The dawn of the twenty-first century has seen the emergence of a major theme in biomedical research: the molecular and genetic basis of what it is to be human. Surprisingly, it turns out that we owe much of our biology and our individuality to the microbes that live on and in our bodies — a realization that promises to radically alter the principles and practice of medicine, public health and basic science. It is therefore appropriate that ever more research is focused on these microbes and their genes, which together are known as the human microbiome1. In this issue, the Human Microbiome Project Consortium2,3 publishes the most extensive catalogue yet of organisms and genes pertaining to our microbiomes.
The first observations of indigenous human microbiota were published more than 300 years ago, soon after the invention of the microscope. Today's view of the microbial world has been radically improved by DNA-sequencing technology. In the wake of the Human Genome Project, calls were issued1,4 for enhanced efforts to be made to characterize the 'second human genome' — the human microbiome. At the end of 2007, the US National Institutes of Health (NIH) launched the Human Microbiome Project (HMP) and, in early 2008, the European Commission and China initiated the Metagenomics of the Human Intestinal Tract (MetaHIT) project. Other countries have begun similar ventures, motivated in part by an interest in better defining their biological heritage.
Two studies, by Huttenhower et al.2 (page 207) and Methé et al.3 (page 215), together with 15 other papers5,6 that are being published simultaneously elsewhere, comprise the first reports of the HMP Consortium research groups. The primary data, as described by Methé and colleagues3, were derived from samples collected from 242 healthy adults in the United States, at 15 (for males) or 18 (for females) body sites — from the skin, nose, mouth, throat, vagina and faeces (to represent the distal gastrointestinal tract). Each person was sampled up to three times over 22 months, generating a total of 11,174 samples.
The consortium researchers obtained the nucleotide sequence of the small-subunit ribosomal RNA — a molecule found in all cellular life — from microorganisms in 5,177 of these samples3. These sequences are commonly used to infer the genetic relationships between organisms. The researchers also surveyed the genomes of the microbes in 681 of the samples3 using a shotgun sequencing approach, which generates random sequences (reads) from a complex pool of DNA molecules. The reads are then assembled on the basis of overlapping sequence similarity, allowing researchers to identify genes and to predict the functions of the proteins that they encode.
The investigators mapped their reads to all available microbial and viral genome sequences to assess community composition — the different types of microbes and their relative abundance — at the various body sites. The researchers also determined the whole-genome sequences of about 800 bacterial strains isolated from humans (from a planned total of 3,000); these sequences have been placed in public databases and can be used as reference genomes for comparative purposes. The consortium authors conclude2,3 that they have identified the majority of the common microbial taxa and their genes present in these 242 healthy humans.
One of the great strengths of the HMP is that samples were collected simultaneously from multiple body habitats of the same individuals. This allowed Huttenhower et al.2 to discover that taxonomic and genetic diversity were greatest in tooth and stool samples, intermediate in skin samples and on the inside surface of the cheek, and lowest in vaginal samples2,3,7 (Fig. 1). The researchers report that each habitat is characterized by a small number of highly abundant 'signature' taxa, but that the relative representation of taxa and genes in each habitat varies considerably between individuals. In most samples, high-abundance taxa are accompanied by low-abundance taxa from the same genus, suggesting that within-community niche specialization occurs. These findings confirm those of an earlier study8, which demonstrated that body habitat accounts for much of the variation in bacterial community composition. Although there is clear evidence for individuality in people's microbiome compositions, the limited temporal scope of the HMP data set prevents a robust analysis of how these communities change over time.
As shown previously for faecal samples9, the relative abundance of microbial genes associated with certain physiological pathways varied less between samples from the same habitat than did the relative abundance of taxa. This suggests that there is functional redundancy between microbial community members. The prevalence of low-abundance genes varied the most between habitats, and Huttenhower and colleagues speculate that the functions of these genes correspond to body-niche-specific activities. Interestingly, the researchers also found evidence of taxonomic co-variation across sites in an individual, such as between the communities of the skin and the saliva2. One possible explanation for this is that interactions between community members are subject to selection pressure by host-specific, and host-wide, environmental factors.
Studying the human microbiome has so far been a lesson in humility. Although the HMP and the MetaHIT project10 are revealing vast amounts of previously uncharacterized microbial diversity within our 'home turf', the functions of these communities remain largely unknown. Moreover, Methé et al.3 report that only 57% of the non-redundant gene families identified by the HMP and MetaHIT researchers were detected by both groups, with 34% of the gene families being detected only in the HMP data and 10% only by the MetaHIT project. This raises questions about the representativeness of the people in the two projects. In both cases, samples were taken from adults in developed nations who have relatively similar lifestyles and, in the case of the HMP, without inflammatory disease. Another recent report reveals11 that populations living in less-developed regions of the world have markedly different microbiomes from those living in the United States.
It is also important to consider the definition of health. The most common cause for exclusion of people from the HMP was chronic gum disease, a condition that is increasingly regarded as 'normal' in developed countries. Furthermore, the prevalence of overweight and obese individuals continues to rise in many populations around the world, the chronic use of prescription drugs is becoming more common, urbanization is increasing, and our natural environment is changing in unexpected ways. Future studies of the human microbiome should accommodate such factors, which are likely to influence our microbial inhabitants.
Many areas of human-microbiome research warrant further investigation, but viruses and small non-bacterial organisms such as fungi deserve special attention, as do questions regarding the functions of the microbiome. We are essentially blind to many of the services that our microbial ecosystems provide — and on which our health depends12 — and investigators desperately need new approaches for studying interactions between members of the microbial community and their human hosts. As Huttenhower and colleagues suggest2, the fact that some of the microbiome's functions are likely to be performed by rare community members or to involve genes that are expressed at low levels will further complicate attempts to decipher their influence.
Despite the valuable initial findings from the HMP and other projects, multiple lines of enquiry remain. Which factors are responsible for the day-to-day or longer-term variation in the composition and functions of a person's microbiome? To what degree are such factors intrinsic to the microorganisms, related to the host, or, indeed, stochastic? Which mechanisms regulate bacterial colonization or invasion of the human microbiome, how does the microbiome respond to disturbance, and to what degree does this response involve the propagation of surviving organisms versus new colonization from outside? What is the basis of resilience in the human microbiome, and can it be predicted and restored? Such questions suggest that the work of Methé et al., Huttenhower et al. and the numerous others studying the human microbiome is only just beginning.
About this article