The human microbiome: eliminating the biomedical/environmental dichotomy in microbial ecology


When a new human being emerges from its mother, a new island pops up in microbial space. Although a human lifespan is a blink in evolutionary time, the human island chain has existed for several million years, and our ancestors stretch back over the millennia in a continuous archipelago. Microbes thrive on us: we provide wonderfully rich and varied habitats, from our UV-exposed, oxic and desiccating skin to our dark, wet, anoxic and energy-rich gut that serves as a home to the vast majority of our 100 trillion microbial (bacterial and archaeal) partners. A sobering or inspiring fact: we contain 10 times more microbial than human cells and an estimated 100 times more microbial genes. How our association with microbes has evolved, the forces that shape it, what about it might be uniquely ‘human’, how changes in our biosphere are affecting it, and how it impacts our health, all are challenging questions for the future because they require a level of engineering and computational sophistication that is still emerging. Our crystal ball sees the epidemiologist of the future describing how changes in kilometre-scale macro-ecosystems affect micrometer-scale microbial ecosystems associated with populations of meter-scale human beings, on time scales of an infection, a human lifespan, or the rise and fall of a society.

Step forward into the world of metagenomics, and we start to see ourselves as supra-organisms whose genome evolved with associated microbial genomes (the microbiome). Although the primate-lineage component of the human genome is decoded, sequencing of the microbiome is just beginning. Our first glance of the microbiota, from recent extensive surveys of its organismal lineages (based on 16S rRNA), and initial DNA-based metagenomic analyses of its microbiome, raises a long list of basic questions: is there a core microbiome shared between humans and passed along as a family heirloom? How does the microbiome of humans differ from those of other animals? How does it change with ageing, travel, marriage, sickness? Is its composition an unappreciated determinant of our well-being, and/or a contributing factor to diseases such as obesity?

We humans have an extraordinary impact on the environment: although we comprise about 0.5% of the total heterotroph biomass on earth, we consume 14–26% of terrestrial net primary production (70% in some regions of south-central Asia) (Imhoff et al., 2004). The remarkable diversity of food sources available to us compared with other species, and our global distribution might be reflected in an exceptionally diverse gut microbiota compared with other species if diet is the primary factor driving diversity in this microbiota and its microbiome. On the other hand, our low levels of (primate lineage) genetic diversity relative to other mammal species (Li and Sadler, 1991; Kaessmann et al., 2001) might suggest that our gut microbiota is relatively impoverished if host genetics (perhaps mediated through the immune system) is the primary determinant of microbial diversity.

To answer these questions about human microbial ecology and its variation requires an integration of data about a microbiota's collective genome, transcriptome and metabolome, the physical and chemical attributes of the host's surface habitats occupied by the sampled microbial community, plus information about the genotype, systems physiology, lifestyle, and living environment of the humans being studied. Initial studies that use monozygotic twin pairs and their mothers will help limit some confounding variables.

Comparisons of the microbiota and microbiome in different groups of humans undoubtedly will be multilayered, making use of data sets of lineage assemblages, gene assemblages and populations of genome types. Integrating these diverse and complex data sets will spark development of advanced computational methods. Our crystal ball shows that the key challenge will be in defining distances: e.g. how far apart are the microbiomes, the transcriptomes, and the metabolomes, of gut microbial communities represented in two stool samples from unrelated, related or the same individual(s)? How can we relate changes in these parameters to changes on different time scales, such as day-to-day diet, health and disease, and the evolutionary ecology of different species? Once these distances are defined, we can use established statistical techniques to place all the data in a single, unified framework that allows us to ask which differences correlate with one another and with human health. A key innovation will be to extend the phylogenetic distance metrics developed for comparing microbial communities using 16S rRNA genes to allow phylogenetic classification of metagenomic samples. The concept of distance is so fundamental a unifying principle that the mantra among students may become ‘distance is to microbial ecology as energy is to physics!’. New, accurate methods of assessing lateral gene transfer that combine information about both the composition and phylogenetic history of each gene, will revolutionize our picture of microbial adaptation to the diversity of ecological niches humans provide.

One major question we will soon be able to answer is whether there is a threshold beyond which phylogenetic relatedness is irrelevant to predicting microbial function: just as pigeons and penguins are both birds but occupy very different niches, we may find it is not correct to assume that relatedness is our best guide to bacterial functions in the human microbiome. Understanding whether rare or abundant microbes play the greatest role in determining the function of the human microbiota and its microbiome, in particular by relating abundance and gene expression in these species to human health, will revolutionize our understanding of the ≥99% of the genes associated with our body that are carried by microbial genomes.

Once these methods are in hand, the epidemiologist of the future can collect metadata to correlate with variability between his subject's microbiomes. The results of epidemiological studies will be translated into therapies. Our medical insurance cards will contain one chip for our primate genome, and one for our microbiome. As part of the annual physical exam, physicians will take a stool sample to update the microbiome profile. Just as today a rise in blood pressure from one visit to the next signals a risk of developing heart disease, tomorrow changes in the microbiome profile will herald a predisposition to diseases such as obesity. Therapeutic intervention will follow, likely a combination of individualized nutrition, deliberate ‘re-programming’ of the microbiota with addition/removal or stimulation of particular lineages or genetic complements within the microbiome, or use of microbial gene products themselves (or their revealed human gene product targets) as part of our 21st century pharmacopoeia.

As our human population increases in size, and as globalization promotes movement of people around the world, and exposure to one another and to new environments (including those changed by our anthropogenic perturbations), our microbiomes proliferate and mix between populations at rates unprecedented in human evolution. The impact of global change on our human microbial ecology (our ‘microevolution’) is not currently known but we are surely not detached from it. It behoves us to establish human microbial observatories just as we have established long-term environmental microbial observatories. It is time to breach the institutionalized dichotomy between environmental science and biomedical research, and to study ourselves as an integral and dependent part of our microbe-dominated world.