Friday, February 26, 2010

Why does genetic determinism persist, in spite of the evidence?

What's so wrong with genetic determinism anyway? Isn't it just political correctness to claim that who we are isn't determined by our genes? Genes are the basis of every trait, aren't they? And, they follow rules so we can predict who'll have what. Lots of diseases are caused by mutational variants at single genes, so traits like behaviors and intelligence and skin color must also be reducible to single genes. And, indeed, we can even retrodict the reason that a given trait evolved, since, obviously, natural selection has molded all living things into what they are today, yes?

 Well, no, even if conventional wisdom is that we can predict pretty much anything from genes. And, people seem to believe it in droves; thousands are happily paying genotyping companies to have their disease risk determined, and at least 15,000 people have volunteered to have their whole genome sequenced by the 100,000 Genome Project. The predictions are often expressed in terms of probabilities, but vended and perceived as being causal -- if you're a regular here, you know what we think about this.

And genetic explanations for many traits, including behaviors, are being adopted far and wide. Just type "gene for" or "genes for" into Google, and see how many hits you get (hint: you'll find that it's close to 10 million, and Donald Trump has the genes for success). Delve even briefly into those results and you see that we now are being told by political scientists that how we vote is genetically determined, not to mention whether or not we vote, economists and psychologists tell us that 'the warrior gene' makes us aggressive (except when it doesn't) and influences whether or not we belong to a gang (if we happen to be male). The list is long and growing -- just do this same search again next week, and you'll see. (By the way, the selection against 'good genes' via the deaths of warriors while cowards stayed home dining well and having the warriors' women has worried societal thinkers from Plato through Darwin and the Nazis to the present. But if that was truly selective at the genetic level, why has society not become all cowards?)

Given how much we've learned about the complex polygenic, gene-by-environment nature of complex traits (e.g., see our discussion of GWAS vs lifestyle questionnaires as predictors of disease risk, and, indeed, a study of 19,000 women published in the Journal of the American Medical Association last week reports that risk of cardiovascular disease can't be predicted from genes identified by GWAS), why is determinism still so appealing? Especially given the potential risk (see our recent posts about eugenics, here and here).Why won't it loose it's grip on contemporary thinking (just as the other side of the pendulum, environmentalism, gripped the post-WWII generation)?

Certainly some of it is just momentum -- when the expensive infrastructure for genetic treasure hunting is built, be it in service to political science departments, psychology, epidemiology, anthropology or economics, it's hard to dismantle, even when the evidence against this approach mounts and the payoffs turn out to be slight. 

And too, it's all too easy to construct scenarios to explain positive results, even non-significant ones. Making selective, after-the-fact claims of success is just what homeopathic medicine, phrenology, and other nonsense therapies do. The temptation seems to be too great for many to resist, particularly in a field like anthropology, with its focus on human evolution and how we got to be what we are. It must be genetic, and it must be due to natural selection. And generally these stories can't be tested, so we can make up anything we want! So, we've got genes for long distance running in West Africans because of their history of cattle rustling, genes for opposable thumbs because, well, because of almost any explanation for why they make us human, and even genes for ping pong prowess in China (although the proposed adaptive reason for this escapes us at the moment, but there is one).

But hold on. These kinds of Just-So stories needn't be so rampant, if researchers would just ask themselves a few simple questions. First, when trying to explain the association of a gene with a trait is a researcher should wonder, if the trait of choice is really so adaptive, (meaning that people with the trait consistently had more children than those without, over many generations), why the trait still varies after so many generations of selective pressure. Perhaps selection wasn't all that strong, and the trait not so adaptive after all. Was the trait present in mammals before there ever were humans? If so, no human-specific arguments bear much weight.

Further, and this is just good science, how can the chosen evolutionary explanation be tested? How would we know why a trait evolved? We can't know enough about environmental pressures at the time the variant arose, nor about how it varied through time, to be truly convinced that we know how or why a trait is selected. So, if you've constructed an explanatory framework that involves assumptions about environmental pressures tens of thousands, or hundreds of thousands of years ago, how will you test it?  And, why should a complex trait like thumbs or language or head size be due to a single gene anyway? If not, what is its genetic and evolutionary basis?

As we say in The Mermaid's Tale, natural selection certainly happens, but in general it's weak, usually very very weak, among other things, rather than a steady, simple driving 'force'. And, the idea that complex traits are due to single genes or even a few genes has been shown over and over not to be so, even if there are determined efforts to make them genetically predictable. So, why do these selective genes-for scenarios persist?

We live in a fundamentalist age, and there's more than a hint of simplistic fundamentalism to these selective scenarios: Darwin was right, and it's heresy to question Darwin. Therefore, everything's due to natural selection -- even though Darwin himself cautioned against that kind of thinking. And, the legacy of success at finding single genes for many (rare, usually pediatric) diseases has lulled us into thinking that all traits should be due to a single gene, so why we're human and no longer chimps is probably due to a single gene, or at least a single trait (the use of language, brain size, uprightedness, thumbs, tool use, and so on), for which we'll surely find the gene. People may deny that their explanations are so simplistic, and say they recognize complexity, but look at what they say and how they act.

Fundamentalism is dangerous, no matter its form. And genetic fundamentalism, as the history of eugenics has taught us, is no exception. Throw in traits fraught with societal import, such as intelligence or criminality -- which can't even be defined meaningfully (does anyone ever look for genes for white collar crime?), much less reduced to single genes -- and it can be explosive.

The underlying problem is a hunger for simple answers to satisfy various desires, such as easy prediction, pharmaceutical bonanza, simple explanations of cause and origin. The scientific challenge is to understand the way genetic mechanisms and genetic variation, as part of a mix of causation, work and how traits affected by such a mix evolve.

 A lot hinges on the meaning of 'determinism'. The usual concept is the same as cause, in the sense that we say gravit y determines how fast something will fall, or temperature when something will melt or boil. But in genetics it can be a confusion of probabilistic association and physical cause. The problem is too great for a post, but probable association -- you have a 27% of being diabetic if you have such-and-such genotype --is a very different kind of determination. It usually is misleadingly simple in many ways, especially when many factors, measured and unmeasured, are involved. Rarely does it mean strict causation.


Harry Banaharis said...

You have identified the "missing heritability" problem that is becoming increasingly apparent as more GWAS are being performed with higher sample numbers: the strength of association between a trait and genotype reaches some statistical significance but has little or no clinical utility.

It's important, however, to note that there is another layer of information within the genome that GWAS do not presently investigate - the epigenome, which carries heritable changes in the DNA and histone proteins that are not associated with DNA sequence. This class of modifications can arise even after a brief exposure to an environmental stressor at developmentally sensitive periods and are the direct imprint of the environment on the genome. Evidence from the science of epigenetics supports a case a against genetic determinism yet interestingly, such genetic changes are easily detectable.

This brings us to the matter of epigenetic determinism and why it is that what your grandmother ate whilst your mother was a foetus - but was also developing her eggs - may determine your risk of metabolic syndrome today.

Ken Weiss said...

We write about the problems with GWAS a lot. There are people trying some clever statistical tricks--ooops! after climategate I better not use that word!--statistical approaches that try to get all the data contained in all the GWAS SNPs. They will find some things, and they'll get as much juice as there is in the SNP-stone. Predictive power will increase but, as a couple of papers have aptly put it, one can't really get much better than the level of heritability (usually way less than 50%) of traits, and even if risk predictions from GWAS genotypes get a lot better, they mostly will not be usefully gene-specific: they'll be aggregate or 'polygenotypic' effects, differing for everyone even with the same risk score

As to epigenetics, one could also enumerate other nonstandard 'genetic' effects, like various complex RNAs (some acting in 'trans' as well as 'cis') that current methods aren't really looking for.

The epigenome means different things to different people. Understanding heritable DNA modification effects that are heritable will be one source of new information--but also new complexity since it won't generally reduce the complexity we already have in DNA sequences (at least, I think that's what'll happen).

Also, many 'genetic' effects are stochastic and somatic, and this applies to mutation and epigenetic changes. Many studies show that stochastic changes during development can have major mosaic effects (classic recent studies involve agouti mice).

But it's not clear to me what you have in mind, because if the stress-related changes arise, will they not be during life and tissue specific? If so, how an you ascertain them? It seems to be similar to trying to ascertain precancerous cells lurking within an otherwise normal tissue.

The final example you gave illustrates what many are certainly finding: early life changes with later life effects. They won't be found by mapping, perhaps, but how we would detect them, much less understand their cause is a question; perhaps you have ideas about this. Or perhaps if we can test the right tissue somehow maybe we don't need to know what the cause was.

Anonymous said...

The problem is people think of genes and the expression of genes as one in the same. Which it isn't, and it's really painfully obvious because as we age our bodies and even minds change, we shed ideas and adopt new ones, and our brains rarely keep much of a record of this. Our brain usually goes "here we are, here's where we've always been".

It's really stupid when you think about it. Do people ever find themselves waking up fat, or waking up magically intelligent? Did Michael Jordan just wake up one morning a basketball phenom. No, he even says so himself in the commercial "Maybe it's my own fault". "Basketball wasn't some God Given gift, it was something I worked for. Every single day of my life".

Bottom line is most people are lazy, usually across the board. Lazy thinkers, lazy workers, and lazy television watchers. They want quick results, so genetic determinism makes this seem like a reality to them. What better way to be a great Politician if you're just born one. What better way to know your son will be quarterback than if his genes say so.

I think really what you're witnessing is that people who work hard are and probably have been few and far between. The age of the internet has brough forward all the whiners and people claiming they're "born this way", because most people who would idle there time on here are the same as those plopped in front of the idiot box (television) and likely the same who wouldn't been Lion chow 10,000 years ago.

Ken Weiss said...

People want simple answers to complicated questions.