7 ms·

The promise and peril of the new science of social genomics

45 points by hunglee2 7y ago

glofish 7y ago
The fundamental problem with the interpretation of this study is the same as with any study of big genomics data: If your dataset it sufficiently large and the phenomena is sufficiently complex you can find myriads of tiny, valid yet uninteresting and ultimately irrelevant effects. It is simply about accuracy - take any genomic data and any visible observable phenotype of the population for the genomic data, if you don't find anything relevant it simply means your method is not yet accurate enough. Will people that like to play pokemon have a genetic marker compared to those that do not? If you average it over hundred of millions of instances you might see a difference. Does it mean anything with respect to one single individual playing pokemon? No, many other factors will down out that contribution. etc. the same is with this study as well, of course every population will eventually develop its own markers unless are being constantly mixing randomly. Does it mean anything at all, beyond what we already know - that a population develops its own markers ...
buboard 7y ago
"social genomics" seems to be the same as population genetics, which is not "new science". I don't see how their findings are surprising. There seems to be a lot of mixing of cause and effect.
DataWorker 7y ago
Some hesitance on the behalf of some researchers shouldn’t be over-interpreted. Any usable science in this domain will be exploited by governments and businesses. Money and power will move on this while academics dither.
allovernow 7y ago
>they have been accused of “opening a new door to eugenics”, according to the title of a 2018 MIT Technology Review article by science historian Nathaniel Comfort. I don't understand this position. Should we just ignore entire branches of science just because their results are inconvenient for some people/narratives? Meanwhile we allow sloppy, one sided science from the socially acceptable camp to influence social policy in a potentially totally broken way. Agenda driven science is bad science. You cannot solve social issues if you refuse to consider where the problems may actually come from. If there's any validity to social genomics, then perhaps our approach of blanket solutions to all groups in society will be shown to be ineffective, and this is a critical avenue to explore because it would indicate that our current allocation of resources is suboptimal.
- vanderZwan 7y ago
  > Agenda driven science is bad science. You cannot solve social issues if you refuse to consider where the problems may actually come from. The problem is that genomics (and anything to do with heredity before it) has been and still is abused in the opposite way: people pretending it to be an inevitable predictor like the law of gravity is, while claiming to not have any social agenda until they have had a glass of wine or two and you start hearing absolutely disturbing things. I've worked for molecular neurobiologists working on brains - the vast majority of them are fine. They research schizophrenia and such that do have genes associated with them, but also get that a lot of the genetics are about incrasing risk factors, not determinism. But once you hit the sub-group that researches the heritability of IQ... well, let's just say that there is a very suspicious aspect to the self-selection bias in the people who decide to research that topic.
  - deogeo 7y ago
    Yet most (western) countries immigration policy is set on the opposite assumption - that personality traits (of which IQ is only one) are not heritable. The caution and "well it's complicated" shouldn't run in only one direction.
    - vanderZwan 7y ago
      > Yet most (western) countries immigration policy is set on the opposite assumption - that personality traits (of which IQ is only one) are not heritable. This is something often said by white people who never had to deal with the many issues around migrating to Western countries, even though we have the entire history of eugenics and "scientific" racism proving the opposite. > The caution and "well it's complicated" shouldn't run in only one direction. That would assume that there are two equally valid opposing views here, and yet there is nothing here to back that up. Science doesn't run on "both sides"-isms.
      - deogeo 7y ago
        > That would assume that there are two equally valid opposing views here, and yet there is nothing here to back that up. So why does the opening of https://en.wikipedia.org/wiki/Heritability_of_IQ https://en.wikipedia.org/wiki/Heritability_of_IQ say "Twin studies of adult individuals have found a heritability of IQ between 57% and 73% with the most recent studies showing heritability for IQ as high as 80% and 86%."? Is the mountain of research on the heritability of psychological traits so obviously 'nothing', you don't even have to bother giving a source for your claim?
gumby 7y ago
This is indeed (as quoted up in the article) very exciting, but as well as very frightening. I notice comments on this thread already ignore the causal aspects (I presume Nature figures its readers will have that in mind). The scary and lazy eugenics argument of course is that some people have an inherent “inferiority” due to their genetic inheritance. But evolutionary pressure at each step applies to local advantage only so the causal chain more likely runs the other way: that 200 years of increasingly centralized industrialization encouraged expression of genes that would favor certain behavior. Plus the reduced cost of transportation would encourage offspring for whom that was not an advantage to move away. Super fascinating. But if seized upon by the modern day phrenologists it could be super dangerous. Edit: a couple of typos
- SamReidHughes 7y ago
  Expensive eugenics babies (maybe just embryo-selected) will get outbred by traditionalists anyway. What will make things really weird is artificial womb tech.
  - shillremover 7y ago
    Very interested in why anyone would downvote this, it seems obviously true.
allovernow 7y ago
Honest question. Should we have banned research in nuclear science because it gave us the bomb? Is social science with two sided implications any different? I feel like the only difference here is that the layman is more likely to understand the implications of misuse of such research, which creates a far stronger social pressure than there was associated with nuclear science in the 40s-50s. But this cat is already out of its bag, and this is research that IMO needs to be done because there are enormous positive consequences for e.g. disease management, human enhancement, and social policy driven by a more complete picture of the origins of human behavior.
- throwawaytkgll 7y ago
  Deriving social policy from popgen results is absolutely insane. Even putting aside the weak correlations, uncertain results and total ignorance of gene/environment interactions (it's not additive, it doesn't make sense to say a trait is x% genetic, especially a behavioral one), it would require having access to the genome of every single person you want to manage. It tantamounts to advocating a totalitarian hierarchical caste system. And no, obviously laymen are not more likely to understand the implication, as evidenced by the "eugenics is good, actually" crowd in this thread.
  - allovernow 7y ago
    Why does this discussion always immediately jump to the absolute worst case scenarios? This should not be such a black/white, us/them topic. We are not advocating for eugenics, or caste systems, and the other protections that we have for human rights in society won't immediately disappear. >it would require having access to the genome of every single person you want to manage Only in this dystopian hellscape you're imagining. This isn't about micromanaging individuals. Its about better understanding the origins of behavior in society. Here's one immediate potential application: how can you be sure that quotas in hiring aren't unrealistic and counterproductive if you refuse to research the effect size of genetics in human interests? My overall point is that we don't know the answer to these questions and there is an emergent social pressure which presumes a worst case application of such research and does not allow for it to be conducted. This leads to assumptions pervasive in social, medical, and academic practices which we don't know to be true. What if effectiveness of psychological interventions hinges on genetics? Should such research be forbidden because of totalitarian fantasies?
readme 7y ago
This kind of research does not have any practical application in a society where we treat each other as equals. Further, it's statistics, so there are outliers. I am 100% certain that lots of the things that determine our lives are heritable, but that doesn't mean I think this kind of research is beneficial. Research should be aimed at improving the general state of humanity. We have already seen what this does. Why would anyone even give this stuff a chance?
- tehjoker 7y ago
  There is far too much friendliness to eugenics and phrenology, though usually not termed as such, in the academy. I saw a presentation where a student was using machine learning to predict intelligence from medical images AND a variety of demographic attributes attributes including sex. I didn't blame the student who seemed to just get assigned the project, but the professor who assigned it and guided it should know better. Scientists should avoid trying to "scientifically" ascertain peoples' place in the world based on inborn traits. Instead, the focus should be on helping people achieve what they want to achieve.
  - shillremover 7y ago
    The job of a scientist is to study objective reality, not be a personal coach.
- ls612 7y ago
  Ignoring real patterns just because we don’t like their implications or fear people will use them for demagoguery doesn’t make them any less real. If things are heritable we should certainly catalog in a rigorous manner how they are heritable.
- SamReidHughes 7y ago
  It's kind of like pro golfers practicing 115 yard shots a lot, until they measured the stats and found their strokes gained is the same as all the other pros'. It does have applicability if you want to know things like what causes crime and how much, how much can we fix generational poverty with moar education, and will legalizing crack cause 50% of the population to get wiped out by addiction.
Gatsky 7y ago
Something like educational attainment is obviously embedded in a particular cultural context. What they are actually saying is these SNPs explain X% of the variation in attainment in a particular system where things are taught a certain way and success is measured by exams, and you need to be good at algebra and sitting still for long periods. You can’t treat it as a universal feature of humanity. Also the education system will look very different in 30 years, who is to say your results are stable? This issue of outcome measures in social genomics seems insurmountable to me.

45 points by hunglee2 7y ago

The fundamental problem with the interpretation of this study is the same as with any study of big genomics data: If your dataset it sufficiently large and the phenomena is sufficiently complex you can find myriads of tiny, valid yet uninteresting and ultimately irrelevant effects. It is simply about accuracy - take any genomic data and any visible observable phenotype of the population for the genomic data, if you don't find anything relevant it simply means your method is not yet accurate enough. Will people that like to play pokemon have a genetic marker compared to those that do not? If you average it over hundred of millions of instances you might see a difference. Does it mean anything with respect to one single individual playing pokemon? No, many other factors will down out that contribution. etc. the same is with this study as well, of course every population will eventually develop its own markers unless are being constantly mixing randomly. Does it mean anything at all, beyond what we already know - that a population develops its own markers ...

buboard 7y ago

"social genomics" seems to be the same as population genetics, which is not "new science". I don't see how their findings are surprising. There seems to be a lot of mixing of cause and effect.

DataWorker 7y ago

Some hesitance on the behalf of some researchers shouldn’t be over-interpreted. Any usable science in this domain will be exploited by governments and businesses. Money and power will move on this while academics dither.

allovernow 7y ago

>they have been accused of “opening a new door to eugenics”, according to the title of a 2018 MIT Technology Review article by science historian Nathaniel Comfort. I don't understand this position. Should we just ignore entire branches of science just because their results are inconvenient for some people/narratives? Meanwhile we allow sloppy, one sided science from the socially acceptable camp to influence social policy in a potentially totally broken way. Agenda driven science is bad science. You cannot solve social issues if you refuse to consider where the problems may actually come from. If there's any validity to social genomics, then perhaps our approach of blanket solutions to all groups in society will be shown to be ineffective, and this is a critical avenue to explore because it would indicate that our current allocation of resources is suboptimal.

vanderZwan 7y ago
> Agenda driven science is bad science. You cannot solve social issues if you refuse to consider where the problems may actually come from. The problem is that genomics (and anything to do with heredity before it) has been and still is abused in the opposite way: people pretending it to be an inevitable predictor like the law of gravity is, while claiming to not have any social agenda until they have had a glass of wine or two and you start hearing absolutely disturbing things. I've worked for molecular neurobiologists working on brains - the vast majority of them are fine. They research schizophrenia and such that do have genes associated with them, but also get that a lot of the genetics are about incrasing risk factors, not determinism. But once you hit the sub-group that researches the heritability of IQ... well, let's just say that there is a very suspicious aspect to the self-selection bias in the people who decide to research that topic.
- deogeo 7y ago
  Yet most (western) countries immigration policy is set on the opposite assumption - that personality traits (of which IQ is only one) are not heritable. The caution and "well it's complicated" shouldn't run in only one direction.
  - vanderZwan 7y ago
    > Yet most (western) countries immigration policy is set on the opposite assumption - that personality traits (of which IQ is only one) are not heritable. This is something often said by white people who never had to deal with the many issues around migrating to Western countries, even though we have the entire history of eugenics and "scientific" racism proving the opposite. > The caution and "well it's complicated" shouldn't run in only one direction. That would assume that there are two equally valid opposing views here, and yet there is nothing here to back that up. Science doesn't run on "both sides"-isms.
    - deogeo 7y ago
      > That would assume that there are two equally valid opposing views here, and yet there is nothing here to back that up. So why does the opening of https://en.wikipedia.org/wiki/Heritability_of_IQ https://en.wikipedia.org/wiki/Heritability_of_IQ say "Twin studies of adult individuals have found a heritability of IQ between 57% and 73% with the most recent studies showing heritability for IQ as high as 80% and 86%."? Is the mountain of research on the heritability of psychological traits so obviously 'nothing', you don't even have to bother giving a source for your claim?

gumby 7y ago

This is indeed (as quoted up in the article) very exciting, but as well as very frightening. I notice comments on this thread already ignore the causal aspects (I presume Nature figures its readers will have that in mind). The scary and lazy eugenics argument of course is that some people have an inherent “inferiority” due to their genetic inheritance. But evolutionary pressure at each step applies to local advantage only so the causal chain more likely runs the other way: that 200 years of increasingly centralized industrialization encouraged expression of genes that would favor certain behavior. Plus the reduced cost of transportation would encourage offspring for whom that was not an advantage to move away. Super fascinating. But if seized upon by the modern day phrenologists it could be super dangerous. Edit: a couple of typos

SamReidHughes 7y ago
Expensive eugenics babies (maybe just embryo-selected) will get outbred by traditionalists anyway. What will make things really weird is artificial womb tech.
- shillremover 7y ago
  Very interested in why anyone would downvote this, it seems obviously true.

allovernow 7y ago

Honest question. Should we have banned research in nuclear science because it gave us the bomb? Is social science with two sided implications any different? I feel like the only difference here is that the layman is more likely to understand the implications of misuse of such research, which creates a far stronger social pressure than there was associated with nuclear science in the 40s-50s. But this cat is already out of its bag, and this is research that IMO needs to be done because there are enormous positive consequences for e.g. disease management, human enhancement, and social policy driven by a more complete picture of the origins of human behavior.

throwawaytkgll 7y ago
Deriving social policy from popgen results is absolutely insane. Even putting aside the weak correlations, uncertain results and total ignorance of gene/environment interactions (it's not additive, it doesn't make sense to say a trait is x% genetic, especially a behavioral one), it would require having access to the genome of every single person you want to manage. It tantamounts to advocating a totalitarian hierarchical caste system. And no, obviously laymen are not more likely to understand the implication, as evidenced by the "eugenics is good, actually" crowd in this thread.
- allovernow 7y ago
  Why does this discussion always immediately jump to the absolute worst case scenarios? This should not be such a black/white, us/them topic. We are not advocating for eugenics, or caste systems, and the other protections that we have for human rights in society won't immediately disappear. >it would require having access to the genome of every single person you want to manage Only in this dystopian hellscape you're imagining. This isn't about micromanaging individuals. Its about better understanding the origins of behavior in society. Here's one immediate potential application: how can you be sure that quotas in hiring aren't unrealistic and counterproductive if you refuse to research the effect size of genetics in human interests? My overall point is that we don't know the answer to these questions and there is an emergent social pressure which presumes a worst case application of such research and does not allow for it to be conducted. This leads to assumptions pervasive in social, medical, and academic practices which we don't know to be true. What if effectiveness of psychological interventions hinges on genetics? Should such research be forbidden because of totalitarian fantasies?

readme 7y ago

This kind of research does not have any practical application in a society where we treat each other as equals. Further, it's statistics, so there are outliers. I am 100% certain that lots of the things that determine our lives are heritable, but that doesn't mean I think this kind of research is beneficial. Research should be aimed at improving the general state of humanity. We have already seen what this does. Why would anyone even give this stuff a chance?

tehjoker 7y ago
There is far too much friendliness to eugenics and phrenology, though usually not termed as such, in the academy. I saw a presentation where a student was using machine learning to predict intelligence from medical images AND a variety of demographic attributes attributes including sex. I didn't blame the student who seemed to just get assigned the project, but the professor who assigned it and guided it should know better. Scientists should avoid trying to "scientifically" ascertain peoples' place in the world based on inborn traits. Instead, the focus should be on helping people achieve what they want to achieve.
- shillremover 7y ago
  The job of a scientist is to study objective reality, not be a personal coach.
ls612 7y ago
Ignoring real patterns just because we don’t like their implications or fear people will use them for demagoguery doesn’t make them any less real. If things are heritable we should certainly catalog in a rigorous manner how they are heritable.
SamReidHughes 7y ago
It's kind of like pro golfers practicing 115 yard shots a lot, until they measured the stats and found their strokes gained is the same as all the other pros'. It does have applicability if you want to know things like what causes crime and how much, how much can we fix generational poverty with moar education, and will legalizing crack cause 50% of the population to get wiped out by addiction.

Gatsky 7y ago

Something like educational attainment is obviously embedded in a particular cultural context. What they are actually saying is these SNPs explain X% of the variation in attainment in a particular system where things are taught a certain way and success is measured by exams, and you need to be good at algebra and sitting still for long periods. You can’t treat it as a universal feature of humanity. Also the education system will look very different in 30 years, who is to say your results are stable? This issue of outcome measures in social genomics seems insurmountable to me.