http://www.ncbi.nlm.nih.gov/pubmed/19297406

There is nothing approaching this in genes for height or â??intelligenceâ? or even what are called â??complex genetic diseasesâ?. Even though much more effort has been put toward them than has been put toward skin color. Skin color is trivial compared to height, and height is trivial compared to â??intelligenceâ?.

Gene skin color: 635 PubMed
Gene height: 2786
Gene intelligence: 2633
Complex genetic diseases: 24889

I think that Greg and I are on exactly the same page, but that is only going on what I have read on his blog. We have never talked or met in meat space (as far as I know), so my knowledge of his positions on things is quite incomplete and I haven’t gone back in his archives to figure it out better. His language is pretty clear. He is not saying that genes play no role, just that no one has shown that genes play as big a role as they claim, and that the data so far seems to indicate (very strongly) that genes play less of a role than people seem to think and claim. I agree with this, not because Greg has said it, but because my independent reading of the literature leads me to the same conclusions.

Why out of the infinite universe of potential hypotheses did you pick those 3? Probably because most everyone these days has a gene fetish. There is lots of neat new gene-crunching equipment, so everyone wants to be â??firstâ? at pushing cells through DNA crunching machines so they can be â??firstâ? at discovering â??the geneâ? for â??everythingâ? so they get a ticket to Stockholm, 11!!!!!

Why out of the very large number of different ways that â??genesâ? and â??environmentâ? could interact you chose this model?

â??But remember, the deviation due to independent variables adds in quadrature. That means that if the environmental variation creates a standard deviation of 5 inches, and if the effect of genetics is equally strong, then the net standard deviation becomes 7 inches. In practice, we couldn’t tell the difference.â?

You are assuming a linear combination. Why are you assuming a linear combination? Only because non-linear combinations are too hard (said with slight nasal whine). Is there any evidence that gene-environment interactions are linear (what ever that actually means)? No, there isn’t for the handful of known gene-environment interactions (PKU for example). How about for the zillions upon zillions of unknown gene-environment interactions? There is not a clue that they are linear. How about for gene1-gene2-gene3-gene4-environment1-environment2-gene5-environment3-gene6 interactions? I don’t know what those 6 genes or 3 environmental factors are, but I am pretty sure (99.999% sure) their interaction is not linear. It pretty much can’t be linear because of the way that DNA is processed. DNA codes for RNA, which is used to make proteins. A discrete number (i.e. an integer) of protein molecules is made each time a gene is transcribed (more or less).

One of the genes I have spent a lot of time thinking about is the MeCP2 gene (because it relates to autism). This codes for the MeCP2 protein and is on the X chromosome. The loss of this gene causes Rett Syndrome in females and fetal death in males. The presence of the extra X chromosome in females rescues the female phenotype even though half the X chromosomes are silenced. RS female usually appear normal until about 18 months when they regress into the RS phenotype. The MeCP2 protein binds to methylated DNA, that is DNA that has been epigenetically programmed by being methylated. There are probably at least thousands of genes and many other spots of DNA that are methylated and so are differentially regulated by the MeCP2 protein, in different cells and in different tissue compartments.

What would be the result of a linear interaction of the MeCP2 deletion with the environment? I have a hard time imagining what a linear model of the interaction of MeCP2 with the environment would even look like. Maybe like a one-dimensional map of the Earth’s surface? Good luck using a one-dimensional map to figure out where you are in four-space.

This is what using a unitary measure for height and IQ is like, a one-dimensional mapping of something much more complicated. Height is the linear sum of the lengths of the leg bones, the pelvis, each spinal vertebra, the skull, the scalp, the foot. The knee fits in there too. Is there a foot-bone-length gene? A skull thickness gene? A leg-bone-length gene?

There are some things that cannot be understood in a simplified form. Many multi-dimensional systems are like that. If you don’t understand a two-dimensional painting, simplifying it to one-dimension isn’t going to make it easier to understand. Many (all?) multi-dimensional non-linear systems are like that.

If you don’t understand something, don’t make something up and pretend that you do.

Skin color is complex, there are unknowns, we’d like to know more, but that allelic variation causes a measurable part of the variation (the “underarm” color, we biological anthropologists call it), what the proteins involved are, and what many of the genes involved are … and, most importantly, the basic biological story … genes to proteins to phenotypes … is understood and modeled.

For something like intelligence, we lack the genes, we lack the proteins ,we lack the developmental mechanism, we lack a measure or even a vague sense of what the different sources of variatoin might be. All we really know is that a huge

HUGE!!!!!!!

amount of variation is caused by environment, and the search for genetic causes has been all about familial studies and phenotypes, not about proteins (gene products) and genes.

In less than 30 seconds I located a dozen papers on genetics of skin color in humans published since 2003, so yeah.

Regarding dogs, you are now using the argument that since I did not spoon feed you data or a reference I must be wrong.

Have you ever seen a dog? Do you know what a mutt is? Ha!

Anyway, the not-as-genetic-as-people-think trait I mentioned in dogs is aggressive behavior. You realy won’t have to work very hard to verify what I said on that.

Of course. Accordingly, I have stated a number of times that the environment does have an effect. I never disputed that. The question is whether genes also have a significant effect.

You are not doing your homework. The secular trend has caused a HUGE variation in height. No genes. Just environment.

But there is a gene for stature .. a gene with two alleles, one for short and one for not short, in humans. I would think that if you were serious about what you are arguing here you’d have google-fu’ed that by now.

I’m not going to address (or even read) this strange sophistic rant on D2Y vs. me.

My google-fu disagrees:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC212702/

Let me be clear: prior to this discussion, I did not know that genes for skin color were hard to find. I looked it up once I realized that it would be a good benchmark for the utility of genetic studies for height. I was open to results that either supported one view or the other.

The result I found was as I stated. There are a few known genes but they account for only a small part of the variation in skin color.

Now, it may be that some progress has been made on it since ’03 but even if so the fact that it was so hard to make that progress shows that genetic studies have a really hard time with finding genes for traits that are determined by numerous genes.

“I have addressed the selective breeding of animals in a number of places.”

Only to give your opinion – with nothing to render it remotely palusible – that humans and animals might be different in this regard. Also, you argued that certain dog traits might not be as genetically determined as people think. Height was not one of them.

“Are you aware of the secular change in height in european and north amnerican populatinos from the 19th trough the mid 20th century?”

Of course. Accordingly, I have stated a number of times that the environment does have an effect. I never disputed that. The question is whether genes also have a significant effect.

daedalus2u wrote: “All I want is for people to keep an open mind and be able to say they don’t know what is causing things.”

That certainly puts you in direct opposition to Greg’s strong claims. I would say, though, that there is enough evidence to go further.

Let me summarize:

There are 3 basic hypotheses about height (and IQ, etc., but let’s stick with height):

G: It is mostly determined by genetics.
E: It is mostly determined by the environment.
B: Both are important (I’ll assume the effects are similar).

There are 4 lines of evidence (and a bonus factoid) that potentially bear on the question:

1) Genetic studies, which have found little.

Result: Inconclusive. Genetic studies don’t work well at all for multi-gene traits.

Conclusion: none

2) The secular increase in height from 19th-20th century.

Result: Rules out G. G is a straw man.

What about E vs B? It really doesn’t tell us anything.

Yes, environmental influence can be strong. And yes, variation in the environment would be enough to create differences between people in the same town.

But remember, the deviation due to independent variables adds in quadrature. That means that if the environmental variation creates a standard deviation of 5 inches, and if the effect of genetics is equally strong, then the net standard deviation becomes 7 inches. In practice, we couldn’t tell the difference.

Conclusion: E or B

3) Height tends to be similar in related individuals.

This would seem to favor genetics, but it has been argued that environment would also tend to be similar for related individuals. I think the trend is far too strong for that excuse to work, but I don’t have proof at hand.

Conclusion: Disfavors E, but not conclusive

4) Animal models; selective breeding has a powerful effect on height in various animals.

It’s true that not all species are the same, and humans could concievably be different from most animals in this regard, but that seems rather unlikely.

Conclusion: Strongly disfavors E (and strongly disfavors G, if we consider the details)

bonus) Hormones and other chemicals affect growth. Genes can certainly the relevant chemicals, as can environmental chemicals.

Conclusion: This is a hand-waving argument, but for whatever it’s worth, it favors B.

Net conclusion: B is the most likely hypothesis.

Height is a good example. There are effects on height that are not due to genetics and are not due to food consumption. But forget about trying to get funding to study them. The secular change in the past 150 years is a good example. People always say â??dietâ?, but people were not starving 50 years ago.

There is a related effect on the size of farm animals that is known to not be genetic and is also not due to food availability. When farm animals are given antibiotics, they grow faster, mature sooner, and end up bigger while eating less feed. It can’t be genetic because it happens to cows, pigs and chickens. All of those animals are fed â??ad libâ?, that is they can eat as much as they want. Why does feeding antibiotics make them grow faster? Good question, that no one knows the answer to.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC153145/?tool=pubmed

I think it relates to nitric oxide, and that antibiotics are doing to farm animals what anionic alkyl sulfonate detergents, conditioning shampoo, and anti-microbial everything is doing to humans, making them grow faster, mature sooner, end up bigger while eating less feed.

Also, my study in dogs is not on the heritability of height but on the heritability on a continuously varying disease trait, which is widely felt to have a polygenic causative component (in addition to environmental causative components). I took your (and other posters) comments to mean that the concept of polygenic inheritance and heritability estimates of ANYTHING in ANY species might be bunk (that you might be challenging the validity of the polygenic model and heritability estimates themselves rather than their application for this particular trait in this particular species – which, if true, would torpedo my thesis a bit, so I wanted to clarify the exact nature of the objection).

You are quite right that the presence of additive variation in one species would not imply it in another (indeed, in dogs I argue that the presence in one BREED does not imply it in another). I apologise for being unclear and for misconstruing (I think!) your argument.

What I find amazing, simply astonishing, is that a phenomenon like that can be observed by everyone, very easily, and the demand for a genetic explanation is so strong still.

It’s like truly, deeply, almost religiously believing that a particular person is going to die of a heart attack. then you see them waking across the street and get flattened, kiled, mushed, by a truck. Then we start looking for an explanation of how a heart attack could have caused the truck to run the person over.

Having said that, I don’t know that this has anytying at all to do with dogs.

There are many species of mammals where we see over secular time a tracking of environment with body size, and many where we see a tracking across space (of habitats) as well (famous examples = African elephants and African buffalo, etc.)

Given that humans clearly, undoubtedly, shift stature over generational time … and I have no doubt that this is a biological phenomenon … without shifting allele frequencies (most of the time, in most cases, though there are “candy bar” exceptions) then perhaps it is possible that some of these other mammals also track the environment and change body size intergenerationally. Perhaps several species do this. Perhaps it is rare. Perhaps its a primate thing. Whatever.

There is no reason to force a genetic explanation on one species becasue one sees it in another. Lots of phenomena on the surface seem very similar but have different causes. It’s called equifinality.

I’d be interested to hear more about your dog study.

or are we saying that until there is direct evidence of exactly what these alleles are (preferably with understanding of how they function biochemically) that we should retain a null hypothesis that variation in human height is entirely environmental in origin, given we KNOW environmental factors can effect human height and we only SPECULATE (with heritability estimates etc.) that there might be additive variation.
(Presumably then heritability estimates as defined by the proportion of additive variation to total phenotypic variation are not enough evidence, either because they are not in of themselves sufficient to override the null hypothesis or they are unable to truly differentiate additive variation from variation arising from the environment – if the latter, is this unique to humans due to family/population structure?).

My thesis in progress involves heritability estimates (in dogs) and so the underlying argument is not trivial to me (even though I don’t care about the inheritance of human height itself. I only started looking into it because I was annoyed with my daughter’s paediatrician). I would therefore really appreciate it if some kind soul could hit me over the head with a clue stick.

Also, where on earth is the money coming from to do studies like the one linked above? I am having a total failure of imagination about the utility of research into human height in the first place.

This is how identical twins can be discordant for gross disorders like anencephaly. There is a lot of epigenetic programming in utero that has life-long effects.