Evolution with a bullet.

There is a lot of buzz about the recent (but still unavailable [update: link]) PNAS paper by John Hawks et al. reporting an accelerated rate of natural selection in humans. This time, I am not going to pick on the media who, predictably, are selling this as a conclusive finding when those of us in the scientific community have not even had a chance to read the paper yet, let alone for anyone to try to critically assess it. It may be fantastic work, and if it holds up I will certainly give it a significant place in my lecture on human evolutionary history. My complaint is about what the authors themselves have been telling the press.

“Ten thousand years ago, no one on planet Earth had blue eyes,” Hawks notes, because that gene—OCA2—had not yet developed. “We are different from people who lived only 400 generations ago in ways that are very obvious; that you can see with your eyes.”

Interesting idea. But I suppose at no time in history could there have been another variant that caused blue eyes? Is this really the sort of thing that can only evolve once and in one way?

“We aren’t the same as people even 1,000 or 2,000 years ago,” he says, which may explain, for example, part of the difference between Viking invaders and their peaceful Swedish descendants.

Um, ok. I wonder if blue eyes make you peaceful?

Harpending says genetic differences among different human populations “cannot be used to justify discrimination. Rights in the Constitution aren’t predicated on utter equality. People have rights and should have opportunities whatever their group.”

And, by implication, the groups are not utterly equal. Care to speculate on which groups are more equal than which others?

The new study comes from two of the same University of Utah scientists – Harpending and Cochran – who created a stir in 2005 when they published a study arguing that above-average intelligence in Ashkenazi Jews – those of northern European heritage – resulted from natural selection in medieval Europe, where they were pressured into jobs as financiers, traders, managers and tax collectors. Those who were smarter succeeded, grew wealthy and had bigger families to pass on their genes. Yet that intelligence also is linked to genetic diseases such as Tay-Sachs and Gaucher in Jews.

No comment.

“History looks more and more like a science fiction novel in which mutants repeatedly arose and displaced normal humans – sometimes quietly, by surviving starvation and disease better, sometimes as a conquering horde. And we are those mutants.”

Michael Crichton’s latest: LACTASE, the harrowing story of a small mutation that conferred a slightly better ability to digest milk and reached a higher frequency in some human populations. Expect the movie in summer 2010.

“Five thousand years is such a small sliver of time – it’s 100 to 200 generations ago,” he says. “That’s how long it’s been since some of these genes originated, and today they are in 30 or 40 percent of people because they’ve had such an advantage. It’s like ‘invasion of the body snatchers.’”

Genes, alleles. Tomayto, tomahto. Either way, they’re out to take us over!

“We are always trying to outrun disease.”

And body snatchers.

“Natural selection cares about how many children you have. People will have kids younger and younger.”

Where’s Bart Simpson when you need him? Natural selection is not conscious. Natural selection is not conscious. Natural selection is not conscious. Natural selection is n…

“Genetic engineering will make all this irrelevant. If people want green-haired kids they will go to the doctor and get them in 100 years.”

No they won’t, because people will be marrying robots by then.

“We are more different genetically from people living 5,000 years ago than they were different from Neanderthals.”

MNSdfnklcn. Oops, sorry… that was Coke sprayed all over my keyboard.

“In the last 40,000 years humans have changed as much as they did in the previous 2 million years.”

Nxjbjbecbc. Dammit… again!

“We found very many human genes undergoing selection,” says anthropologist Gregory Cochran of the University of Utah, a member of the team that analyzed the 3.9 million genes showing the most variation. “Most are very recent, so much so that the rate of human evolution over the past few thousand years is far greater than it has been over the past few million years.” [emphasis added]

Really? A few million years ago there were no humans at all.


12 thoughts on “Evolution with a bullet.

  1. I’ve read the paper. You can find it [here].

    There’s no data in the paper. It simply contains a summary of the number of “selection events” they found in each of four populations. Most of the paper is an outline of the mathematical tests they performed on that data set.

    Unless you’re willing to spend an inordinate amount of time doing some background reading it’s impossible to judge the quality of their analyses. Even with the background reading it’s impossible to judge the quality of their data.

    The author quotations in the various press releases are enough to make me highly suspicious. If that’s an example of the quality of their science then the paper dos not deserve the attention it’s getting.

    This looks very much like a self promotion campaign where the actual scientific paper does not measure up to the hype in the press releases.

  2. I don’t like the way Hawks is hyping the paper either but

    1) They used (and cited) the public HapMap, not some secret data set. So it’s not fair to claim “There’s no data in the paper”.

    2) All papers outside one’s immediate field require “inordinate amount of time doing some background reading” to understand. That’s why peer reviewers are generally people working on very similar problems.

    3) “Most of the paper is an outline of the mathematical tests they performed on that data set”. What’s wrong with that? Welcome to modern biology — where analysis of data is as much science as bench experiments.

  3. Don’t PNAS papers usually state how the paper was submitted, i.e. through full peer-review or through membership of the academy?

    This paper doesn’t say. There is a footnote that says “Insert ’This paper was submitted directly to the PNAS office.’ when applicable.”

    There is another footnote marked as placeholder, so I guess this is not the finalised PDF yet…

  4. Jonathan Badger says,

    1) They used (and cited) the public HapMap, not some secret data set. So it’s not fair to claim “There’s no data in the paper”.

    The “data” is the 11,439 “selective events” that they pulled out of HapMap. I have no way of evaluating that dataset by looking at it myself to see if it looks credible.

    I can’t tell, for example, what the average size of a “selective event” is and how many SNPs it covers. I can’t tell how much overlap there is between the CHB and JPT groups—that’s something I’m curious about.

    In many papers like this the authors give us at least a few typical examples of their data.

    2) All papers outside one’s immediate field require “inordinate amount of time doing some background reading” to understand. That’s why peer reviewers are generally people working on very similar problems.

    I understand that, Jonathan. The problem here is that there are only a handful of people in the world who have the expertise to understand this paper. Probably fewer, since we can’t see the data.

    That’s a big problem, especially since the claims are extraordinary. It looks and smells like a snow job to me.

    3) “Most of the paper is an outline of the mathematical tests they performed on that data set”. What’s wrong with that? Welcome to modern biology — where analysis of data is as much science as bench experiments.

    I’ll tell you what’s wrong with that. We can evaluate the tests but we can’t evaluate the results of applying those tests to their data. Have you heard of the expression “garbage in, garbage out?”

    Jonathan, I’m not nearly as stupid as you make out. I’ve spent a good part of my life analyzing sequence datasets using some fairly sophisticated techniques. I learned one thing. It doesn’t matter how good your programs are if your sequence alignments are crap.

  5. The “data” is the 11,439 “selective events” that they pulled out of HapMap. I have no way of evaluating that dataset by looking at it myself to see if it looks credible.

    That “pulling out” (as you put it) was a large part of what the authors did; it’s one thing to argue that the methods to do this weren’t detailed enough to replicate; it just seems odd to call these intermediate results the “data” rather than what they started with.

    I’ll tell you what’s wrong with that. We can evaluate the tests but we can’t evaluate the results of applying those tests to their data. Have you heard of the expression “garbage in, garbage out?”

    Of course their methods might be flawed. Given that results are surprising, it’s certainly worth considering. But if there’s anything lacking for replication it would be in the description of the tests themselves.

  6. The claims in this paper may be correct or they may be proven incorrect.

    But your snide commentary just sounds pedantic and petty.

  7. Are you talking to me, “tensai”? If so, then please let us know which of the comments said by the authors to the press you agree with.

  8. Besides you, very few evolutionary biologists, science bloggers, or science journalists (with the exception of amateur hereditarians like Andrew Sullivan, William Saletan, and Steve Sailer) have noticed an obvious implication of the PNAS authors’ interpretations of and speculations about their data. These implications are in the paper, some authors’ statements to the press, their 2007 AAPA abstracts, and Cochran and Harpending’s history of statements on these matters.

  9. It doesn’t matter how good your programs are if your sequence alignments are crap.

    Well, technically, they’re not working with DNA sequences that need to be aligned. They’re using genotyped SNPs.

    I think the complaint that they didn’t use the phased hapmap data (because they weren’t available at the time the data were analyzed) is more important.

  10. Although without a doubt the press gets carried away with research findings (to be duplicated) as the hundredfold acceleration of human evolution, we scientists might wish not only to criticize but also to educate the press.

    Moreover, occasionally one may find some of the press-coverage of Post-ENCODE Genomics rather insightful.

    For instance, what would this blog audience (and with their help, the public) think about an accelerated evolution evidence vis-à-vis the coverage of ENCODE-release by The Economist
    ?

    [The Economist] What is being proposed is the inheritance of characteristics acquired during an individual’s lifetime, rather than as the result of chance mutations. This was first suggested by Jean Baptiste Lamarck, before Charles Darwin’s idea of natural selection swept the board. However, even Darwin did not reject the idea that Lamarckian inheritance had some part to play, and it did not disappear as a serious idea until 20th-century genetic experiments failed to find evidence for it.

    The wiggle room for the re-admission of Lamarck’s ideas comes from the discovery that small RNAs are active in cells’ nuclei as well as in their outer reaches. Greg Hannon, of the Cold Spring Harbor Laboratory in New York State, thinks that some of these RNA molecules are helping to direct subtle chemical modifications to DNA. Such modifications make it harder for a cell’s code-reading machinery to get at the affected region of the genome. They thus change the effective composition of the genome in a way similar to mutation of the DNA itself (it is such mutations that are the raw material of natural selection). Indeed, they sometimes stimulate actual chemical changes in the DNA—in other words, real mutations.

    Even this observation, interesting though it is, does not restore Lamarckism because such changes are not necessarily advantageous. But what Dr Hannon believes is that the changes in question sometimes happen in response to stimuli in the environment.

    The chances are that even this is still a random process, and that offspring born with such environmentally induced changes are no more likely to benefit than if those changes had been induced by a chemical or a dose of radiation. And yet, it is just possible Dr Hannon is on to something. The idea that the RNA operating system which is emerging into view can, as it were, re-write the DNA hard-drive in a predesigned way, is not completely ridiculous.

    This could not result in genuine novelty. That must still come from natural selection. But it might optimise the next generation using the experience of the present one, even though the optimising software is the result of Darwinism. And if that turned out to be commonplace, it would be the paradigm shift to end them all. [End of quote from The Economist]

    pellionisz@junkdna.com

  11. RPM says,

    Well, technically, they’re not working with DNA sequences that need to be aligned. They’re using genotyped SNPs.

    I know that.

    I was referring to the kind of work I’m familiar with and the point was that we need to see the data.

Comments are closed.