Feeling sleepy?

Share on FacebookShare on Google+Email this to someoneTweet about this on Twitter

Nature Genetics this week has published a genome-wide association study of narcolepsy in the Japanese population. The finding in the paper is a variant that confers a modest risk of narcolepsy, but personally, I was blown away by Figure 1, reproduced above. The figure shows the strength of association of each of 500,000 SNPs with narcolespy, and the novel reproducible finding is on chromosome 22 (ie, it doesn’t stand out all that impressively in this plot). The major signal, absolutely swamping everything else, is in fact in the MHC region (called HLA in humans).

This region, of course, contains risk factors for type I diabetes, crohn’s disease, and most (all?) other autoimmune diseases. A quick google confirms that, indeed, thinking of narcolepsy as an autoimmune disease is not new, but it’s definitely new to me, and it’s pretty striking to see just how much more important the risk factors in HLA are compared to everything else.

Labels: ,


  1. Paging Gregory Cochran and Paul Ewald to the white courtesy phone…

  2. Matt, I’m going down a totally different track. Maybe this is old stuff to you gene guys, but I don’t recall it being discussed on this blog: The idea of overclocking the immune system. Certainly we’ve experienced a lot of sudden evolutionary pressure on the immune system in general, and it would seem to me that the built-in advantage of rare alleles in the immune system would make it especially subject to overclocking.

  3. David,  
    Maybe. At the very least, the radically polymorphic nature of immune system loci makes deleterious gene-interaction stories a lot more plausible there than anywhere else. But when I see autoimmune my first thought is “molecular mimicry” because it’s more elegant and we have documented instances.

  4. overclocking the immune system 
    But calling it overclocking would imply that this galaxy of disorder-associated SNPs is a recent response to increased selection for disease resistance. That *could* be true (the selection pressure has certainly been there) – but it’s also possible that the massively polymorphic MHC/HLA is simply the best setup for an immune system in social mammals generally, that is, this situation could be tens of millions of years old.

  5. Anyone interested in narco should check out the hypocretin findings. Fascinating story.

  6. A narcoleptic layman asks: does this research prove that narcolepsy is congenital, i.e., does it disprove the trauma theory of narcolepsy causation?

  7. At my company there is a scientist who does nothing but analyze SNP’s in the HLA region–mostly for childhood leukemia and sex-selection during pregnancy (more boys than girls conceived but more boys miscarry so birth rates even out). I was surprised to find out how much is going on at that locus, but a link to narcolepsy is certainly surprising and interesting.