From Microbiota to Microbiome: Understanding Our Greater Selves 

A party fact I like to cite is that humans have, on average, 1-2 kg of bacteria in them, roughly the mass of the human brain ¹. That’s three pints and tens of thousands of unique species dwelling on our skins, forming complex ecosystems in our gastrointestinal tracts, and modulating the environments of our oral cavities, among other things². In most contexts (even in Boston, where I live), this is met with some mix of disgust and incredulity, despite the public’s general acknowledgement that humans are not alone in (or on) our skins. However, it’s one thing to know that we’re not alone and another thing entirely to understand exactly what that means for us. Is it good? Is it bad? Do we need these microbes? Do they need us? Well, in a word, yes. 

Now, I might not be welcome at many parties, but these microbes certainly are. With roles spanning immune regulation, gut-brain axis modulation, maternal health, and a whole host of other activities, these microbiota are more than simple passengers. Indeed, these communities are often termed the “hidden organ” because, while they don’t occupy the obvious physical space of something like the brain, kidney, or pancreas, they are an essential part of our physiology. That’s why we’ve transitioned from not only exploring the “microbiota” – the organisms inhabiting a space – to also examining the “microbiome” – those organisms along with their collections of genes, metabolites, and functions that together form a complex superstructure of multiple species. 

The growing quantity of research bears this out: in the last 20 years, articles with the topic “microbiome” have multiplied, with a 48% yearly growth rate in publications with that keyword to over 3400 last year alone and a concomitant increase in placements in top-tier journals (Fig. 1, own analysis). The public’s perception has similarly magnified: searches for “microbiome” have expanded from essentially zero traffic to 25% of the search volume of probiotics, which is usually the first association the public has with the microbiome. It’s clear that we’re no longer conceiving of our microbiome as a passive entrant into our digestive tracts, but as an active and diverse participant in virtually every element of human health. 

My colleague Alessandro Busetti has written a comprehensive review focusing on our current understanding of the microbiome in relation to human health. He makes three major points. First, an organism does not exist in a vacuum, instead, it is a complex ecosystem – a “meta-organism”³ – where many different, unique species make up one larger, functioning whole. Second, this complexity presents a highly dimensional problem, and measuring the relationships of these organisms is hardly a trivial task. Third, he enumerates some of the major ways we’re leveraging our knowledge about the microbiome, including the skin microbiome, nutrition, drug development, and immune modulation. The throughline here is that, well, more than 10,000 species grappling for purchase on a planet we thought was fundamentally our own is an intrinsically difficult thing to understand and has ramifications far beyond our initial conceptions. 

Surprise party 

We’ve known we’re not alone since Antonie van Leeuwenhoek’s microscopic investigations of his own microbiota in the 1670s-1680s ⁴. More recently, Barry Marshall won the Nobel Prize in medicine for characterizing Helicobacter pylori’s role in stomach ulcers and dysbiosis in 2005, identifying a specific (if incomplete) link between a microbial species and human health ⁵. On the flipside, we’re left wondering what life looks like without all these possible invaders. To that end, in the late 1800s, the first germ-free rodent was delivered via caesarian section, and – consistent with Louis Pasteur’s 1885 hypothesis – the animals didn’t do so well⁶. Ernst Kuster’s 1915 germ-free goat lasted a mere 34 days, and a litany of other health defects seemed to follow the embryos born in isolation⁷. It’s not only common to mammals: synthetic crops tend to be less nutrient-efficient and resilient than their conventional cousins⁸. It’s clear that our microbiomes are both essential and integrated, but also difficult to line up with a history of pathogenesis and infection from unwanted hangers-on. 

So, how do we square this massive, alien circle? In a previous post, I discussed how multiomic methods are accelerating innovation through complex physiological conditions. While the Human Genome Project may not have solved our own blueprints, it has led to a booming industry of lower-cost sequencing, and our understanding of our own metaorganism might be the main beneficiary. We now know that our microbiomes contribute roughly 150 times the genetic information that our native genomes do, inferring a massive extension on our own metabolic and phenotypic potentials⁹. We’ve made good use of this extension, too, characterizing 26,970 species bins in the popular and state-of-the-art MetaPhlAn 4 database¹⁰. That’s a far cry from Leeuwenhoek’s five “animalcules” and implies a lot more function than Marshall’s H. pylori. Likewise, the development and standardization of germ-free animals gives us ideal, cough, guinea pigs to test our hypotheses on. But 26,970 is a huge number to rigorously test in traditional scientific null-hypothesis methods, not to mention the combinatorial expanse that follows, so we need a better way. 

Where the wild things are 

While we can’t exhaustively test and enumerate our microbiomes, we can try to measure our inhabitants and link them to real function. These microbes stand, along with our own cells, on the interface between us and our environment, impacting the ways we eat, breathe, smell, and respond to a whole world of stimuli. In many ways, they’re our first and best friends, introduced to us first at birth through our mother’s vaginal microbiome. This last point poses another functional difficulty for using germ-free models, which are necessarily germ-free from birth, in turn limiting later-life germ-free efforts to antibiotic-aided studies¹¹. Instead, we focus on collecting reference states and then identifying phenotypic departures, collecting data as inclusively as possible, and attempting to parse out the results afterward. 

While sequencing and advances in culturing techniques afford us the latitude to take snapshots of conditions, it’s another thing entirely to quilt those snapshots into a causal and cohesive story. Individual microbiomes differ significantly by age, nutrition, medication, and a myriad of other predictors. While some microbes, like Clostridium difficile, are found in antibiotic-administered patients¹², show obvious and predictable changes based on patient state, many others do not. Ideally, we might pull biological specimens from organisms and culture the microbes, but we’ve found quickly that organisms tend to function quite differently outside of the body than in it, with some important microbes recalcitrant to grow at all¹³. 

Instead of disturbing the careful ecosystem in the service of measurement, we can now rely on genetic sequencing to provide snapshots of microbial diversity and membership. Whether using preserved regions through 16S sequencing or relying on deeper shotgun sequencing, comparing the read amounts to established databases can give us an idea of the composition of an individual’s microbiome. We can now say “what” is there by quantifying the microbiota and, based on alignment with other studies, what that “what” might be doing, inferring the microbiome. That is to say that we can start to parse out patterns and relationships based on phylogeny, previous studies, and metabolic potential to hypothesize the relationships between host, microbiome, and environment. This can be further extended by looking at individual genes in deeper sequencing studies, yielding a more complete suggestion on the activities that might govern a certain outcome. 

Easier said than done 

At this point, we’ve identified that we have this large, multispecies organ that we’re better off with than without. We can sequence biological specimens with 16S or shotgun sequences to generate FASTQ files and use specialized bioinformatics pipelines and, preferably, cloud or high-performance computing to form these into biologically relevant data. Getting to this point is nontrivial, usually requiring the efforts of at least one dedicated bioinformatician to build and maintain the pipeline. Even more, the results are not a straightforward indication of the biological underpinnings of the interaction. At its most standard, we receive an abundance table with rows as organismal lineages and columns as samples with values corresponding to the relative (percent) abundance of that lineage, or taxon, in that sample. Typical results include a cascade of lineages that recursively include major taxonomic levels from kingdom to species. At this point, we have a direct idea of what and how much is there. 

Unfortunately, the line between biological function is still obscured by the sheer quantity and variety of organisms present in a sample. That is to say that we continue, even in reference to other studies, to be confined to quantifying microbiota but not microbiome function. It’s not straightforward to determine, for example, whether our listed Escherichia coli is of the K-12¹⁴, Nissle 1917¹⁵, or EDL933¹⁶ variety simply by read binning at shallow depths. That’s a big problem; in this case, the first is a docile bacterium found broadly in labs as a model organism, the second is a probiotic strain originally isolated from a German soldier in 1917, and the third is pathogenic and enterohemorrhagic, and what people typically associate E. coli with. Naming the microbiota, organizing them phylogenetically, and assigning abundance values is a useful but incomplete analysis of their function. 

We know enough to delineate E. coli strains because they’re very domesticable. They leave an obvious mark on human physiology, they culture outside of a host, and there is a growing body of research and genetic tools that can be applied to them. We’re not so lucky with most species that live within us,  instead, we’re stuck in the hypothesis phase. We can use diversity metrics, cladograms, differential statistics, and enrichment against databases to start to organize our thoughts, but even that relies on strong experimental design, follow-up studies, or previous knowledge. For recalcitrant, uncharacterized strains at low abundance, we need other strategies. 

At the interface 

Having collected data on both the host and the microbiota inside it, the missing piece to understanding the microbiome as a collective is the collaboration and communication between the host, microbiota, and their environment. At this interface is a complex molecular interplay, most of which is conducted through small nutrient and signaling molecules¹⁷. Metabolomics is the ideal candidate for measuring this relationship and quantifying the specific interactions that occur within this complex ecosystem. 

Annie Evans, Head of Research and Development at Metabolon, says it well in a recent blog post: metabolomics allows us to “look around at ‘the now’”. We’re able not only to investigate the genetic potential of both systems operating separately, but we can sample molecular phenotypes from them at the moment of physiological outcome. Metabolomics also helps us identify the inputs from the environment and exposure that might not be identifiable from genomics alone. This is further established by the fact that microbiota do a lot of chemical reactions that humans simply cannot and by partnering metabolomics with metagenomics and human genomics, we can get a good idea of what organisms are producing what non-human chemicals. Metabolomics, then, is the ideal method for getting us from simple characterization to microbiome-focused, hypothesis-driven discovery. 

Metabolon is the world’s leading provider of quality metabolomics research at scale. At its core, Metabolon seeks to quantify metabolites according to the world’s largest proprietary library of annotated analytes, which is particularly necessary when many of the molecules at this complex interface might be confused with unqualified fragments by other providers. Now, by incorporating metagenomic data into its well-established metabolomic workflows, Metabolon is poised to push the field even further, connecting the dots between the diverse landscape of the microbiome and the individual it inhabits.

Metabolon isn’t just a data provider, having had its data and analysis incorporated into a long history of microbiome studies. As a field leader in metabolomic and microbiome analysis, Metabolon can now boast a suite of gold-standard metagenomic tools to evaluate metagenomic data without data engineering, management, or bioinformatic pipeline building. Not only that, but this suite of tools seamlessly integrates metadata, metabolomics, and metagenomics for the best-in-class end-to-end microbiome research product. We’re excited to see what you can discover and we’re excited to be a part of it.  

References

1. Flint, H. J. The impact of nutrition on the human microbiome. Nutr Rev 70, (2012). 
2. Ogunrinola, G. A., Oyewale, J. O., Oshamika, O. O. & Olasehinde, G. I. The Human Microbiome and Its Impacts on Health. Int J Microbiol 2020, (2020). 
3. Berg, G. et al. Microbiome definition re-visited: old concepts and new challenges. Microbiome vol. 8 Preprint at https://doi.org/10.1186/s40168-020-00875-0 (2020). 
4. Nonia Pariente. A field is born. Nature Research (2019). 
5. NobelPrize.org. Barry J Marshall — Facts. Nobel Prize Outreach 2025 (2025). 
6. Qv, L. et al. Methods for Establishment and Maintenance of Germ-Free Rat Models. Front Microbiol 11, (2020). 
7. Maggie Zhou. Germ-Free Animals. Embryo Project Encyclopedia (2024). 
8. Ansabayeva, A. et al. Plant Growth-Promoting Microbes for Resilient Farming Systems: Mitigating Environmental Stressors and Boosting Crops Productivity—A Review. Horticulturae vol. 11 Preprint at https://doi.org/10.3390/horticulturae11030260 (2025). 
9. Qin, J. et al. A human gut microbial gene catalogue established by metagenomic sequencing. Nature 464, 59–65 (2010). 
10. Blanco-Míguez, A. et al. Extending and improving metagenomic taxonomic profiling with uncharacterized species using MetaPhlAn 4. Nat Biotechnol 41, 1633–1644 (2023). 
11. Bongers, K. S. et al. Antibiotics cause metabolic changes in mice primarily through microbiome modulation rather than behavioral changes. PLoS One 17, (2022). 
12. Piccioni, A. et al. Gut Microbiota and Clostridium difficile: What We Know and the New Frontiers. International Journal of Molecular Sciences vol. 23 Preprint at https://doi.org/10.3390/ijms232113323 (2022). 
13. Rekdal, V. M., Bess, E. N., Bisanz, J. E., Turnbaugh, P. J. & Balskus, E. P. Discovery and inhibition of an interspecies gut bacterial pathway for Levodopa metabolism. Science (1979) 364, (2019). 
14. Browning, D. F., Hobman, J. L. & Busby, S. J. W. Laboratory strains of Escherichia coli K-12: things are seldom what they seem. Microb Genom 9, (2023). 
15. Sonnenborn, U. Escherichia coli strain Nissle 1917—from bench to bedside and back: history of a special Escherichia coli strain with probiotic properties. FEMS Microbiol Lett 363, fnw212 (2016). 
16. Saile, N. et al. Escherichia coli O157:H7 strain EDL933 harbors multiple functional prophage-associated genes necessary for the utilization of 5-N-acetyl-9-O-acetyl neuraminic acid as a growth substrate. Appl Environ Microbiol 82, 5940–5950 (2016). 
17. Silpe, J. E. & Balskus, E. P. Deciphering Human Microbiota-Host Chemical Interactions. ACS Cent Sci 7, 20–29 (2021)