Are Our Microbiomes Racial?

Imagine…that all human bodies which exist looked alike, that on the other hand, different sets of characteristics seemed, as it were, to change their habitation among these bodies. Such a set of characteristics might be, say, mildness, together with a high pitched voice, and slow movements, or a choleric temperament, a deep voice, and jerky movements, and such like. Under such circumstances, although it would be possible to give the bodies names, we should perhaps be as little inclined to do so as we are to give names to the chairs of our dining-room set. On the other hand, it might be useful to give names to the sets of characteristics, and the use of these names would now roughly correspond to the personal names in our present language. –Ludwig Wittgenstein (1933-1934)

By an order of magnitude—10 to 1—most of the cells in our bodies aren’t even our own. We handle the indignity by assuming ourselves the ecological stage across which ‘our’ microorganismal visitors must mindlessly interact. Like we were gods looking down upon subjects so puny they didn’t know we existed.

As if our consciousness was synonymous with control. As if the quorum effects routinely documented in microbes couldn’t possibly include a distributed if insentient cognition, or, perhaps more disturbingly, in an ironic reversal, a sentience so unearthly we wouldn’t recognize it if it were staring us in the face.

Maybe the animists have it right after all, material mechanisms aside. Whether we sense it or not perhaps we are routinely and multiply possessed. And when we are sickened by infection our ill at ease arises in part out of a sinking feeling we are momentarily inhabited by a diffuse and alien being.

The nastier visitors live a lifetime folded inside our own: birth upon infection, kids by transmission, and death on immunal clearance or—bummer—our own passing (although some outlive their hosts, making it to the next susceptible by virtue of killing us).

Within hosts and across populations disparate pathogens can converge on locale-specific disease guilds, a possession by committee, with each member cutting ground for the other. HIV, KSHV and TB, for instance, arrived at a passing sub-Saharan detente, producing an epidemiological mutualism scientists have found embodied by specific molecular mechanisms. In reciprocal activation, two bugs’ proteins set off molecular pathways in the other.

New work by Peer Bork’s group shows a trinity of such communities among humanity’s natural gut fauna, which typically aid our digestion and produce vitamins. With twenty-two new fecal metagenomes added to those already in the literature, sequencing every microbial gene found in the gut, the team identified by cluster and principle components analyses three basic enterotypes across our species’ microbiomes.

Type 1 is largely dominated by Bacteroides bacteria, type 2 by Prevotella, and type 3 by Ruminococcus. Each major actor is accompanied by a unique entourage at the exclusion of other taxa. The Bacteroides of type 1, for instance, is positively associated with Clostridiales and Parabacteroides, and negatively associated with six others.

The three enterotypes, however, aren’t stratified by country or continent, nor by host sex, health, age, or, as previously shown, weight. They do differ in the mix of enzymes they produce. Enterotypes 1 and 2, for instance, generate energy by way of enzyme clusters involved in fermenting carbohydrates and proteins and degrading mucin glycoproteins, respectively. Fascinating differences in their divisions of labor. Some of these individual protein clusters evidently differ by a number of demographics. Enzymes involved in starch degradation, such as glycosidases and glucan phosphorylases, increase with age.

As the study’s subjects were entirely drawn from industrial countries, the work is incomplete. Other enterotypes may yet segregate by food regimen or locale, for instance.

Recent work by Eric Alm’s group shows microbiomes generally group more by ecological niche than host genotype and locality, and do so even across host species! With dumbfounding implications, the team showed livestock- and human-associated bacteria recently shared forty-two antibiotic-resistance genes. In some profoundly fundamental ways, industrial pigs and people are becoming more closely related to each other than either of us is to our indigenous cousins.


But assuming functionality as an all-encompassing cause—the Bork group alludes to the possibility of some unidentified host or environmental differences—may be a matter of putting the cart before the horse. While enterotyping’s applications are apparent—in everything from diet to drug intake to disease diagnostics—non-adaptationist explanations for such ecological assemblages need to be put into the mix:

The Bork work doesn’t clarify whether the enterotypes are inherited. Setting aside for now the problem of horizontal gene transfer, even if they were, at the population level the propagation of one enterotype over another may be from our perspective near-random. The enterotypes may crisscross human families in lineages outside the scope of their hosts’ genetic inheritance. Familial correlation may be more a matter of environmental proximity than host genetics.

The assembling itself, on the other hand, is most certainly functionally locked in, a taxon here and there notwithstanding given the inevitable differences in individual host guts and the occasional stochastic burps across enteric environments. Once one key bacterium taxon colonizes a newborn gut by chance, even if those odds are weighted by some familial factor, including exposure at birth, the pioneer prepares the gut for its enterotypic successors.

What we’re getting at here is that we need to offer the concepts of contingency and historical constraint the kind of room in our explanation natural selection is oft-afforded. And if enterotypes move from person to person with little or no reference to their host’s identity, turning people, at least in this domain, into Wittgenstein’s nameless, or into races defined by the content of their gut rather than the color of their skin, then causality, and our sense of self, must be situated as much in the field as it is in the object.

Indeed, as we are co-dependent on these microbes for survival, perhaps we are as much their visitors as they ours. Each newborn must meet and greet its group of commensurate bacteria. And these unicellular confederates in turn emerge out of a historical trajectory microbe and human mold off-and-on together.


Cutting edge work, however, is by definition in flux. Conclusions are routinely overturned (only to be overturned again). Work published just this past month by the Human Microbiome Project documenting the microbiome over a larger population and across multiple body parts, including the gut, indicates that, well, in fact, at least in this new study, the human microbiome is racially segregated (here, here and here).

HMP screened the microbial genomes of nearly 5000 specimens from 242 healthy Americans. Samples were drawn from across 18 body habitats: oral cavity and oropharynx, saliva, inside cheek, gums, palate, tonsils, throat, tongue, tooth biofilm above and below the gum line, behind the ears, inner elbows, a nostril, a stool sample, and for the lucky ladies, three vaginal specimens. For 131 individuals a second sample was taken from each part at a subsequent time point to test the stability of microbial community structure.

The consortium isolated and sequenced sections of all 16S ribosomal RNA (rRNA) it detected in each sample. Species across the tree of life can be differentiated by their 16S, which consist of highly conserved regions alternating with variable sequences.

By a second protocol the genomes of the entire microbial community were profiled by whole-genome shotgun sequencing, within specified limits in detection, sequencing depth and statistical power. Community membership was identified by comparing the resulting reads against reference sequences available for bacteria, archaea, viruses and microeukaryotes. Finally, the genes identified were assigned as best as possible to protein families and annotated for gene function.

Ostensibly the aim of the project to this point is to produce reference microbial profiles of ‘healthy’ individuals against which to compare those subsequent of sick people. Do specific illnesses correlate with shifts in microbial taxonomic and functional profiles in any particular body part or across parts as smaller studies so far suggest? A fascinating and fundamental question indeed.

But what HMP did find in these reference individuals is in and of itself illuminating:

  • Oral and stool microbial communities were particularly diverse in community taxonomy. Body parts, however, differed in their alpha and beta diversities. Individual saliva samples, for one, displayed great diversity in their taxonomy but differed little among individual subjects.
  • Variation within individuals over time differed less than between-subjects in both taxonomy and function. That is, albeit over only two time points, individual microbial communities appear to remain relatively stable. More sampling across longer periods is likely to capture profile shifts, even in individual hosts who remain in good health. Across individuals, the study showed depending on the body part microbial configurations either locked into discrete blocks or varying continuously.
  • No microbial taxa were observed present in all body parts or in all individual subjects. In other words, HMP found strong niche differentiation within and among subjects. Some metabolic pathways, however, were universally present, indicating multiple taxa functionally converging on core housekeeping tasks, including, in the gut, spermidine biosynthesis, methionine degradation and hydrogen sulphide production.
  • Although no ubiquitous taxa were apparent, each body part hosted primary signature species. Lots of Streptococcus across body parts, for instance. Haemophilus in the inside cheek. Actinomyces in the tooth biofilm above the gum. Even so, within this methodological context microbiome profiles proved deeply personalized, particularly in the make-up of less abundant species. In other words, microbial succession isn’t wholly deterministic. Within specific taxa, across individuals, there also appeared considerable genetic variation, perhaps, as the consortium surmises, a result of host selection pressures and the functional trade-offs across taxa, but perhaps also, as we discussed in our review of the Bork work, from contingent founder events.
  • Cholera, TB and salmonella and other NIAID class A-C pathogens weren’t detected in the study population. On the other hand, genes from the PATRIC database were found in relative abundance across hosts and habitats. The result suggests the health/disease divide may depend on the state of the microbial community rather than the mere presence of an etiological agent, including shifts in the relative abundance and functional interactions among the community’s constituents. HMP hypothesizes that if suppressor taxa shift, then a “recessive” disease agent may become microecologically active.


And race? Do our microbial communities differ by particular host characteristics? Do age, body mass index, ethnicity, and other clinical metadata correlate with microbial taxonomy and metabolism?

As with Bork’s group, the consortium found most of the relationships, for instance with body mass index, were weak. HMP hypothesizes other factors as yet unstudied may be important, including short- and long-term diet, daily metabolic cycles, and, as we discussed, founder effects such as mode of transmission and individual host genetics.

A small number did stand out, however. As previous research showed, vaginal pH appeared correlated with microbial shifts, including at higher pH a drop in Lactobacillus and an increase in overall diversity. Increasing age appeared associated with greater diversity in metabolic pathways on the skin.

Finally, the big reveal, unlike the Bork work subject ethnicity strongly dropped out across body habitats. Ethnicity showed 266 associations with various microbiome data, more than any other demographic/clinical factor. Asians, Mexicans and Whites, for instance, showed large relative abundances in ornithine and histidine biosynthesis in the tongue and small abundances in Proteobacteria and Gammaproteobacteria clades in the nares and inner elbow. Blacks and Puerto Ricans showed the reverse.

Perhaps with good sense, HMP offered nary a word about why. But it’s clear, depending on the resulting applications, we find ourselves on the eve of the next battle over the meaning of racial differences in biomedical outcomes.

Ten years ago the essentialist typology underlying race-based drugs such as BiDil set off a firestorm. As Troy Duster described it early on,

In what has been touted as “the first ethnic drug,” the biotech firm NitroMed received a green light from the Food and Drug Administration in March 2001 to proceed with a full-scale clinical trial, “the first prospective trial conducted exclusively in black men and women suffering from heart failure.”…

BiDil is a drug designed to restore low or depleted nitric oxide levels to the blood to treat or prevent cases of congestive heart failure. It was originally designed for a wide population base, and race was irrelevant. But the early clinical studies revealed no compelling results, and an FDA advisory panel voted 9 to 3 against approval.

In a remarkable turn of fate, however, BiDil was suddenly born again as a racialized intervention…

[NitroMed claimed] that BiDil has a special effect greater on African Americans than whites. The clinical trials now under way are not designed to test that hypothesis. Rather, by concentrating only on blacks, the study can have little or nothing compelling to say about comparative results, by race.

Developmental biologist Armand Leroi retroactively certified such expediency. As race can affect medical treatment, “many new drugs are now labeled with warnings that they may not work in some ethnic or racial groups.” That such effects need not be predominantly biological in origin—or even exist—mattered little.

Leroi cagily conceded differences among races arise from population averages alone. But as we are unlikely to sequence individuals’ genomes at scale any time soon, he argued, and presumably won’t be able to individualize medical treatment that way, we’ll just have to accept a racialized medical genetics.

Even many among those who have accepted racial genomics as biologically real have since rejected such crass calculation,

In the case of BiDil, [Craig Venter and his colleagues] note, the manufacturer “is voluntarily financing a study to investigate the genetic basis for the response to the drug.” But in general, “Once a race-based drug has been developed, there is a possibility that a drug company may terminate its research and not pursue follow-up studies into the underlying cause. This could stunt medical care with race-based medicine, rather than personalized medicine.”

Venter’s position here puts him at odds with his own creed that, “Science is the business right now. If the science works, the business works, and vice versa.” Accordingly, NitroMed’s quarterly margins would sanctify the science behind the drug, even as the science here was found more in the marketing than the biology. As Jonathan Khan discovered,

[No] firm evidence exists that BiDil actually works better or differently in African-Americans than in anyone else… So how did BiDil become tagged as an ethnic drug and the harbinger of a new age of medicine? The story of the drug’s development is a tangled tale of inconclusive studies, regulatory hurdles and commercial motives.

The drug’s effects were more financial than biological. “Why the mistake, and what is at stake?” Duster asked,

Part of the answer lies in the role of prospective markets for biotech products. While the new mantra of biotechnology is to claim that pharmaceuticals will someday soon be marketed to individuals based upon their DNA, the fundamental truth is that selling drugs is about markets. These markets are not about individual designer drugs, but about groups and population aggregates that become the target market.

In a classical piece of epidemiological research, Michael Klag and his associates showed a decade ago that, in general, the darker the skin color, the higher the rate of hypertension for American blacks, even inside the African American community. Klag indicated that the issue was not biological or genetic in origin, but biological in effect due to stress-related outcomes of reduced access to valued social goods such as employment, promotion, housing stock, etc. The effect was biological, not the origins.

NitroMed repackaged a failed drug as racially based. And Leroi used NitroMed’s marketing to confirm race’s biological reality.

The problem extends beyond racial drugs to medical genomics itself, which has largely fallen flat delivering on its prospectuses. However amenable they are to commoditization, our genomes aren’t our health destinies. But there remains considerable weight in such an ideological imperative. Given science’s present economics, increasingly capital-financed, method and interpretation alike appear now turning to the new metagenomics for recourse.

“Research into the nature of the human microbiome has yielded many surprises,” writes Jennifer Ackerman,

[No] two people share the same microbial makeup, for instance—even identical twins. This finding may help unravel a mystery presented by the Human Genome Project, which confirmed that the human DNA of all the people the over is 99.9 alike. Our individual fates, health and perhaps even some of our actions may have more to do with the variation in the genes found in our microbiome than in our own genes.

Capitalism’s ingenuity can be found in repeatedly cornering epigenetic biology into preformationist precepts. With dividends in systemic legitimization. Nature is made capitalist in nature, making capitalism natural. Health disparities arise out of our genes or our guts rather than systems of apartheid.


New epistemological boss same as the old epistemological boss. Such essentialisms deserve continued rebuke, in whatever forms they transmogrify. There may be a more positive, if not also positivist, response, however.

We do risk letting metagenomics change the topic. By way of their emotional and material deprivations, racism and other sources of population-level stress have definitional effect on individuals as early as conception, as well as on the populations of which they are a part. Racism shapes ontogeny, regardless of allelic frequencies. And it’s to that relationship and its overdetermination of health outcomes by race where research and social action need to be directed.

But it may be worth the risk reappropriating the terms of the debate. By a true social science of the microbiome we can investigate host determinants under a more sophisticated rubric than clunky clinical metadata.

For instance, how is microbiome diversity geographically distributed? Are microbiomes racially segregated more in racially segregated neighborhoods, cities or countries than in integrated ones? Or from pixel to pixel or block to block? Does racial oppression, and its material and sociopsychological stresses and deprivations, select for particular microbial profiles over others?

Do microbiomes display a class structure? Is there such a thing as a Great Recession microbiome? Do the discrete blocks vs. continuous diversity the HMP study found map to different parts of this social multispace? Do profiles from body part to body part depend on the kinds of environmental insults found locally? And if so, across which aspects of the social and physical environment? What is the shape and nature of the population social network across which commensal microbes are transmitted?

All such questions would require rethinking people less as clinical subjects from which to tap samples—a Gray’s Physiology—and more as socially active beings, part and parcel of particular populations at particular localities, shaped by an idiosyncratic history, and differentially exposed in time and space to the greater world.

As a result, each of our microbiomes may indeed deserve a personal name, representing more than a single person but in addition the histories each of us shares with whom and with where we live. I’ve taken to greeting mine Ludwig St. Paul. What it may call me in return is beyond our present understanding.


3 Responses to “Are Our Microbiomes Racial?”

  1. To: Host of Ludwig St. Paul
    From: Hostess with the Mostess

    Re: Race and the Human Microbiome

    Wow. This is mind-shifting stuff. I appreciate your writerly approach. Thanks for posting this link on Spirit of 1848 list-serve, and please continue to do so!

  2. rgwallace Says:

    Thanks, Christine! And I like your MB handle. LSP, 10-4 out.

  3. Thank you for this piece. It is very informative yet not so scientific in presentation that the average reader gets lost. Most presentations I have heard about the microbiome have been from doctors and nutritionists. It was refreshing to hear the discussion in this context.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: