“Heritability is the proportion of phenotypic variation in a population that is due to genetic variation between individuals.” -wikipedia
Not all genetic variation is inherited. We used to think that essentially genotype=phenotype (i.e., having a gene meant you had whatever trait goes along with it). But now we know that all kinds of more complicated stuff is going on that determines whether genes are activated. So, there isn’t the direct correspondence between genes and proteins like we previously thought.
Further, heritability is typically estimated from twin studies, by comparing DZ and MZ twins. However, MZ twins likely have a more similar environment than do DZ twins. MZ twins often share a placenta. In addition, I have observed that they are often treated differently than DZ twins. For example, I suspect it’s more common to dress MZ twins alike, to assume they like the same things, etc. MZ twins also look more alike, and there is no doubt that appearance affects how people are treated. If it’s true that MZ twins are treated more similarly than DZ twins, we might expect estimates of heritability to increase with age, which is exactly what we find in practice.
Agreed – I think obesity is a great case study here for intuition-buliding. Some populations experience more obesity than others controlling for environment, but people from the same genetically homogeneous population will experience different rates of obesity depending in different environments (a la Jungle Effect) – and most complicated of all, different populations respond differently to changes in environment. The cognitive stuff interacts with purely physical stuff.
The land outside of SNPs is almost unfathomably complicated (and there’s even surprising complication with those puppies) – but a la Dunning-Kruger, you have to understand a lot about it in order to understand how complicated it is.
After everything that i being uncovered by epigenetics, we can never again return to what we used to think of “heredity”. not even when we can locate a gene and map it to a trait can we ever say that something is genetic or heritable. there are cool little things that pin open the DNA or close it shut. this determines IF the gene will ever be expressed. these little DNA holders are controlled by environment- not by some deterministic DNA-> RNA-> to protein programming. It simply does not work that way. To complicate it more, the environment that changes the opening or closing of DNA (gene expression) is handed down generation after generation. but it is NOT “genetic”. it is environmental control of genes and is somewhat plastic. more complicated still…. it is the genes (that create organisms) that create the environment that select for the genes and so on and so on. we can never think about genes in the same way- ever again. and when we talk about iq— it’s a loaded gun. soon people will be talking about heredity in the way the neo- reductionists talk about the mind/body/world. you cannot reduce things down to their parts because it is ALWAYS an interaction. it’s so simple, yet we keep positing that reduction is possible. we think we can locate x gene and tie it to x trait. but, unless the conditions in the environment that interact with the gene are right, x trait never appears. it’s an outdated view that still hold a prominent place in academia. paradigms take a long time to shift.
[...] toolkit genes, genetic switches and how most conversations about heritability are flawed. I learned a lot about imprinted genes from Charlene Lewis (especially BDNF), about [...]
A interesting study of identical twins who were falsely classified as fraternal. Turns out that they are even more similar than normal identical twins!
“Results suggest that heritability estimates may be higher if we deploy comparisons among twins who misperceive their zygosity – but mainly for behavioral phenotypes. While we may not make much of these differences, they at least give us comfort that by setting the GE covariance term to zero in standard heritability models, we are not significantly biasing results. A number of approaches—ranging from the misclassification strategy pursued here to
using IBD sibling resemblance models—seem to be converging on the results that the old narrow-sense heritability estimates are not far off. This assumes, of course, that the other assumption of random mating holds. However, if parents tend to be more alike genetically than they would be if mating were random (a likely case, especially if we believe genes are related to phenotypes and the same phenotypes that researchers tend to study are those on which mates also sort), then heritability estimates would be downwardly biased.”
http://www.nber.org/papers/w16711