This reasoning is a fallacy.

You can make precisely the same argument about the last common ancestor between humans and chimps, or the last common ancestor of humans and neandertals. In a general context, chimp brain is as complex as ours. Yet evolution happened in between, we can track it, and we can see that it did in fact modify the cognitive functions of that ancestor; chimps are not humans. And hell, even the developmental machinery that makes an egg develop into an adult vertebrate is complex and interdependent, and if what you are quoting were true, one would expect all vertebrate life remain at the stage of a fish.

The actual reason might be much more mundane: the initial small population of modern humans expanded so rapidly that any resultant genetic differences between populations are the result of neutral evolution (like genetic drift) rather than natural selection. This is also why genetic diversity is inversely correlated with geographic distance to Africa. Essentially, we are all still this same small initial population, but we expanded like a balloon, taking down - directly or indirectly - any other populations that might have existed at times (like the neandertals, denisovians and many, many other hominins).

Not even that. I would rather say: humans with these mutations had a higher chance to leave offspring. It's not like we are talking about a single mutation that is present in all humans native to Siberia, but rather that a frequency of certain genotype is higher in these areas.

Yes, it is the same mutation you are talking about. The associated mutations (or "snips", SNP -- single nucleotide polymorphisms) are all the same, even in the West African tribes, and are thought to be of a common origin.

However, there actually is a known case convergent evolution of lactase persistence, fully described in this Nature Genetics paper: http://www.nature.com/ng/journal/v39/n1/full/ng1946.html . The authors analysed genotypes of East African pastoral tribes where lactase persistance is also widely spread, and found several alternative mutations in the same regulatory region. The most common of these mutations is thought to be ~ 7000 years old.

Of course not only in the mitochondria. Reactive oxygen species are also one of the first lines of defense against bacteria; macrophages generate reactive oxygen in the phagosomes - when macrophages ingest a live bacterium, it then becomes surrounded by a membrane, forming a phagosome. It is best explained with a picture: http://textbookofbacteriology.net/Phago.jpeg

Next, the phagosome changes into what is called phagolysosome, which is like a death cage for the bacterium. All sorts of nasty enzymes and molecules get injected in this space, including enzymes producing reactive oxygen species.

However, immune cells can also release oxygen directly into bloodstream. Given the cytotoxicity of ROS, this is like detonating nukes in your blood vessels, and can result in collateral damage. However, this stuff totally happens. Neutrophils can undergo a sort of harakiri, releasing DNA-bound bacteriocidal proteins in form of a sticky NET ("Neutrophil Extracellular Traps"). Proteins in NET generate extracellular ROS. It's a bit like gutting yourself and strangling the enemy with your own intestines.

So yeah, our own immune system produces ROS when it is fighting bacteria. The desired effect is local and limited to bacteria, but collateral damage is known to exist, and antioxidants can help to contain it.

That said, the hypothesis is not even that (a proper hypothesis), but more like a vague idea: how could you test it scientifically?

Post scriptum: erratum and an additional explanation

"cognitive and brain development" -> "cognitive functions and brain development"

"gene is not important" -> gene itself might be important, but its precise sequence might not matter because different variants are able to fulfill the same function. The problem is that a synonymous mutation usually is not visible to natural selection, but that doesn't mean that a non-synonymous mutation is always visible; many non-synonymous mutations are effectively neutral.

Also note that the whole story works only for protein coding genes, because we can easily tell "important" (non-synonymous) from "non-important" (synonymous) changes. However, first of all there are many important genes which do not encode proteins, for example the regulatory microRNAs or structural RNAs. It is not easy to tell which mutations are neutral, and which aren't. Second, there are regulatory regions that can matter a lot; again, it is hard to tell which mutation will have an effect. For example, the famous lactase persistence mutation is a mutation in a regulatory region, not in the gene itself; it messes up the switch of the gene that all other mammals use to turn off lastase production in adult animals. It does not alter the sequence of the protein coding region.

Such cases can still be worked on, for example by looking whether there are mutations in regions that are otherwise conserved in other species. Unfortunately, this is nowhere as sensitive or specific as considering the dN/dS ratio.

1. Just like the article says and unlike the Slashdot summary suggests, shiver-free thermogenesis is old and all mammals share it.

2. The researchers found traces of positive selection in a gene involved in shiver-free thermogenesis.

3. How do you look for traces of selection? A mutation in a DNA fragment coding for a protein can have two effects: either it changes the corresponding amino acid in the protein sequence (non-synonymous mutation), or it does not (synonymous mutation). This is because genetic code is redundant and different codons code for same amino acids, so a change from one codon to another does not have to change the protein. Synonymous mutations are assumed to be neutral for evolution (although they are not, not always).

Now, if you look at many possible variants of a gene and collect many different mutations, you can calculate whether the ratio of non-synonymous to synonymous mutations (called the dN/dS ratio) is (i) higher (ii) lower or (iii) quite like expected. Depending on the outcome of the test, you can say:

- if it is higher than expected, then there is a positive selection force at work (the gene is pushed towards change)
- if it is lower than expected, then we have a case of purifying selection; the gene is being actively maintained as it is, and any non-synonymous mutations are being removed from the population
- if it is neither lower nor higher, the gene is just not important

4. So, nice, you found that a gene related to non-shiver thermogenesis shows traces of positive selection. So what?

The answer is, not much. You do not always know which mutation was the one being selected. And even if you can pinpoint it, very often you will not be able to say what it actually does. So fine, you have a leucine replaced by arginine at position 186 in a protein chain; you might be able even to model the new sequence and see a delicate shift in the structure of the protein. How does it relate to the protein function? What has been modified or improved? No idea.

5. OK, why is that important? It is important because much of the genetic variability of the humans that we know is thought to have been fixated by genetic drift and other neutral evolutionary effects (like surfing the wave of colonization) - rather than selection. There are few examples of selection known. Light skin is one of them, and is thought to be an adaptation to the vitamin D deficiency caused by lack of sun at high latitudes. Mutation that keeps lactase being produced throughout life is another one. There were independent (convergent) events in both cases, by the way.

Look, humans are special. Special in the sense that humans are genetically extremely uniform, and the genetic differences between, say, native Australian, a blond-haired, blue-eyed Swede and a member of the Mbuti people from Africa are all together much smaller than between two chimpanzee individuals from groups living a few hundred kilometers apart. And moreover, these few mutations specific for some people but not for other seem to be more or less neutral in their character.

Finding differences that are *not* neutral, that are actually doing something is therefore an interesting thing. Notably, the few existing differences like that are linked to mundane things like metabolism or immune response (yes, some people are special because they don't fart after drinking milk, how is that for a superior race), and not, for example, to cognitive and brain development. The latter differences are found between humans and other primates.

As mentioned, this is philosophy, not science, so it is hard to confront different and contradictory views.

For example, there will be some who will tell you that the popperian concept of falsification is one of the most influential ideas of the XX century, responsible for the dramatic progress in understanding of our world that happened in the XX century. Not least because of the statistical hypothesis-testing framework, and there is little doubt about the influence of XX century statistics on progress in science.

But yes, it does not fit the copernican revolution.

Oh, what a beautiful quote, I will steal it for my students. It is simply grand... given the career and persona of J.B.S. Haldane, son of J.S. Haldane, who was one of the most influential evolutionary biologists and geneticists of all time.

Firstly, please note that Thomas Kuhn's view of how science happens is one of many possibilities. On one side of the spectrum, you have Popper and his younger collegue, Imre Lakatos; on the other end, you have Feyerabend and his "everything goes". Unfortunately, all that is philosophy, so itself is not science and cannot be verified experimentally or backed up with meaningful statistics. Thus, depending on whom you talk to, you will find arguments for Popper or for Lakatos or for Feyerabend or for Kuhn, all coming from the same field of science.

Personally, I value the popperian hypothesis-falsification paradigm a lot, especially since it fits so nicely with classical statistical hypothesis testing, and I insist on teaching it to students (I am a biologist), but I am well aware of its limitations.

Unfortunately, when reading texts of the great philosphers of science, one has the impression that all they really wanted to explain was "the big stuff", the grand theories, the grand revolutions or paradigm shifts. It is easy to argue for paradigm shifts if you focus on Copernicus and Einstein. It is much harder to immerse yourself in the day-to-day reality of scientific work, the millions of manuscripts generated, the propagation of ideas, their deeply intertwined relationships, as no idea, however genial, ever materializes itself from nothing.

I would not agree with that. Yes, very often, cat is not necessary, strictly speaking; for example, the list of arguments could be directly passed to grep in this case. However, personally I consider it to be bad style. A verbose cat (i) doesn't cost anything in terms of memory or CPU (if it does, then 1981 called, they want their computer back), (ii) facilitates or invites additional filters, for example adding sort -u (or sort | uniq, as I would rather have it), (iii) makes the whole pipeline more readable. I find it especially funny when people who do not hesitate to launch a terrible misuse of resources (also know as a "Web browser") hesitate to put an additional cat command to a long pipeline.

Since the bad guys have the code anyways, they should immediately publish the code as Open Source. Chances are, someone from the community will find the exploits before the guys who have stolen it.

This incident might also be used as an argument for open sourcing even critical code.

j.

Never done.
Yep, sounds informative.

j.

Reminds me of an old joke:

"So, how do you do these ships in bottles?"

"Well, I take some glue, some wooden sticks and some cord and throw it all into a bottle. Then I shake the bottle vigorously until the glue settles. Many funny things come out. Some of them are ships."

j.

As a Linux user, I was about to ignore the article when I glanced over the sentence "It affected not only Sun's Java but other implementations such as OpenJDK, on multiple platforms, including Linux and Windows. "

If I understand it correctly, all Java implementations have this flaw, so why write that it is a "MacOS vulnerability" and not "Java vulnerability"?

I want to know more how it affects my Ubuntu box!

j.