Saturday, April 4, 2015
I question the thesis of an article "How Europeans Evolved white skin" presented by the American Association of Anthropologists
I decided to scrutinize a story, by an AAAS contributor for SCIENCE, itself based on a scientific article that appeared on an AAAS web site. The article was based on a study done by Iain Mathieson, and David Reich, and it is apparently in pre-publication at an archival server called bioRxiv. This is an excerpt from the AAAS article which I took issue with on the blog http://news.sciencemag.org/archaeology/2015/04/how-europeans-evolved-white-skin.
"Now, a new study from the same team drills down further into that remarkable data to search for genes that were under strong natural selection—including traits so favorable that they spread rapidly throughout Europe in the past 8000 years. By comparing the ancient European genomes with those of recent ones from the 1000 Genomes Project, population geneticist Iain Mathieson, a postdoc in the Harvard University lab of population geneticist David Reich, found five genes associated with changes in diet and skin pigmentation that underwent strong natural selection." -SCIENCE
I evaluated this paper based on causality analysis, and I believe that the central hypothesis that it is based on has serious issues. My disagreement is not with the genetic data itself, it is the interpretation of the data that I take issue with.
I was hoping to have a response from the Harvard authors of this study specifically, Iain Mathieson and population geneticist David Reich, but I ended up engaging a bunch of others in simultaneous debate.
And this is what I wrote (in italics below):
The thesis of the article's paper, however is more than simply (about) the data. It is the interpretation that the present genes sequenced are present because they conferred a selective advantage to their hosts. That, as I've said below, fails to acknowledge the selectability of the other genes, such as Bcl2 and P53, certainly present 40,000 years ago and found in all mammals. Skin color has no selectability for these genes specifically, and yet tthey co-develop along with the other 30,000 genes in the human genome.
I received some replies but nothing convincing to me that there is an answer to my questions. I'll just call some of the other responses "some other replies" and you can see them for yourselves on the actual blog but I'll spare readers details here:
A well meaning scientist who described himself as a quantitative geneticist, with a PhD, replied, and I'll paraphrase here, that "the author's in the article were somewhat "lazy" in not differentiating alleles from genes, (which implies that the Bcl2 and P53 are not selectable in this case as they are too large). He also stated in so many words, that I was unclear on the concept of what a polymorphism is, and thus, the two genes that I described above, would not be relevant to selection in skin specifically, as they are much larger genes and critically he argued, ANY slight change in these genes would be terminal.
Needless to say this thinking is simply not backed up by experimental evidence, which I'm fully aware of, but this is what I wrote:
I do understand the distinction you make about polymorphisms. A significant change to Bcl2 or P53 would be terminal whereas the genes [of skin pigmentation] discussed here are perhaps "less critical" they can vary much more in sequence without drastic consequences. However, I would disagree that these are entirely non-polymorphic, since Bcl2 in particular is believed to be directly responsible for cancer resistance in addition to mediating the onset of cancer in ways that aren't understood. P53 is a similar gene, and I'm speaking again about its cancer relatedness with respect to how it controls transformation. Because cancer is prevalent, afflicting 1 in three people, these are subtle polymorphisms would be selected for by a variety of factors. I don't agree with the "these genes are untouchable". Every gene or even allele has a history. The point is that skin cancer is intricately related to sun exposure, but also genetics and the operation of the tumor suppression capabilities of bcl 2 and P53, I could argue that a frequency of cancer would in fact be selectable. Skin cancer would be highly detrimental to fitness. My other point is why don't we all have cancer? a rather extreme view, but hypothetically a good question. You, a quantitative geneticist see function of genes, I see chemistry. But the fact that cancer is so complex and effects people differently, (has different triggers sun exposure, etc) says that genes like Bcl2 and P53 are mutating enough to create such divergence and hence fitness. Genes express proteins, they don't "know" if they're polymorphic or not.
You might ask OK what does polymorphic mean? So for readers, who aren't in the "know" what he's saying is actually a well known convention of evo theory, and that is to say that certain genes are generally "off limits" they're "safe" so to speak, not because they can't change, mind you, but because if they did even change even slightly, it would be terminal to that organism at birth or even before, so in other words the genes are never carried forward, neither would the modification to Bcl2 or P53.
I continued with that point in another response:
Those two genes I mentioned are specific to skin's function, like all other cells. Hence, they are relevant to the discussion. My point is that these other gene's presence in descendants (the ones you mention) are equally critical to the thesis that some factor such as vitamin D, caused their frequency to increase. Bcl2 and P53 are critical to skin cell's ability to avoid and fight cancer by regulation. The unanswered question that I have already posed is "Why aren't
these genes also mentioned in the paper?" They are more critical to skin health and survival of their hosts than the genes discussed i.e. SLC24A5 and SLC45A2. So I would like an explanation as to how these genes were also selected for along with the genes mentioned in the paper.
You would not be reproducing without intact and highly active forms of these two genes, period. I've yet to have a reply from a scientist here on this point,which it seems you’ve missed entirely. We would not conclude that skin color was the advantaged trait, as the authors do, since these other genes which
confer no color or visible trait, (other than not being alive perhaps?) were obviously more important to vital function of skin cells and the organ and had to be inherited in any individual. Of course it would be ideal to have the authors of the paper, Iain Mathieson or David Reich provide an explanation on this specific point.
Another responder, (which I'll again paraphrase) provided an explanation for the change in pigmentation based again, on the already well accepted model that Vitamin D is a necessary nutrient, and thus, the possession in the skin of a gene which makes it, would be an advantage. He also made a comparison between this selection mechanism and that of bear's coats (white in Northern latitudes for camouflage/hunting advantage on ice) vs other bears "darker coats" in warmer regions. Of course, that part about Vitamin D, is exactly what the article is already stating.
This was my reply:
The argument that you present here is "smoking gun" for teleological, "bears have heavy coats because they 'must' stay warm in the winter." Dawkins in "selfish gene" theory avoids this argument by use of gene "selfishness". In nature there is no 'must' only chemistry and probability. The basic thesis of the article's paper, however is more than simply the data that specific genes were traced through varioius generations. It is the interpretation and CONCLUSION that the present genes sequenced are present because they conferred a selective advantage to their hosts. That is what is critical about this paper. That, as I've said below, fails to acknowledge the selectability of the other genes, also involved critically in skin tone survival, such as Bcl2 and P53, certainly present 40,000 years ago and also found in all mammals if not other vertebrates. Skin color has no selectability for these genes specifically, Bcl2, at least as stated in the paper, and yet tthey co-develop along with the other 30,000 genes in the human genome.
Many other commentators simply resorted to responses that were either meant to censor my posting or to drive me away from the site, and (for now) I won't bother reproducing here. So in reply to those types of posts, which are attempted to attack one's argument by "poisoning the well" this is basically where I left it:
I'm a PD/PI, that would be a professional scientist, commenting on the specific AAAS article above "How Europeans Evolved White Skin" by author Ann Gibbons (correspondent author for Science magazine) with a specific question relating to Bcl-2 and P53 gene function, which relate to the health of all cells, including those in skin, and which appear to be neglected in this referenced paper, and yet would be vitally important to the question of skin pigment development via their regulation of cellular mutation response. How does selection of Vitamin D account for them? Unfortunately, the only response to my question thus far, comes from a mutant troll on this site, who rambles on about the "bible" and who knows what else. So, are there any REAL scientists who would like to engage in a discussion about this AAAS referenced article? And if not, what is it that you fear about scientific questions? My question is for the moderator/editor Ann Gibbons. Let's raise it a level shall we?
In summary, the central thesis that the five genes, primarily genes relating to Vitamin D are somehow causative is what I am questioning in regards to the other critical genes Bcl2 and P53, which are never mentioned in their paper.
I did not receive a succinct response to my argument addressing the articles central thesis. And what I'd hoped might generate a "breakthrough" on this point, perhaps in a live forum of dialogue, fizzled. But I will update this post if a cogent response from the authors does come through.