)"
posted by clockzero at 5:44 PM on October 11, 2015 [2 favorites]
Also, my bad for assuming you meant the genomic imprinting vs behavioral imprinting. That said, behavioral imprinting is also a really weird place to jump for other reasons. I think that's what you were getting at, though?
Yes, I think we have a distinct language problem here. I was speaking of the process by which animals acquire their mother and mating imprints through exposure to stimuli in early life. I am quite certain a similar process, though not as sharp and easily observable as it is in precocial species, is at work in all macrofauna species including ourselves. There is simply no genetic component to it at all because there is no room for it in the genome and no reason for it to be there.
But I am also having a very unpleasant flashback to the 1980's here and will bow out for good now that I've had such a good look at the undercarriage of a bus.
posted by Bringer Tom at 6:08 PM on October 11, 2015
Yes, I think we have a distinct language problem here. I was speaking of the process by which animals acquire their mother and mating imprints through exposure to stimuli in early life. I am quite certain a similar process, though not as sharp and easily observable as it is in precocial species, is at work in all macrofauna species including ourselves. There is simply no genetic component to it at all because there is no room for it in the genome and no reason for it to be there.
But I am also having a very unpleasant flashback to the 1980's here and will bow out for good now that I've had such a good look at the undercarriage of a bus.
posted by Bringer Tom at 6:08 PM on October 11, 2015
No room for it in the genome? I don't even know where to start with that one. You're just.. you're just wrong on so many levels and in so many ways. en forme de poire explained, right here, two comments above yours, why we do actually know something about kink and orientation.
posted by feckless fecal fear mongering at 6:22 PM on October 11, 2015 [2 favorites]
posted by feckless fecal fear mongering at 6:22 PM on October 11, 2015 [2 favorites]
I don't even know where to start with that one.
You might try information theory. Seven gigabytes really isn't a lot of information when you're talking about something as complex as a human being.
We actually do have a pretty good idea how the genome builds the brain; it's a fractal structure and that has all kinds of interesting implications. One of those implications, if you've ever delved into that math, is that there is no such thing as fine control. There is no such thing as a small change. You can build a fantastically complex scaffold but you can't populate it with detailed instructions.
As far as kink vs orientation, we went through this before when everyone was welcome at gay pride day until there was a little respect and the folks with leather and other stuff going suddenly weren't welcome. I have typed the next sentence about fifty times and just fucking give up.
posted by Bringer Tom at 6:51 PM on October 11, 2015
You might try information theory. Seven gigabytes really isn't a lot of information when you're talking about something as complex as a human being.
We actually do have a pretty good idea how the genome builds the brain; it's a fractal structure and that has all kinds of interesting implications. One of those implications, if you've ever delved into that math, is that there is no such thing as fine control. There is no such thing as a small change. You can build a fantastically complex scaffold but you can't populate it with detailed instructions.
As far as kink vs orientation, we went through this before when everyone was welcome at gay pride day until there was a little respect and the folks with leather and other stuff going suddenly weren't welcome. I have typed the next sentence about fifty times and just fucking give up.
posted by Bringer Tom at 6:51 PM on October 11, 2015
What on earth are you on about? Are you even bothering to read what actual people who actually work in this actual field are trying to tell you?
Kink and orientation are different things that appear to arise through different mechanisms.
The whole respectability at Pride thing has exactly not one iota to do with this discussion.
posted by feckless fecal fear mongering at 7:01 PM on October 11, 2015 [1 favorite]
Kink and orientation are different things that appear to arise through different mechanisms.
The whole respectability at Pride thing has exactly not one iota to do with this discussion.
posted by feckless fecal fear mongering at 7:01 PM on October 11, 2015 [1 favorite]
Kink and orientation are different things that appear to arise through different mechanisms.
Cite? And no, that one with the rats does not cut it. That is in fact exactly the sort of thing anti-homosexual people used to gin up to say that teh gay was voluntary.
posted by Bringer Tom at 7:15 PM on October 11, 2015
Cite? And no, that one with the rats does not cut it. That is in fact exactly the sort of thing anti-homosexual people used to gin up to say that teh gay was voluntary.
posted by Bringer Tom at 7:15 PM on October 11, 2015
"different things that appear to arise through different mechanisms" does not in any way, by even the most ungenerous reading, remotely resemble "voluntary".
posted by Etrigan at 7:17 PM on October 11, 2015 [2 favorites]
posted by Etrigan at 7:17 PM on October 11, 2015 [2 favorites]
Cite? And no, that one with the rats does not cut it. That is in fact exactly the sort of thing anti-homosexual people used to gin up to say that teh gay was voluntary.
So you want a cite and one has been given and somehow it's not good enough? You must be getting quite the workout moving those goalposts.
The point of en forme de poire's statement about the rats is precisely that homosexuality does not appear to be voluntary, and does not arise in the same way that kinks do.
And lest you forget, I'm queer as a three dollar bill. Bent as a corkscrew. Sworn in ass pirate on the MV Buttfuckery. I'm also a kinky motherfucker. So please stop straightsplaining my sexual orientation to me, alright? Please start listening to the actual experts right here in this thread who have tried rather patiently to explain to you where you are in error and why.
posted by feckless fecal fear mongering at 7:26 PM on October 11, 2015 [2 favorites]
So you want a cite and one has been given and somehow it's not good enough? You must be getting quite the workout moving those goalposts.
The point of en forme de poire's statement about the rats is precisely that homosexuality does not appear to be voluntary, and does not arise in the same way that kinks do.
And lest you forget, I'm queer as a three dollar bill. Bent as a corkscrew. Sworn in ass pirate on the MV Buttfuckery. I'm also a kinky motherfucker. So please stop straightsplaining my sexual orientation to me, alright? Please start listening to the actual experts right here in this thread who have tried rather patiently to explain to you where you are in error and why.
posted by feckless fecal fear mongering at 7:26 PM on October 11, 2015 [2 favorites]
They were measuring involuntary behaviors in that study, and the mechanism is, literally, exactly what you were suggesting to be true for kink (i.e. associative learning) so your reaction to it confuses me. At this point I think you're just reflexively dismissing the arguments people are presenting to you without engaging with them.
posted by en forme de poire at 7:27 PM on October 11, 2015 [3 favorites]
posted by en forme de poire at 7:27 PM on October 11, 2015 [3 favorites]
I'd also add that two of the experts who are explaining things to you are, likewise, people who carry copies of the Gay Agenda* in their wallets.
* 7am: crap it's time to wake up again
7:25: ok ok get out of bed
7:45: breakfast
8:00: Why didn't you let the dog out?
9:00: Work, ugh...
posted by feckless fecal fear mongering at 7:30 PM on October 11, 2015
* 7am: crap it's time to wake up again
7:25: ok ok get out of bed
7:45: breakfast
8:00: Why didn't you let the dog out?
9:00: Work, ugh...
posted by feckless fecal fear mongering at 7:30 PM on October 11, 2015
Let's recap - smart young gay researcher investigates a question relating to his own life, gets interesting results. Press and public go berserk based on a 10 minute conference presentation of a pilot study. Many seek to vilify the researcher himself for his findings (despite zero evidence of malice or malfeasance on his part). Researcher is appalled by implications of the whole mess, and chooses to withdraw from academia entirely. (CITE)
This did not go well.
posted by synapse at 8:08 PM on October 11, 2015 [2 favorites]
This did not go well.
posted by synapse at 8:08 PM on October 11, 2015 [2 favorites]
That does suck, though I noted that even while he was defending this work, he was already hanging out a shingle as a data analyst for hire (see his Twitter profile, e.g.), which implies to me that he was already in the process of leaving academia at that time. So I don't know that this was necessarily the precipitating event (the order of events is not very clear from the BuzzFeed article).
posted by en forme de poire at 7:36 AM on October 12, 2015 [3 favorites]
posted by en forme de poire at 7:36 AM on October 12, 2015 [3 favorites]
Yeah, I think I'm done engaging. I'm really tempted to write up even more lists of "this is wrong, this is a hole in your understanding, this is not relevant, here's why kink and sexual orientation are not comparable, we do actually know about how changes in gene regulation have distinct and wide-ranging changes in social behavior (google vasopressin receptor distributions in prairie voles! or hell, licking and grooming behavior and maternal care in rats!)", but Bringer Tom, I don't think you're really engaging with what we're saying at this point... and you're certainly moving goalposts here. It's especially weird to be basically accused of furthering scientific homophobia as a queer scientist (which I said! right up there!). If anyone else wants more detail into what we know about research like this and where this work is coming from, I'll be happy to elaborate. Not convinced anyone's reading, though.
posted by sciatrix at 9:36 AM on October 12, 2015 [7 favorites]
posted by sciatrix at 9:36 AM on October 12, 2015 [7 favorites]
Ngun mentions leaving academia "last week" in news pieces published October 8th, which is the day ASHG put out their press release, so I don't think you can get causation between the media coverage of Ngun's work and his decision to leave academia. There's definitely room to blame poor and sensationalist science reporting here, or ASHG's choice to publicise this study in particular knowing the coverage it was likely to attract, but many of the scientists I have seen discussing this are taking aim at science they think is dubious and unsound and not targeting the researcher, and I don't think you need evidence of malice on the part of the researcher to scrutinise their methods. There are doubtless other small pilot studies with statistically iffy methods that do not get this level of scrutiny, and that does seem harsh on Ngun.
Andrew Gelman has some thoughts on the study, including a followup tackling Ngun's defense of his study and lack of correction for multiple testing. It strikes me that a lot of the discussion here is based on really fuzzy information - a press release based on the abstract for a talk, information about the methods coming out through tweets at the talk and then back-and-forth as to whether the reports were missing key details, etc. I think Gelman is spot-on to suggest that releasing a preprint with details of the work would have cut out a lot of these issues.
posted by penguinliz at 9:45 AM on October 12, 2015 [2 favorites]
Andrew Gelman has some thoughts on the study, including a followup tackling Ngun's defense of his study and lack of correction for multiple testing. It strikes me that a lot of the discussion here is based on really fuzzy information - a press release based on the abstract for a talk, information about the methods coming out through tweets at the talk and then back-and-forth as to whether the reports were missing key details, etc. I think Gelman is spot-on to suggest that releasing a preprint with details of the work would have cut out a lot of these issues.
posted by penguinliz at 9:45 AM on October 12, 2015 [2 favorites]
[Me:] I'd actually expect science to analyze data without a preference for what they hope to see. No one expected tooth cavities to play a role in heart disease, and yet the data proves they do.
MisantropicPainforest : Lemme guess, you aren't a scientist? Because lemme tell you, its never done that way. You are always testing hypotheses, selecting among a variety of models, etc. There is a lot of subjectivity in the actual doing of science.
You'd guess wrong, and have thus provided an example of what's faulty in this complaint. Throwing data out because it doesn't match your expectations - which is what this complaint aims for ("you'd expect epigenetic variants influencing sexuality to occur in the brain") - is the opposite of science.
It's cherry-picking results, and it's not what actual scientists should be doing. The authors are reporting a statistically significant finding; while there are valid criticisms, none of them are in the form of "We don't expect these results because x, therefore we don't believe them".
quaking fajita : Sure, except they didn't find any statistically significant differences.
Yes, they did. If you disagree, show your math.
posted by IAmBroom at 2:09 PM on October 12, 2015
MisantropicPainforest : Lemme guess, you aren't a scientist? Because lemme tell you, its never done that way. You are always testing hypotheses, selecting among a variety of models, etc. There is a lot of subjectivity in the actual doing of science.
You'd guess wrong, and have thus provided an example of what's faulty in this complaint. Throwing data out because it doesn't match your expectations - which is what this complaint aims for ("you'd expect epigenetic variants influencing sexuality to occur in the brain") - is the opposite of science.
It's cherry-picking results, and it's not what actual scientists should be doing. The authors are reporting a statistically significant finding; while there are valid criticisms, none of them are in the form of "We don't expect these results because x, therefore we don't believe them".
quaking fajita : Sure, except they didn't find any statistically significant differences.
Yes, they did. If you disagree, show your math.
posted by IAmBroom at 2:09 PM on October 12, 2015
So all Bayesian statistical analysis that uses informative priors is unscientific? Good to know.
posted by MisantropicPainforest at 2:48 PM on October 12, 2015
posted by MisantropicPainforest at 2:48 PM on October 12, 2015
IAmBroom
I'd actually expect science to analyze data without a preference for what they hope to see. No one expected tooth cavities to play a role in heart disease, and yet the data proves they do.
MisantropicPainforest
Lemme guess, you aren't a scientist? Because lemme tell you, its never done that way. You are always testing hypotheses, selecting among a variety of models, etc. There is a lot of subjectivity in the actual doing of science.
IAmBroom
You'd guess wrong, and have thus provided an example of what's faulty in this complaint. Throwing data out because it doesn't match your expectations - which is what this complaint aims for ("you'd expect epigenetic variants influencing sexuality to occur in the brain") - is the opposite of science.
I'm not sure you two are actually disagreeing with each other. IAB is (I think) merely pointing out that the process of turning data into statistical models and explanation more broadly shouldn't be colored by bias toward any specific outcomes, and that unexpected discoveries are no less valid for entailing possibilities that researchers didn't anticipate because they didn't know enough about their object of study yet.
MPF is likewise (I think) pointing out that scientists go into their investigations with very specific, theoretically- and empirically-informed frameworks, questions and predictions, but this does not preclude the possibility of totally unexpected results.
posted by clockzero at 6:28 PM on October 12, 2015 [2 favorites]
I'd actually expect science to analyze data without a preference for what they hope to see. No one expected tooth cavities to play a role in heart disease, and yet the data proves they do.
MisantropicPainforest
Lemme guess, you aren't a scientist? Because lemme tell you, its never done that way. You are always testing hypotheses, selecting among a variety of models, etc. There is a lot of subjectivity in the actual doing of science.
IAmBroom
You'd guess wrong, and have thus provided an example of what's faulty in this complaint. Throwing data out because it doesn't match your expectations - which is what this complaint aims for ("you'd expect epigenetic variants influencing sexuality to occur in the brain") - is the opposite of science.
I'm not sure you two are actually disagreeing with each other. IAB is (I think) merely pointing out that the process of turning data into statistical models and explanation more broadly shouldn't be colored by bias toward any specific outcomes, and that unexpected discoveries are no less valid for entailing possibilities that researchers didn't anticipate because they didn't know enough about their object of study yet.
MPF is likewise (I think) pointing out that scientists go into their investigations with very specific, theoretically- and empirically-informed frameworks, questions and predictions, but this does not preclude the possibility of totally unexpected results.
posted by clockzero at 6:28 PM on October 12, 2015 [2 favorites]
« Older In Battlestar Galactica, fracking causes pregnancy... | Shade Court Is Adjourned Newer »
This thread has been archived and is closed to new comments
posted by quaking fajita at 11:40 AM on October 10, 2015