Nature has two studies on the difference in gut microbe composition between lean and obese mice and people (mice study, human study). Read their news department's review here to get the skinny (c'mon, you had to have seen that one coming!). In brief, the balance between two divisions of bacteria -- Bacteroidetes (B) and Firmicutes (F) -- is pronouncedly different in lean and obese mice and people, such that lean ones have a greater proportion of B microbes and a lesser proportion of F microbes compared to obese ones. The F to B ratio in obese humans was about 12 times that of the lean humans, but decreased over the course of a year-long diet (though not reaching the levels of lean subjects). And after mice raised in sterile conditions were given gut microbes from lean or obese mice, the ones given obese microbes gained more weight than those given lean microbes within just 2 weeks (1.3 +/- 0.2 g vs 0.86 +/- 0.1 g, respectively), even though the two groups did not differ in the amount of food they ate -- the obese microbe group simply absorbed more energy from food. Though small absolutely, this difference in weight gain presumably accumulates over longer periods of time.

I blogged here about some studies earlier this year that investigated obesity from an infectious angle, which implicated a human adenovirus. Let me reiterate my skepticism regarding thrifty gene ideas about obesity (more below the fold).

The basic idea is that in times of famine, an allele that increased an individual's ability to harvest more energy from food would have been advantageous, although when thrust into a new environment of abundance, the allele will tend to make its bearers obese. The prediction, then, is that populations for whom famine has been the strongest burden over time should have higher frequencies of such alleles, and thus in the modern world should have a higher frequency of obesity. Because famine arose from agriculture and civilization, the concrete prediction is that Eurasians should be most obese, and non-Eurasian populations falling below depending on when they adopted agriculture. However, in the US, Europeans are least likely to be obese, followed by African-Americans, and then Native Americans and Mexicans. The latter two (New World) groups have not been practicing large-scale agriculture long at all, and thus have not been as subjected to famine as other groups, yet they are the most likely to be obese. See here for obesity rates by gender and race.

Thus, obesity looks like an infectious disease whose microbes have resided longer in Eurasia than in sub-Saharan Africa or the New World.

Note that the microbes involved need not have been disease-causing in Europeans -- if the two Nature studies are to be believed, obesity is associated with a bacterial balance being thrown outta whack. That suggests that the "microbiome" in the gut is like one large co-adapted gene complex. Just as hybrid depression can result from disrupting such complexes, co-habitation of races separated for a long time might introduce gut microbes from one group into another, disrupting the balance of the latter's gut microbiome and/or its interaction with the human system around it. That is, if human guts co-evolve with gut microbes, Native American guts may not know how to work optimally with the new European gut microbes. This clearly isn't all that happens, though, since people of European and Native American ancestry have been living next to each other in the US for centuries, yet obesity has only recently flared to epidemic proportions. But something like this is likely at work, and likely accounts for some of the racial variation in obesity.