john hawks weblog

paleoanthropology, genetics and evolution

heritability

  • Link parade, 2

    Tue, 2012-10-23 23:43 -- John Hawks

    Ben Phelan at Slate writes about the recent evolution of lactase persistence: "The Most Spectacular Mutation in Recent Human History".

    The plot is still fuzzy, but we know a few things: The rise of civilization coincided with a strange twist in our evolutionary history. We became, in the coinage of one paleoanthropologist, “mampires” who feed on the fluids of other animals. Western civilization, which is twinned with agriculture, seems to have required milk to begin functioning. No one can say why. We know much less than we think about why we eat what we do. The puzzle is not merely academic. If we knew more, we might learn something about why our relationship to food can be so strange.

    I wanted to quote that passage because it was my friend Greg Cochran's son Roddy who coined the term "mampires", which is exceptionally clever. On the article as a whole, I think Phelan makes too much of the "mystery" aspect of the advantage of lactase persistence. There are really only two serious hypotheses and none of the possible explanations are mutually exclusive. I would have liked to see the article devote more attention to the multiple lactase persistence mutations in other populations, which together point to the very great advantage of the trait in association with dairying.

    David Dobbs writes in the New York Times about the genetics of intelligence and what we know (and don't know) about it: "If Smart Is the Norm, Stupidity Gets More Interesting". The piece emphasizes that geneticists haven't had much luck finding genes that explain the heritability of intelligence. The problem of "missing heritability" has loomed over complex trait genetics for the last several years, meaning that we can estimate the heritability of traits with twin studies and other traditional pedigree approaches, but single gene loci do not account for much of the variance of these traits. One possibility is that common genes have such small effects that they are statistically difficult to find.

    Another possibility is that very rare genes of small effect -- or new mutations -- may explain the heritability of such traits within families. The most likely reason for large-effect mutations to be rare is if they are deleterious. Across a population, this hypothesis of many rare deleterious mutations is called "genetic load":

    But in some other genetic realms we do differ widely, for example, mutational load — the number of mutations we carry. This tends to run in families, which means some of us generate and retain more mutations than others do. Among our 23,000 genes, you may carry 500 mutations while I carry 1,000.

    Most mutations have no effect. But those that do are more likely to bring harm than good, Dr. Mitchell said in an interview, because “there are simply many more ways of screwing something up than of improving it.”

    This is a nicely balanced treatment and emphasizes evolutionary approaches in an accessible way for Times readers.

    From the San Jose Mercury News, a story by Lisa Krieger: "Open-source science helps San Carlos father's genetic quest".

    One tiny flaw in one gene in one little girl. That explains why Beatrice Rienhoff, 8, is so lean and leggy.

    ...

    No one else in her family had such a syndrome. In fact, apparently no one else in the world did either.

    Rienhoff -- a biotech consultant trained in math, medicine and genetics at Harvard, Johns Hopkins and the Fred Hutchinson Cancer Research Center in Seattle -- launched a search.

    Yes, you can do this now. This father is now making transgenic mice with his daughter's mutation to better understand its effects.

    (via Gene Expression)

    Ken Weiss writes about some of the reasons a family medical history is a better predictor of individual health than genotyping: "23andLess".

    The most likely truth at this stage is that such common traits like heart disease or how tall or heavy you are, are determined by a very large number of genes, mostly with individually very small effects. Each person with the 'same' trait--each diabetic, say--has that trait for a different genetic reason. Individual genetic variants may be causal contributors, but they are not very important.

    I agree with his point...although as I was reading the post, it occurred to me that doctors treat family history as if it were much more effective than it should be, if causal variants really have small effect sizes. Complex disorders are not the same as Mendelian disorders with low penetrance. Having a grandfather with heart disease, for example, should mean substantially less to you than having a grandfather who is tall.

    Tags: 
    links
    lactase
    open science
    heritability
    personal genomics
    intelligence

  • Teeth and teaching

    Sun, 2012-09-30 14:21 -- John Hawks

    Razib Khan has a short but worthwhile post about dental health and heritability: "The moral measure of bad teeth".

    As someone who is quite conscious of the power of genetics, I was quite taken aback by this blind spot. I realized that not only did I attribute my own rather fortunate dental health (so far) to my personal behaviors, but, I had long suspected those with dental issues of less than optimal habits. Obviously environment (e.g., high sugar diet) does matter. But apparently a great deal of the variation in the trait is heritable.

    Tooth pathologies are a great example of heritability, because they provide a discrete character (cavities or no cavities) with an age-dependent component, a meristic (how many cavities) component, and several well-characterized environmental components. They're wonderfully multifactorial -- unlike, say, obesity, which is usually understood only along a single factor dimension. Plus, teaching students about this stuff actually helps them address their own health needs, as well as those of their future children.

    Tags: 
    teeth
    heritability
    teaching

  • Twins separated at sport

    Tue, 2012-08-21 10:57 -- John Hawks

    The Shortcuts Blog discusses British runner Mo Farah, his identical twin Hassan, the heritability of extreme performance: "Could Mo Farah's talent have run in the family?". The twins were raised apart after age 8.

    But, he added, as kids "Mo and I were on a par as runners. Sometimes I would beat him, sometimes he would beat me. He has had the most technically advanced training and advice available in the world ... and I have had nothing. Who knows what I could have become? We could have been famous twin Olympic athletes."

    I generally give my first Anthro 105 lecture on sport, and I think this is a great topic to spark discussion.

    Tags: 
    sports
    heritability

  • Steampunk genetics

    Thu, 2011-09-15 10:42 -- John Hawks

    An article in European Journal of Human Genetics that came out a couple of years ago has always impressed me, and I just noticed that it has gone to open access: "Predicting human height by Victorian and genomic methods" [1].

    The premise is that Galton's method of predicting stature from relatives still gives substantially better predictions than genotyping a collection of known variants that influence stature. It's basically a restatement of the "missing heritability" problem in more concrete (and colorful) terms. Here's a passage from the abstract:

    For highly heritable traits such as height, we conclude that in applications in which parental phenotypic information is available (eg, medicine), the Victorian Galton's method will long stay unsurpassed, in terms of both discriminative accuracy and costs. For less heritable traits, and in situations in which parental information is not available (eg, forensics), genomic methods may provide an alternative, given that the variants determining an essential proportion of the trait's variation can be identified.

    A great illustration. We know the trait is heritable, but the heritability is spread across many, many loci most of which remain unidentified. Hence, we can't predict the stature well from genotypes.

    References

    1. Aulchenko YS, Struchalin MV, Belonogova NM, Axenovich TI, Weedon MN, Hofman A, Uitterlinden AG, Kayser M, Oostra BA, van Duijn CM, et al. Predicting human height by Victorian and genomic methods. European Journal of Human Genetics. 2009;17(8):1070 - 1075.
    Tags: 
    stature
    missing heritability
    heritability
    GWAS

  • Heritability and stature

    Mon, 2011-09-12 01:42 -- John Hawks
    Synopsis: 
    Heritability is the proportion of variance in a phenotype explained by additive genetic variance.

    Tall parents tend to have tall children.

    That's a simple generalization, not an absolute statement. You may be a short person with two tall parents. If you know many families, you'll probably know someone who is an exception to the rule. Two short parents may have a very tall daughter, and two siblings may be very different heights.

    Still, if we look at many families we will find that the stature of the parents lets us make some fair predictions about the statures of their children. We can look at data to quantify just how parent and offspring heights are related.

    For example, here is a plot of the statures of students in my Anthropology 105 class in 2010, compared to the mean of their mothers' and fathers' statures. The average of mother and father's heights are the midparent stature.

    Young men in this class have a taller average stature than young women. The picture separates men and women, and both taller sons and daughters tend to come from taller parents.

    We can do a bit better than this to quantify the relationship of the parents' and sons' and daughters' statures: We can put lines on the graph to show how the sons' and daughters' heights tend to increase with the midparent stature:

    Each of these lines is called a linear regression between midparent stature and offspring stature. The linear regression is the line that has the smallest squared distance from all of the points, added together.

    The squared distance is special. In statistics, the average squared distance from all points to the mean is called the variance. When we consider a trait like stature, there are many possible biological causes that can contribute to individuals being taller or shorter than the average. In statistical terms, these causes are all factors that contribute to the variance of stature.

    In the chart above, the linear regression represents the amount of the variance of the stature of the offspring that was contributed by the variance of their midparent statures. Parent statures can predict offspring statures and the linear regression is the prediction.

    It's not a perfect prediction. Take a look at the midparent value of 160 cm. When the midparent average is 160 cm, the regressions predict that daughters will be 156 cm and sons will be 168 cm. Out of Anthropology 105 students last year, two men had parents with an average of 160 cm, and both of them were very close to 168 cm in height — a good prediction! But the two women with parents this stature have very different heights. One is only 148 cm, the other 162, both more than 2 inches from the predicted value, in different directions. The regression gives the best prediction we can make from the parents, but there is obviously a lot of variation that can't be predicted in that way.

    Let's look more closely at the female students. The following chart adds together females from several years of Anthropology 105, with their midparent statures.

    The slope of the regression across these 300 women is 0.72. That means that roughly 72 percent of the variance of the students' stature can be attributed to variance in their midparent stature.

    This regression is special in genetics. Daughters resemble their parents because they inherit genes from them. The regression between the midparent and daughters' statures gives us a way to estimate the effects of those genes. In fact, the proportion of variance in a trait that can be explained by genes is the same as the slope of the regression. Geneticists call this proportion the heritability of stature. In my Anthropology 105 classes over the last few years, we would estimate the heritability of stature as 0.72, or 72%.

    Again, there are exceptions. Sometimes parents give other things to their children besides genes. The right foods, the right resources can make a difference to growth and development. And some genes can have unexpected effects. Most obviously, some genes make sons a lot taller than daughters, which is why we've considered the two sexes separately.

    The heritability of a trait is a powerful concept. It's important to understand some of its strengths and limits:

    1. Heritability refers to a population. A female in my Anthropology 105 class may be taller or shorter than her parents. We can't predict 72% of that student's stature; instead, 72% of the variance in the students can be explained by the variance among their parents.
    2. Traits with higher heritability have a lower influence from the environment. Traits that are highly influenced by the environment have a lower heritability.
    3. The response of a population to selection on a trait is determined by the heritability.
    4. Geneticists can look at other relatives besides parents to estimate heritability. Some of the strongest estimates come from comparing identical twins (who have the same genes) to fraternal twins (who share on average half their genes).
    5. Estimating heritability doesn't require us to know anything about the actual genes themselves.

    The last point is very important. We can quantify the influence of genes without knowing anything about which genes even affect a trait. Francis Galton invented the parent-offspring method of estimating heritability, more than twenty years before the word "gene" was coined. Remembering this helps to remind us that the concept of heritability is limited. It is not a guide to how genes work, it is a simple scale of which traits have a stronger or weaker genetic influence.

    Because it is a proportion, heritability varies only between zero and one. A trait with zero heritability is one for which none of the variance can be explained by the parents' variance.

    Heritability may be very low because the individuals in the population have little genetic variability. For example, an orchard of apple trees may consist of genetically identical individuals that are clones of a single parent tree. The apples (hopefully) all taste the same. But the trees may be very different in height, branch form, and the number of apples. These traits may be strongly influenced both by the environments of individual trees and by chance. By contrast, wild populations of oak trees are not made up of clones. The heights of individual trees are still strongly influenced by environments, but they may also be influenced by differences in genes.

    Study questions: 
    1. Make a list of three traits that have low heritability in humans. Why are these mostly influenced by the environment and not genes?
    2. Are there environments inhabited by human populations that would make the influence of genes seem to be less on some traits?
    3. Suppose that the linear regression between midparent and offspring for a trait has a slope of 0.56. What would you estimate as the heritability of this trait?
    Study terms: 
    stature
    heritability
    variance
    Tags: 
    heritability
    explainer
    Anthropology 105
    stature

  • Brain scans and gene scans

    Sat, 2011-04-02 08:20 -- John Hawks

    Wray Herbert notes the fallacy of interpreting fMRI and other brain imagery as especially meaningful: "The Brain Is Not an Explanation". I'm pointing to this because of the similarities between brain scanning in neuroscience and genotyping in genetics. The result certainly looks objective, but what is actually there?

    The problem is that the final product—the brain image—looks like a photograph, and that’s how most readers take it, as a simple snapshot of the brain in action. That’s in part because the simplicity of the message is appealing: Complicated behavior X lights up brain area Y. But such reductionism, Beck argues, lacks any explanatory power. Consider the chocoholic example again: Leaving aside the fact that chocoholic is not a recognized diagnosis, what does this study actually show? It shows that people who define themselves as chocolate cravers have more activity, relative to people who do not define themselves as chocolate cravers, is certain pleasures centers of the brain. That is, the sight and taste of chocolate activated the brain’s reward system in cravers, documenting . . . what? Well, documenting that some people find chocolate more rewarding than others. As Beck notes, we probably don’t need a brain scan to corroborate what most people probably already believe anyway.

    A study would be powerful if it gave a way of predicting phenotypes or behaviors from the scans (or the genotypes). But finding a correlate of a behavior doesn't make it more real than it already is. Remember that we can already predict many phenotypes from family history much more accurately than we can from genotypes. As Galton discovered, the additive genetic component of variance exists even if we know nothing about the mechanisms of transmission.

    But then, if the modality didn't matter, we wouldn't need to bother with chocolate at all. We could just stimulate the pleasure center directly.

    Tags: 
    brain
    neuroscience
    behavior
    heritability
    testing

  • Sports and genetics

    Sat, 2010-12-18 16:46 -- John Hawks

    Sports Are 80 Percent Mental has an interview with Peter Vint of the U.S. Olympic Committee: "Do Young Athletes Need Practice Or Genetics? A Conversation With Peter Vint". Vint does a good job of describing the complexity of performance -- most sports require a combination of physical and mental skills that are developed through learning and practice:

    For example, a pilot controlling an automated aircraft may need only nominal motor skill to press a button, but will require substantial mental and perceptual skill to understand what happens when the automation switches from one mode to another. On the other hand, a basketball player will require extensive motor skill in executing a drive to the basket but will, though to a lesser extent, also involve perceptual and mental skills. Good examples of the world's best players in sport (especially team sports) seem to have exceptionally well developed perceptual skills which allow them to "see the field" better than others and "know where players will be before they even arrive".

    The broader topic of the interview is, what is the relationship between practice, genetics, and talent? Sport is interesting for its diversity. Even a pure running race involves both mental and physical skill, and the mental game is much greater for some sports. Many sports have limiting physical requirements -- elite distance cyclists, for example, are in a very narrow range of body mass compared to the average man. Even so, the subset of people who make it as elite cyclists is so small that a wide range of performance and personality traits can influence the set. At one level, the personality traits are primary -- to compete at an elite level, you need the support of a team, which may mean paying your dues as a domestique for several years.

    One reason it's hard to talk about genetic influences on sport performance -- the point at which most people care whether genes make a difference is long after many genes have the opportunity to matter. Whether genes make a difference to middle school sports learning and performance or not -- that's just not a burning topic of study. Whether genes make a difference at an elite level is, with a few exceptions, a harder kind of genetic question, because the gene-environment interaction may be so strong by the time people reach the elite level.

    A comparable problem: finding genes that influence recovery time after heart surgery. Recovery time may be influenced by several genes of strong effect, but these genes need not have anything to do with the initial risk of cardiovascular disease. Still, everyone who got the surgery must first have gotten cardiovascular disease -- the initial risk genes influence the composition of the sample. That's the relation between elite athletes and beginners -- nobody gets to the Olympic level without a long period of being pretty good at their sport, with all the practice that entails.

    (via Neuroanthropology)

    Tags: 
    behavior genetics
    sports
    heritability
    testing

  • Heritability and genetic test essay

    Fri, 2010-09-10 11:40 -- John Hawks

    Neuroscientist Dorothy Bishop provides a student-level opinion piece in the Guardian that addresses the "missing heritability" problem without using the term: "Where does the myth of a gene for things like intelligence come from?".

    It's an unfortunate headline, because she doesn't disagree with a strong genetic influence on personality, intelligence or other behavioral traits. Bishop merely explains that they are polygenic, with few genes of strong effect. The "myth" is the assumption that there's "a gene for" a trait, instead of many genes influencing a trait.

    Consider one of the more reliable associations between genes and behaviour: a gene known as KIAA0319 which has been found to relate to reading ability in several different samples. In one study, an overall association was reported with a p value of 0.0001, indicating that the likelihood of the association being a fluke is 1 in 10,000. However, this reflected the fact that one gene variant was found in 39% of normal readers and only 25% of dyslexics, with a different variant being seen in 30% of controls and 35% of dyslexics.

    Some commentators have argued that such small effects are uninteresting. I disagree: findings like this can pave the way for studies into the neurobiological effects of the gene on brain development, and for studies of gene-gene and gene-environment interactions. But it does mean that talk of a "gene for dyslexia", or genetic screening for personality or ability, is seriously misguided.

    This is far from saying there's "no gene" for dyslexia; but the public could use much more discussion of what these genetic studies actually show, and much less sensationalizing of the "scientists find gene for dyslexia" variety.

    I know, I'm preaching to the choir here. What I don't understand is why we keep having to make this point, after thirty years of headline writers getting this wrong time and time again. I won't blame science writers for this one, because the real problem is the sound byte.

    I don't fully agree with the piece, in that I think Bishop underestimates the prospect of genetic tests. Sure, testing a single SNP does no good for most traits, but in the long run we will see many quantitative traits for which multi-locus combinations predict phenotypes a lot better than chance.

    Still, the general point holds. Right now, family history is a better predictor of most traits than any genetic assay. Having a twin gives you the best prediction of all -- which is another way of saying that the heritability estimates from twins and other pedigrees that Bishop discusses are the evidence we need to account for genetically.

    Tags: 
    heritability
    genetics and society
    testing
    genetics

  • Genetics of brain surface area and cortical thickness

    Tue, 2009-12-29 01:02 -- John Hawks

    The field of brain imaging has progressed remarkably during the past several years. I follow the literature because as the study of heritability of brain structure becomes more precise, it may start to be possible to study the polymorphisms that explain the genetic variation of the brain. Twin designs are the most powerful ways of assessing the heritability, although they do present certain weaknesses. A fairly comprehensive recent review of imaging studies of brain structure and heritability in twins is the paper by Peper and colleagues (2007), cited below. Most volumetric measures of the brain and its parts have heritabilities greater than 50 percent; total brain volume is about as heritable as stature in most studies, upward of 85 percent.

    Fifteen years ago, the state of the art was MRI studies of a dozen or so pairs of both monozygotic and dizygotic twins. By the end of the 1990's, studies started to appear that used more than 50 of each type of twin. With sample size, the power of these studies increased. Moreover, advances in imaging technology and software began to allow researchers to focus on smaller structures, first whole lobes, and later smaller parts of the brain.

    Much of your thinking happens in the cerebral cortex, which is a relatively thin layer of tissue, folded around the outside surface of the brain. Lately, MRI studies have begun to explore the structure of the cortex itself, by measuring the surface area of cortex in different regions of the brain, as well as the thickness of the cortex in the same areas.

    Panizzon and colleagues (2009) examined MRIs from the Vietnam Era Twin Study of Aging, numbering 110 monozygotic and 92 dizygotic twin pairs. That sample size makes the study quite powerful for testing heritability, and they find that additive genetic (heritable) factors account for 89 percent of the variation in cortical surface area, and 81 percent of the variation in average cortical thickness. The gray matter volumes of different lobes range from 31 percent to 88 percent additive variance, with unique (non-shared) environmental factors generally accounting for a bit over half the remainder.

    Despite being relatively large compared to most earlier imaging twin studies, the sample of around 200 twin pairs is still relatively weak for testing correlations among brain areas. This limitation is important to keep in mind, because one of the important genetic questions is the extent to which separate developmental modules may be involved in overall brain development. The operation of common factors across different brain regions would be evidence against extreme modularity; independence of different brain measures argues for a modular developmental model.

    After controlling for brain volume, the thickness of the cortex is negatively correlated with surface area in this sample, although non-significantly so. This suggests that these two features of the cortex are relatively independent, instead of being determined by shared factors. That's after controlling for endocranial volume, however, which is strongly positively correlated with surface area. There were very few significant correlations found between different sections of the cortical surface, in either area, thickness, or volume.

    The paper takes the relative independence of cortical area and thickness to be its major finding:

    That surface area and thickness are genetically distinct from one another has numerous implications for continued investigations into the genetic influences of brain structure. Perhaps the most significant of these is the need to explore the genetics of surface area to a greater degree. Like thickness, surface area is a highly heritable construct; yet, it has been largely overlooked in human imaging genetics research. Specific mutations in humans have been linked to excessive gyrification of the cortex as well as an increase in the cortical surface area (Piao et al. 2004; Jansen and Andermann 2005). Animal studies have also demonstrated that manipulation of specific genes can result in dramatic changes in arealization and expansion of select areas of the cerebral cortex like the primary visual area and the primary somatosensory area (Bishop et al. 2000; Mallamaci et al. 2000). Findings such as these would appear to suggest that the genes that influence surface area are critical to the early growth and development of the brain. The observed genetic relationship between total surface area and intracranial volume lends support to this conclusion. If this is the case, then a more focused examination of the genetics of surface area may produce new insights into disorders believed to have early developmental origins, such as schizophrenia (Panizzon et al. 2009: 5-6).

    On one level, these kinds of studies involve incredible technology and massive investment in recruiting and following research subjects. But on another, they are very simple -- measure volumes and areas and plug them into an equation. Total brain volume is among the most heritable gross anatomical measures on the human body, and the lobar divisions of the neocortex are nearly as heritable -- while, interestingly, some smaller parts (such as the hippocampus) are apparently more influenced by unique environmental factors. Surface area and cortical thickness, despite the thin, sheet-like nature of the cortex itself, apparently go along with total brain volume with their relatively high heritabilities.

    References:

    Panizzon MS, Fennema-Notestine C, Eyler LT, Jernigan TL, Prom-Wormley E, Neale M, Jacobson K, Lyons MJ, Grant MD, Franz CE, Xian H, Tsuang M, Fischl B, Seidman L, Dale A, Kremen WS. 2009. Distinct genetic influences on cortical surface area and cortical thickness. Cerebral Cortex (advance) doi:10.1093/cercor/bhp026

    Peper JS, Brouwer RM, Boomsma DI, Kahn RS, Hulshoff Pol HE. 2007. Genetic influences on human brain structure: A review of brain imaging studies in twins. Human Brain Mapping 28:464-473. doi:10.1002/hbm.20398

    Tags: 
    genetics
    twins
    brain
    heritability
Subscribe to heritability

Neandertals

For years, I've worked on their bones. Now I'm working on their genes. Read more about the science studying these ancient people.

Denisova

From a finger bone of an ancient human came the record of a completely unexpected population. My lab is working on the science of the Denisova genome.

Acceleration

The advent of agriculture caused natural selection to speed up greatly in humans. We're uncovering some of the ways that populations have rapidly changed during the last 10,000 years.

Malapa

Just outside Johannesburg, the Malapa site is producing some of the most exciting finds in human evolution. This site is the headquarters of the Malapa Soft Tissue Project.