"it seems that the Duffy null phenotype is a recent adaptation to malaria among West Africans."


I'm beginning to doubt this. The gene frequency is too high. Assuming that Duffy has any cost at all (which looks likely considering the leukopenia and geographically restricted distribution), you'd expect the frequency to stall out at a lower level - people without Duffy would then coast on crowd immunity.

A virulent African zoonosis that Duffy blocks might make for a better explanation. The protection against vivax malaria would then be a coincidence...

Gcochran, you point out that a varmint with a nonhuman reservoir would be likelier to drive duffy-null near fixation - likelier than vivax. But it may be roughly as parsimonious to posit that the varmint was in fact vivax, but it has an undiscovered animal reservoir - or had one that no longer exists.

Even if no animal reservoir has ever existed, http://www.ajtmh.org/cgi/content...t/full/75/4/ 575
">this rather tentative 2006 paper reports that certain duffy-negative individuals are sparsely infected by probable vivax. Even if the fitness burden is zero in duffy-null carriers, their presence (assuming they really exist) might still impact the degree of herd immunity. I would think the degree of herd immunity would have to be rather high to bring the allele into balanced polymorphism. But of course it depends on the cost of the neutropenia etc.

The posited zoonosis could be a lot more lethal than vivax.

This is kind of off-topic, but I had this question ever since I first heard about Duffy, and given the number of knowledgeable readers here I thought I'd see if anyone would know.

Despite the well-recognized role of Duffy as an obligatory, or near obligatory, receptor for malaria, and given the immense public expenditure on research related to this disease, I cannot manage to find one single high-throughput (or even "low"-throughput) screen for small molecule Duffy antagonists in the literature.

This is all the more surprising given that Duffy is related to the GPCR superfamily that is well-known for containing excellent drug targets, and also that a similar strategy with the HIV receptor CCR5 is widely considered viable. I'm wondering if it's been tried and no hits have turned up.

On average, prospects for inhibiting protein-protein interactions with small molecules have been considered questionable. Apparently the set of successful examples, including the CCR5 antagonists, is quite small - though some say an acceleration is afoot.

Apologies if this is too off-topic, but whenever HIV articles come up, I always wonder about Duesberg and the other HIV-is-a-harmless-retrovirus folks. What do GNXP folks think of that stuff? I would be especially curious to here from gcochran.

Duesberg is crazy. AIDs looked like a contagious disease from day one and the agent was clearly identified not much later. And the drugs (protease inhibitors and such) designed to block HIV greatly alleviate the disease: what more could you want, egg in your beer?

Looking at Wiki, I found a wonderful sentence:

"For example, the magazine Continuum, which consistently denied the existence of HIV/AIDS, shut down when its editors all died of AIDS-related causes."

Only fair.

The FAQ on Duesberg's webpage has a high concentration of very low quality arguments - by far the highest I have ever seen from a person of his credentials. If Duesberg turns out to be right I will boil and eat my hat, then take my pants off and jump-kick a hive of killer bees. I recommend reading Steve Harris on the subject. He's a powerful and fair reasoner who likes to hash it out with AIDS denialists in great detail.

http://groups.google.com/groups/...harris& safe=off

You didn't have to know anything about molecular biology, epidemiology, or virology to smell the rat in Duesberg's argument. If AIDS was the result of recreational drug use, then it should have been rock musicians who were dropping like flies, not ballet dancers.

The sad case of Christine Maggiore and her daughter should have been the final, tragic nail in that coffin.

Razib- Does the contact authors link work? I sent you a note regarding a paper you may be interested in. Don't know if you received the email.

ghoghogol, it works now. thanks for telling about the problem.

To Eric: I've heard that about protein-protein interactions, but I just assumed people wouldn't write off the possibility here, because many GPCRs bind peptides yet are still "druggable".

And about Duesberg--I have had a chance to speak with him in person, and what I noticed was his tendency to use analogies that weren't based on any real conceptual similarity. For instance, he compares the genome to a book, with the chromosomes as chapters whose pages, or genes, are organized in a particular order that determines physiology on a big-picture level. He gives no evidence from studies of gene expression to suggest that they are actually ever read in this order, though, most likely because there isn't any.

If AIDS were the result of toxic antiviral drugs, we'd also expect it to have occurred in the animals we used for testing. It did not.

For Duesberg to be correct, HIV would have to be harmless, both under normal and immune-depressed conditions, for rats, monkeys, and all great apes other than man - and stay that way. And somehow become virulent only in humans once a previously-undetected immune depression caused by drug use gave it a foothold.

The real problem isn't that this scenario is ludicrously implausible, it's that there's never been a bit of evidence that supported it.

You're right, I didn't realize things like chemokines had GPCR receptors. I looked it up and some of them are 75 residues in their mature secreted form. Vasoactive intestinal peptide is 28 residues, enkephalins five. I have no clue which ones have been drugged though, and of course there are lots of much smaller GPCR ligands.

A lot of people say many more protein-protein drugs have been published in recent years than in the past. So here's hoping. It would certainly open up infinite possibilities. But that was also hoped in the past about combinatorial chemistry (which you gotta admit is an awesome idea) and other stuff.

I wonder if one factor isn't cost. Malaria drugs pretty much have to be extremely cheap. Many people think protein-protein inhibitors will turn out to average larger than the average drug, though I don't know if that is true of the ones that exist so far. Large drugs, if they have to be synth'd, are likely to be far dearer than small ones. And when you mention HTS I'm guessing you mean libraries that are primarily synthetic? My grasp of these matters is loose but I hear natural products are not really going strong these days.

One of the CCR5 inhibitors for HIV is only the size of tetracycline. Certainly smaller than taxol. But both of the latter are semi-synthed, like most antibacterials and opiates, and the antimalarial artemether.

I was mistaken, natural products are quite far from total eclipse:

http://pipeline.corante.com/ arch...d_unnatural.php