Statistically learning the functional landscape of microbial communities

Soft Modes as a Predictive Framework for Low-Dimensional Biological Systems Across Scales

All biological systems are subject to perturbations arising from thermal fluctuations, external environments, or mutations. Yet, while biological systems consist of thousands of interacting components, recent high-throughput experiments have shown that their response to perturbations is surprisingly low dimensional: confined to only a few stereotyped changes out of the many possible. In this review, we explore a unifying dynamical systems framework-soft modes-to explain and analyze low dimensionality in biology, from molecules to ecosystems. We argue that this soft mode framework makes nontrivial predictions that generalize classic ideas from developmental biology to disparate systems, namely phenocopying, dual buffering, and global epistasis. While some of these predictions have been borne out in experiments, we discuss how soft modes allow for a surprisingly far-reaching and unifying framework in which to analyze data from protein biophysics to microbial ecology.

Microbiome determinants of productivity in aquaculture of whiteleg shrimp

Aquaculture holds immense promise for addressing the food needs of our growing global population. Yet, a quantitative understanding of the factors that control its efficiency and productivity has remained elusive. In this study, we address this knowledge gap by focusing on the microbiome determinants of productivity, more specifically animal survival and growth, for one of the most predominant animal species in global aquaculture, whiteleg shrimp (Penaeus vannamei). Through analysis of the shrimp-associated microbiome from previous studies across Asia and Latin America, we established the presence of core phylogenetic groups, widely prevalent across aquaculture conditions in disparate geographic locations and including both pathogenic and beneficial microbes. Focusing on the early stages of growth (larval hatcheries), we showed that the composition of the microbiome alone can predict a remarkable fraction of the variation in shrimp larvae survival rates (ca. 50%). Taxa associated with high survival rates share recently acquired genes that appear to be specific to aquaculture conditions. These genes are involved in the biosynthesis of growth factors and protein degradation, underscoring the potential role of beneficial microorganisms in nutrient assimilation. By contrast, the predictability of the microbiome on the adult shrimp weight in grow-out farms is weaker (10%-20%), akin to observations in the context of livestock. In conclusion, our study unveils a novel avenue for predicting productivity in shrimp aquaculture based on microbiome analysis. This paves the way for targeted manipulation of the microbiome as a strategic approach to enhance aquaculture efficiency from the earliest developmental stages.

IMPORTANCE

Aquaculture is a rapidly growing industry essential for global food security, yet its productivity is often constrained by high mortality rates and inefficient growth. While the microbiome is known to influence host health and nutrient assimilation, its broader role in animal production remains poorly understood. Here, we take a data-driven approach to address this gap by systematically analyzing shrimp-associated microbiomes across hatcheries and farms. By integrating microbiome data with machine learning, we demonstrate that microbial communities are powerful predictors of key production outcomes, shaping shrimp survival and growth. Our findings suggest that the microbiome could serve as a diagnostic tool for assessing production conditions and optimizing management strategies. In addition, machine learning techniques offer a promising avenue for identifying beneficial microbes and developing targeted microbiome therapies to enhance aquaculture sustainability and efficiency.

Low-dimensional genotype-fitness mapping across divergent environments suggests a limiting functions model of fitness

A central goal in evolutionary biology is to be able to predict the effect of a genetic mutation on fitness. This is a major challenge because fitness depends both on phenotypic changes due to the mutation, and how these phenotypes map onto fitness in a particular environment. Genotype, phenotype, and environment spaces are all extremely complex, rendering bottom-up prediction unlikely. Here we show, using a large collection of adaptive yeast mutants, that fitness across a set of lab environments can be well-captured by top-down, low-dimensional linear models that generate abstract genotype-phenotype-fitness maps. We find that these maps are low-dimensional not only in the environment where the adaptive mutants evolved, but also in more divergent environments. We further find that the genotype-phenotype-fitness spaces implied by these maps overlap only partially across environments. We argue that these patterns are consistent with a "limiting functions" model of fitness, whereby only a small number of limiting functions can be modified to affect fitness in any given environment. The pleiotropic side-effects on non-limiting functions are effectively hidden from natural selection locally, but can be revealed globally. These results combine to emphasize the importance of environmental context in genotype-phenotype-fitness mapping, and have implications for the predictability and trajectory of evolution in complex environments.

Replicating community dynamics reveals how initial composition shapes the functional outcomes of bacterial communities

Bacterial communities play key roles in global biogeochemical cycles, industry, agriculture, human health, and animal husbandry. There is therefore great interest in understanding bacterial community dynamics so that they can be controlled and engineered to optimise ecosystem services. We assess the reproducibility and predictability of bacterial community dynamics by creating a frozen archive of hundreds of naturally-occurring bacterial communities that we repeatedly revive and track in a standardised, complex resource environment. Replicate communities follow reproducible trajectories and the community dynamics closely map to ecosystem functioning. However, even under standardised conditions, the communities exhibit tipping-points, where small differences in initial community composition create divergent compositional and functional outcomes. The predictability of community trajectories therefore requires detailed knowledge of rugged compositional landscapes where ecosystem properties are not the inevitable result of prevailing environmental conditions but can be tilted toward different outcomes depending on the initial community composition. Our results shed light on the relationship between composition and function, opening new avenues to understand the feasibility and limitations of function prediction in complex microbial communities.

Collective Microbial Effects Drive Toxin Bioremediation and Enable Rational Design

The metabolic activity of microbial communities is essential for host and environmental health, influencing processes from immune regulation to bioremediation. Given this importance, the rational design of microbiomes with targeted functional properties is an important objective. Designing microbial consortia with targeted functions is challenging due to complex community interactions and environmental heterogeneity. Community-function landscapes address this challenge by statistically inferring impacts of species presence or absence on function. Similar to fitness landscapes, community-function landscapes are shaped by both additive effects and interactions (epistasis) among species that influence function. Here, we apply the community-function landscape approach to design synthetic microbial consortia to degrade the toxic environmental contaminant bisphenol-A (BPA). Using synthetic communities of BPA-degrading isolates, we map community-function landscapes across increasing BPA concentrations, where higher BPA means greater toxicity. As toxicity increases, so does epistasis, indicating that collective effects become more important in degradation. Further, we leverage landscapes to rationally design communities with predictable BPA degradation dynamics in vitro. Remarkably, designed synthetic communities are able to remediate BPA in contaminated soils. Our results demonstrate that toxicity can drive epistatic interactions in community-function landscapes and that these landscapes can guide microbial consortia design for bioremediation.

Competition for shared resources increases dependence on initial population size during coalescence of gut microbial communities

The long-term success of introduced populations depends on both their initial size and ability to compete against existing residents, but it remains unclear how these factors collectively shape colonization dynamics. Here, we investigate how initial population (propagule) size shapes the outcome of community coalescence by systematically mixing eight pairs of in vitro microbial communities at ratios that vary over six orders of magnitude, and we compare our results to neutral ecological theory. Although the composition of the resulting cocultures deviated substantially from neutral expectations, each coculture contained species whose relative abundance depended on propagule size even after ~40 generations of growth. Using a consumer-resource model, we show that this dose-dependent colonization can arise when resident and introduced species have high niche overlap and consume shared resources at similar rates. Strain isolates displayed longer-lasting dose dependence when introduced into diverse communities than in pairwise cocultures, consistent with our model's prediction that propagule size should have larger, more persistent effects in diverse communities. Our model also successfully predicted that species with similar resource-utilization profiles, as inferred from growth in spent media and untargeted metabolomics, would show stronger dose dependence in pairwise coculture. This work demonstrates that transient, dose-dependent colonization dynamics can emerge from resource competition and exert long-term effects on the outcomes of community coalescence.

Molecules-mediated bidirectional interactions between microbes and human cells

Complex molecules-mediated interactions, which are based on the bidirectional information exchange between microbes and human cells, enable the defense against diseases and health maintenance. Recently, diverse single-direction interactions based on active metabolites, immunity factors, and quorum sensing signals have largely been summarized separately. In this review, according to a simplified timeline, we proposed the framework of Molecules-mediated Bidirectional Interactions (MBI) between microbe and humans to decipher and understand their intricate interactions systematically. About the microbe-derived interactions, we summarized various molecules, such as short-chain fatty acids, bile acids, tryptophan catabolites, and quorum sensing molecules, and their corresponding human receptors. Concerning the human-derived interactions, we reviewed the effect of human molecules, including hormones, cytokines, and other circulatory metabolites on microbial characteristics and phenotypes. Finally, we discussed the challenges and trends for developing and deciphering molecule-mediated bidirectional interactions and their potential applications in the guard of human health.

Ecology and Evolutionary Biology as Frameworks to Study Wine Fermentations

Winemaking has leveraged microbiology to enhance wine quality, typically by engineering and inoculating individual yeast strains with desirable traits. However, yeast strains do not grow alone during wine fermentation, rather they are embedded in diverse and evolving microbial communities exhibiting complex ecological dynamics. Understanding and predicting the interplay between the yeast community over the course of the species succession and the chemical matrix of wine can benefit from recognising that wine, like all microbial ecosystems, is subject to general ecological and evolutionary rules. In this piece, we outline how conceptual and methodological frameworks from community ecology and evolutionary biology can assist wine yeast researchers in improving wine fermentation processes by understanding the mechanisms governing population dynamics, predicting and engineering these important microcosms, and unlocking the genetic potential for wine strain development.

The architecture of theory and data in microbiome design: towards an S-matrix for microbiomes

Designing microbiomes for applications in health, bioengineering, and sustainability is intrinsically linked to a fundamental theoretical understanding of the rules governing microbial community assembly. Microbial ecologists have used a range of mathematical models to understand, predict, and control microbiomes, ranging from mechanistic models, putting microbial populations and their interactions as the focus, to purely statistical approaches, searching for patterns in empirical and experimental data. We review the success and limitations of these modeling approaches when designing novel microbiomes, especially when guided by (inevitably) incomplete experimental data. Although successful at predicting generic patterns of community assembly, mechanistic and phenomenological models tend to fall short of the precision needed to design and implement specific functionality in a microbiome. We argue that to effectively design microbiomes with optimal functions in diverse environments, ecologists should combine data-driven techniques with mechanistic models - a middle, third way for using theory to inform design.

Synthetic Ecosystems: From the Test Tube to the Biosphere

The study of ecosystems, both natural and artificial, has historically been mediated by population dynamics theories. In this framework, quantifying population numbers and related variables (associated with metabolism or biological-environmental interactions) plays a central role in measuring and predicting system-level properties. As we move toward advanced technological engineering of cells and organisms, the possibility of bioengineering ecosystems (from the gut microbiome to wildlands) opens several questions that will require quantitative models to find answers. Here, we present a comprehensive survey of quantitative modeling approaches for managing three kinds of synthetic ecosystems, sharing the presence of engineered strains. These include test tube examples of ecosystems hosting a relatively low number of interacting species, mesoscale closed ecosystems (or ecospheres), and macro-scale, engineered ecosystems. The potential outcomes of synthetic ecosystem designs and their limits will be relevant to different disciplines, including biomedical engineering, astrobiology, space exploration and fighting climate change impacts on endangered ecosystems. We propose a space of possible ecosystems that captures this broad range of scenarios and a tentative roadmap for open problems and further exploration.

Soft Modes as a Predictive Framework for Low Dimensional Biological Systems across Scales

All biological systems are subject to perturbations: due to thermal fluctuations, external environments, or mutations. Yet, while biological systems are composed of thousands of interacting components, recent high-throughput experiments show that their response to perturbations is surprisingly low-dimensional: confined to only a few stereotyped changes out of the many possible. Here, we explore a unifying dynamical systems framework - soft modes - to explain and analyze low-dimensionality in biology, from molecules to eco-systems. We argue that this one framework of soft modes makes non-trivial predictions that generalize classic ideas from developmental biology to disparate systems, namely: phenocopying, dual buffering, and global epistasis. While some of these predictions have been borne out in experiments, we discuss how soft modes allow for a surprisingly far-reaching and unifying framework in which to analyze data from protein biophysics to microbial ecology.

Variability of functional and biodiversity responses to perturbations is predictable and informative

Perturbations such as climate change, invasive species and pollution, impact the functioning and diversity of ecosystems. However diversity has many meanings, and ecosystems provide a plethora of functions. Thus, on top of the various perturbations that global change represents, there are also many ways to measure a perturbation's ecological impact. This leads to an overwhelming response variability, which undermines hopes of prediction. Here, we show that this variability can instead provide insights into hidden features of functions and of species responses to perturbations. By analysing a dataset of global change experiments in microbial soil systems we first show that the variability of functional and diversity responses to perturbations is not random; functions that are mechanistically similar tend to respond coherently. Furthermore, diversity metrics and broad functions (e.g. total biomass) systematically respond in opposite ways. We then formalise these observations to demonstrate, using geometrical arguments, simulations, and a theory-driven analysis of the empirical data, that the response variability of ecosystems is not only predictable, but can also be used to access useful information about species contributions to functions and population-level responses to perturbations. Our research offers a powerful framework for understanding the complexity of ecological responses to global change.

Diet, Microbiome, and Inflammation Predictors of Fecal and Plasma Short-Chain Fatty Acids in Humans

BACKGROUND

Gut microbes produce short-chain fatty acids (SCFAs), which are associated with broad health benefits. However, it is not fully known how diet and/or the gut microbiome could be modulated to improve SCFA production.

OBJECTIVES

The objective of this study was to identify dietary, inflammatory, and/or microbiome predictors of SCFAs in a cohort of healthy adults.

METHODS

SCFAs were measured in fecal and plasma samples from 359 healthy adults in the United States Department of Agriculture Nutritional Phenotyping Study. Habitual and recent diet was assessed using a Food Frequency Questionnaire and Automated Self-Administered 24-h Dietary Assesment Tool dietary recalls. Markers of systemic and gut inflammation were measured in fecal and plasma samples. The gut microbiome was assessed using shotgun metagenomics. Using statistics and machine learning, we determined how the abundance and composition of SCFAs varied with measures of diet, inflammation, and the gut microbiome.

RESULTS

We show that fecal pH may be a good proxy for fecal SCFA abundance. A higher Healthy Eating Index for a habitual diet was associated with a compositional increase in fecal butyrate relative to acetate and propionate. SCFAs were associated with markers of subclinical gastrointestinal (GI) inflammation. Fecal SCFA abundance was inversely related to plasma lipopolysaccharide-binding protein. When we analyzed hierarchically organized diet and microbiome data with taxonomy-aware algorithms, we observed that diet and microbiome features were far more predictive of fecal SCFA abundances compared to plasma SCFA abundances. The top diet and microbiome predictors of fecal butyrate included potatoes and the thiamine biosynthesis pathway, respectively.

CONCLUSIONS

These results suggest that resistant starch in the form of potatoes and microbially produced thiamine provide a substrate and essential cofactor, respectively, for butyrate synthesis. Thiamine may be a rate-limiting nutrient for butyrate production in adults. Overall, these findings illustrate the complex biology underpinning SCFA production in the gut. This trial was registered at clinicaltrials.gov as NCT02367287.

Neuronal firing rate diversity lowers the dimension of population covariability

Populations of neurons produce activity with two central features. First, neuronal responses are very diverse - specific stimuli or behaviors prompt some neurons to emit many action potentials, while other neurons remain relatively silent. Second, the trial-to-trial fluctuations of neuronal response occupy a low dimensional space, owing to significant correlations between the activity of neurons. These two features define the quality of neuronal representation. We link these two aspects of population response using a recurrent circuit model and derive the following relation: the more diverse the firing rates of neurons in a population, the lower the effective dimension of population trial-to-trial covariability. This surprising prediction is tested and validated using simultaneously recorded neuronal populations from numerous brain areas in mice, non-human primates, and in the motor cortex of human participants. Using our relation we present a theory where a more diverse neuronal code leads to better fine discrimination performance from population activity. In line with this theory, we show that neuronal populations across the brain exhibit both more diverse mean responses and lower-dimensional fluctuations when the brain is in more heightened states of information processing. In sum, we present a key organizational principle of neuronal population response that is widely observed across the nervous system and acts to synergistically improve population representation.

Quantifying microbial interactions: concepts, caveats, and applications

Microbial communities are fundamental to every ecosystem on Earth and hold great potential for biotechnological applications. However, their complex nature hampers our ability to study and understand them. A common strategy to tackle this complexity is to abstract the community into a network of interactions between its members - a phenomenological description that captures the overall effects of various chemical and physical mechanisms that underpin these relationships. This approach has proven useful for numerous applications in microbial ecology, including predicting community dynamics and stability and understanding community assembly and evolution. However, care is required in quantifying and interpreting interactions. Here, we clarify the concept of an interaction and discuss when interaction measurements are useful despite their context-dependent nature. Furthermore, we categorize different approaches for quantifying interactions, highlighting the research objectives each approach is best suited for.

Food web context modifies predator foraging and weakens trophic interaction strength

Trophic interaction modifications (TIM) are widespread in natural systems and occur when a third species indirectly alters the strength of a trophic interaction. Past studies have focused on documenting the existence and magnitude of TIMs; however, the underlying processes and long-term consequences remain elusive. To address this gap, we experimentally quantified the density-dependent effect of a third species on a predator's functional response. We conducted short-term experiments with ciliate communities composed of a predator, prey and non-consumable 'modifier' species. In both communities, increasing modifier density weakened the trophic interaction strength, due to a negative effect on the predator's space clearance rate. Simulated long-term dynamics indicate quantitative differences between models that account for TIMs or include only pairwise interactions. Our study demonstrates that TIMs are important to understand and predict community dynamics and highlights the need to move beyond focal species pairs to understand the consequences of species interactions in communities.

Global epistasis and the emergence of function in microbial consortia

The many functions of microbial communities emerge from a complex web of interactions between organisms and their environment. This poses a significant obstacle to engineering microbial consortia, hindering our ability to harness the potential of microorganisms for biotechnological applications. In this study, we demonstrate that the collective effect of ecological interactions between microbes in a community can be captured by simple statistical models that predict how adding a new species to a community will affect its function. These predictive models mirror the patterns of global epistasis reported in genetics, and they can be quantitatively interpreted in terms of pairwise interactions between community members. Our results illuminate an unexplored path to quantitatively predicting the function of microbial consortia from their composition, paving the way to optimizing desirable community properties and bringing the tasks of predicting biological function at the genetic, organismal, and ecological scales under the same quantitative formalism.

Challenges in quantifying functional redundancy and selection in microbial communities

Microbiomes can exhibit large variations in species abundances but high reproducibility in abundances of functional units, an observation often considered evidence for functional redundancy. Based on such reduction in functional variability, selection is hypothesized to act on functional units in these ecosystems. However, the link between functional redundancy and selection remains unclear. Here, we show that reduction in functional variability does not always imply selection on functional profiles. We propose empirical null models to account for the confounding effects of statistical averaging and bias toward environment-independent beneficial functions. We apply our models to existing data sets, and find that the abundances of metabolic groups within microbial communities from bromeliad foliage do not exhibit any evidence of the previously hypothesized functional selection. By contrast, communities of soil bacteria or human gut commensals grown in vitro are selected for metabolic capabilities. By separating the effects of averaging and functional bias on functional variability, we find that the appearance of functional selection in gut microbiome samples from the Human Microbiome Project is artifactual, and that there is no evidence of selection for any molecular function represented by KEGG orthology. These concepts articulate a basic framework for quantifying functional redundancy and selection, advancing our understanding of the mapping between microbiome taxonomy and function.

Data-driven prediction of colonization outcomes for complex microbial communities

Microbial interactions can lead to different colonization outcomes of exogenous species, be they pathogenic or beneficial in nature. Predicting the colonization of exogenous species in complex communities remains a fundamental challenge in microbial ecology, mainly due to our limited knowledge of the diverse mechanisms governing microbial dynamics. Here, we propose a data-driven approach independent of any dynamics model to predict colonization outcomes of exogenous species from the baseline compositions of microbial communities. We systematically validate this approach using synthetic data, finding that machine learning models can predict not only the binary colonization outcome but also the post-invasion steady-state abundance of the invading species. Then we conduct colonization experiments for commensal gut bacteria species Enterococcus faecium and Akkermansia muciniphila in hundreds of human stool-derived in vitro microbial communities, confirming that the data-driven approaches can predict the colonization outcomes in experiments. Furthermore, we find that while most resident species are predicted to have a weak negative impact on the colonization of exogenous species, strongly interacting species could significantly alter the colonization outcomes, e.g., Enterococcus faecalis inhibits the invasion of E. faecium invasion. The presented results suggest that the data-driven approaches are powerful tools to inform the ecology and management of microbial communities.

More than the sum of its parts: uncovering emerging effects of microbial interactions in complex communities

Microbial communities are not only shaped by the diversity of microorganisms and their individual metabolic potential, but also by the vast amount of intra- and interspecies interactions that can occur pairwise interactions among microorganisms, we suggest that more attention should be drawn towards the effects on the entire microbiome that emerge from individual interactions between community members. The production of certain metabolites that can be tied to a specific microbe-microbe interaction might subsequently influence the physicochemical parameters of the habitat, stimulate a change in the trophic network of the community or create new micro-habitats through the formation of biofilms, similar to the production of antimicrobial substances which might negatively affect only one microorganism but cause a ripple effect on the abundance of other community members. Here, we argue that combining established as well as innovative laboratory and computational methods is needed to predict novel interactions and assess their secondary effects. Such efforts will enable future microbiome studies to expand our knowledge on the dynamics of complex microbial communities.

Strategies for tailoring functional microbial synthetic communities

Natural ecosystems harbor a huge reservoir of taxonomically diverse microbes that are important for plant growth and health. The vast diversity of soil microorganisms and their complex interactions make it challenging to pinpoint the main players important for the life support functions microbes can provide to plants, including enhanced tolerance to (a)biotic stress factors. Designing simplified microbial synthetic communities (SynComs) helps reduce this complexity to unravel the molecular and chemical basis and interplay of specific microbiome functions. While SynComs have been successfully employed to dissect microbial interactions or reproduce microbiome-associated phenotypes, the assembly and reconstitution of these communities have often been based on generic abundance patterns or taxonomic identities and co-occurrences but have only rarely been informed by functional traits. Here, we review recent studies on designing functional SynComs to reveal common principles and discuss multidimensional approaches for community design. We propose a strategy for tailoring the design of functional SynComs based on integration of high-throughput experimental assays with microbial strains and computational genomic analyses of their functional capabilities.

Microbial community composition predicts bacterial production across ocean ecosystems

Microbial ecological functions are an emergent property of community composition. For some ecological functions, this link is strong enough that community composition can be used to estimate the quantity of an ecological function. Here, we apply random forest regression models to compare the predictive performance of community composition and environmental data for bacterial production (BP). Using data from two independent long-term ecological research sites-Palmer LTER in Antarctica and Station SPOT in California-we found that community composition was a strong predictor of BP. The top performing model achieved an R2 of 0.84 and RMSE of 20.2 pmol L-1 hr-1 on independent validation data, outperforming a model based solely on environmental data (R2 = 0.32, RMSE = 51.4 pmol L-1 hr-1). We then operationalized our top performing model, estimating BP for 346 Antarctic samples from 2015 to 2020 for which only community composition data were available. Our predictions resolved spatial trends in BP with significance in the Antarctic (P value = 1 × 10-4) and highlighted important taxa for BP across ocean basins. Our results demonstrate a strong link between microbial community composition and microbial ecosystem function and begin to leverage long-term datasets to construct models of BP based on microbial community composition.

Solving polymicrobial puzzles: evolutionary dynamics and future directions

Polymicrobial infections include various microorganisms, often necessitating different treatment methods than a monomicrobial infection. Scientists have been puzzled by the complex interactions within these communities for generations. The presence of specific microorganisms warrants a chronic infection and impacts crucial factors such as virulence and antibiotic susceptibility. Game theory is valuable for scenarios involving multiple decision-makers, but its relevance to polymicrobial infections is limited. Eco-evolutionary dynamics introduce causation for multiple proteomic interactions like metabolic syntropy and niche segregation. The review culminates both these giants to form evolutionary dynamics (ED). There is a significant amount of literature on inter-bacterial interactions that remain unsynchronised. Such raw data can only be moulded by analysing the ED involved. The review culminates the inter-bacterial interactions in multiple clinically relevant polymicrobial infections like chronic wounds, CAUTI, otitis media and dental carries. The data is further moulded with ED to analyse the niche colonisation of two notoriously competitive bacteria: S.aureus and P.aeruginosa. The review attempts to develop a future trajectory for polymicrobial research by following recent innovative strategies incorporating ED to curb polymicrobial infections.

An evolution-based framework for describing human gut bacteria

The human gut microbiome contains many bacterial strains of the same species ('strain-level variants'). Describing strains in a biologically meaningful way rather than purely taxonomically is an important goal but challenging due to the genetic complexity of strain-level variation. Here, we measured patterns of co-evolution across >7,000 strains spanning the bacterial tree-of-life. Using these patterns as a prior for studying hundreds of gut commensal strains that we isolated, sequenced, and metabolically profiled revealed widespread structure beneath the phylogenetic level of species. Defining strains by their co-evolutionary signatures enabled predicting their metabolic phenotypes and engineering consortia from strain genome content alone. Our findings demonstrate a biologically relevant organization to strain-level variation and motivate a new schema for describing bacterial strains based on their evolutionary history.