News & AnalysisMicrobiology

Microbial Survey of Human Body Reveals Extensive Variation

See allHide authors and affiliations

Science  15 Jun 2012:
Vol. 336, Issue 6087, pp. 1369-1371
DOI: 10.1126/science.336.6087.1369

The Human Microbiome Project charted the bodily whereabouts of thousands of bacteria, including this gut microbe, Enterococcus faecalis.


Though perhaps less contemplated than the navel, the inside of the elbow seems nonetheless comfortingly familiar. Don't be fooled. Our bodies are home to trillions of microbes, and that patch of skin hosts an invisible ecosystem whose microbial members were virtually unknown until now. This week, a consortium involving 200 investigators from a variety of disciplines describes that elbow-dwelling microscopic community as well as more than a dozen others as part of the most thorough look ever at the microbial world on and within us.

By the time it is completed next year, this Human Microbiome Project (HMP) will have spent $170 million cataloging the microorganisms—and to a lesser extent, their genes—that live on 18 human body sites, sequencing 3000 relevant bacteria, and determining differences between the microbiomes of healthy and unhealthy people. “It's a great framework, reference, and tool to really understand the microbiome,” says microbial ecologist Maria Gloria Dominguez-Bello of the University of Puerto Rico in San Juan.

The U.S. National Institutes of Health (NIH) started HMP in 2007 to jump-start a young field. When the effort was first conceived, researchers had hopes of generating a reference healthy microbiome, on par with the reference human genome, that the community could use to assess the role of microbes in health and disease. Two reports this week in Nature and 15 others in PLoS ONE and other journals, all from the HMP consortium, speak to the naïveté of that idea. “There's not a single reference human microbiome that we can compare everybody to,” says George Weinstock, an HMP researcher at Washington University in St. Louis in Missouri. “You can have two healthy people whose microbiomes are fairly different.”

The research released this week comes just as NIH is debating whether to launch an HMP follow-up and European nations are considering a successor to their own ambitious human gut microbiome effort, Metagenomics of the Human Intestinal Tract (MetaHIT) (Science, 8 June, p. 1246). For the HMP studies just published, researchers took samples from more than a dozen body sites covering the skin, nose, mouth, and gut in 242 healthy people. (In men, 15 sites were sampled, in women, an extra three in the vagina.) They excluded participants with any signs of illness, even minor gum disease, as they wanted to get a sense of what a healthy microbiome looks like. Some people provided up to three samples over a 22-month period to assess whether their microbiomes were stable over time. In each sample, the researchers probed for copies of a DNA sequence, the 16S ribosomal subunit, that is used to identify organisms. And in 1300 of the 11,000 samples, they sequenced all DNA present to get a sense of what genes were at a particular body site.

Spice of life.

The different body sites studied by the Human Microbiome Project show extensive variety in their community compositions.


HMP to date has generated 3.5 terabases of data, more than 1000 times the amount produced by the original Human Genome Project. “HMP is about the same size [as] all the previous microbiome studies put together,” says Curtis Huttenhower, a computational biologist at Harvard School of Public Health in Boston. “No computational methods existed to analyze that scope of data.”

“It's an impressive data set,” adds Dusko Ehrlich, who coordinates Europe's MetaHIT, whose results were published last year.

Despite the lack of a clear-cut reference of a healthy microbiome, some patterns have emerged from the HMP data. They provide “dramatic and gratifying degrees of support for what were just some early and tantalizing leads” about our microbial partners, says David Relman, a microbiologist and infectious diseases clinician at Stanford University in Palo Alto, California, who is not part of HMP.

For one, no two microbiomes are alike. “There is a personalization of the microbiome that takes place in every individual,” says HMP investigator Barbara Methé, a microbial ecologist at the J. Craig Venter Institute in Rockville, Maryland. There's also an incredible diversity of life across the body of each individual. The microbes on the forearm are starkly different from those on the teeth and equally different again from those in the gut or in the vagina.

Yet there are similarities across different people in, say, the forearm microflora, indicating that our microbial partners are not randomly distributed. Each body site had a few core or “signature” bacteria living there, with characteristic genes linked to that particular body site. “Different parts of the body see different configurations of organisms,” Methé says.

Once they established the community makeup of each body site using the 16S data, the HMP researchers examined whether any of the microbes were known pathogens hiding out on or in the healthy subjects. They found no 16S genes from deadly germs such as those that cause cholera or botulism. But lurking in some places in some people were about 100 opportunistic pathogens, microbes that are known under certain circumstances to become dangerous. For example, they confirmed previously discovered hideouts of Staphylococcus aureus, infamous for having antibiotic-resistant strains: 30% of the subjects had this bacterium in their noses and 5% had it behind their ears. And they discovered hideaways of other less-well-known pathogens, Huttenhower says, such as Gardnerella vaginalis, Haemophilus influenzae, and Treponema denticola.

For a subset of the body samples, the researchers went beyond simply doing 16S identification of the microbes present. They fully sequenced most of the sample's DNA, trying to reveal the full array of genes available to microbes. Huttenhower and his colleagues assigned those genes to various cellular or metabolic pathways and thus were able to reconstruct how these microbes were functioning on the body.

That analysis revealed that even if the specific microbes present differ between body sites, core sets of microbial functions are shared by those sites. At each one, “the bugs are very different but the pathways are very similar,” Huttenhower says. “If you go to different cities, you have different people with different last names and different backgrounds, but every city has people who are the bankers and the people who run the subway.”

Some body sites had more bankers than subway workers, however. There were many microbial genes for simple carbohydrate processing in the mouth, where food flushes through quickly, for example. In contrast, the gut had more microbial pathways for processing complex carbohydrates.

It may eventually be possible to diagnose healthy microbiomes by the metabolic pathways they have functioning in them and to treat unhealthy microbiomes by restoring missing pathways or targeting pathways that are out of whack. “It could be a process that could be carried out by different bugs in different people,” Huttenhower says, so it may be most efficient to go after pathways rather than the microbes themselves.

Some researchers contend that there's too much variety among human microbiomes to think in terms of a reference microbiome. But Methé argues that the HMP data establish the right range of organisms and pathways for each body site and thus do establish a reference of sorts. “If we can begin to understand what we think should be there and in what proportions, you can begin to tell the difference between a healthy state and a disease state,” she says.

Because the functioning of our bodies is so intimately linked to the microbiome, the latter's DNA has been referred to as the “second human genome.” But deciphering this second genome has proven a much tougher challenge than our first, Relman says. HMP “is the Human Genome Project to the 1000th power.”

View Abstract

Navigate This Article