Periodic Reporting for period 1 - ECOPROSPECTOR (Mapping vast functional landscapes with single-species resolution: a new approach for precision engineering of microbial consortia)
Berichtszeitraum: 2023-03-01 bis 2025-08-31
This project (ECOPROSPECTOR) focuses on addressing a key open question in microbial biotechnology: Given a library of candidate species, which should we inoculate together in a bioreactor to form a community that optimizes a biotechnological process of interest, such as the production of high-value molecules? Traditional approaches to answer this question have struggled due to the complexity and astronomical number of potential interactions among multiple species. For example, choosing optimal communities from a pool of just 100 microbial species involves navigating approximately 10³⁰ possible assemblages. As a result of these interactions, it has been difficult to predict how adding or removing a single species might boost or hinder the overall performance of the communities.
ECOPROSPECTOR addresses this bottleneck by leveraging conceptual and methodological tools of Quantitative Genetics, a field that has tackled the similar problem of similarly vast combinatorial spaces (of mutations, in that case) and similarly complex potential interactions (which are known in genetics as “epistasis”). The objective of ECOPROSPECTOR is to extend the theory of epistasis in genetics to ecology, using it to develop a new theoretical and computational paradigm for predicting the quantitative properties (or functions) of microbial communities based on their composition at the single-species level. This approach is supported by preliminary findings suggesting that the effect of adding a species to a community can often be forecasted by a simple, yet powerful, mathematical relationship that mirrors similar findings in genetics.
The project begins with an empirical focus on a tractable library of soil bacteria, aiming to identify and model these predictive relationships between community composition and function. Machine learning tools will then be used to reconstruct the high-dimensional landscape connecting community composition to ecosystem function (the production and secretion of biotechnologically relevant molecules, such as enzymes or peptides). This theoretical map will guide the search for optimal communities. Genetic, environmental, and modeling techniques will be employed to mechanistically explain the discovered patterns and link them to specific species traits.
The anticipated impact will be large, both practically (by enabling rational design of high-functioning microbial consortia for industrial or environmental use) and theoretical—laying the groundwork for a unifying framework that bridges evolutionary and ecological theory, with wide-reaching implications for both fields. We envision three important axis of impact
Expected Impact
• Fundamental scientific knowledge: By determining how species interactions shape community function, we’ll help push microbial ecology to a new level of quantitative understanding.
• Industrial and clinical applications: Our methods can benefit areas like biotechnology, where microbial consortia are key (e.g. fermentation, food production, antibiotic resistance research).
• Innovation: If we can reliably predict community behavior, this will pave the way to new collaborations, industrial partnerships, and potentially spin-off applications (like improved microbial production platforms).
ECOPROSPECTOR addresses this bottleneck by leveraging conceptual and methodological tools of Quantitative Genetics, a field that has tackled the similar problem of similarly vast combinatorial spaces (of mutations, in that case) and similarly complex potential interactions (which are known in genetics as “epistasis”). The objective of ECOPROSPECTOR is to extend the theory of epistasis in genetics to ecology, using it to develop a new theoretical and computational paradigm for predicting the quantitative properties (or functions) of microbial communities based on their composition at the single-species level. This approach is supported by preliminary findings suggesting that the effect of adding a species to a community can often be forecasted by a simple, yet powerful, mathematical relationship that mirrors similar findings in genetics.
The project begins with an empirical focus on a tractable library of soil bacteria, aiming to identify and model these predictive relationships between community composition and function. Machine learning tools will then be used to reconstruct the high-dimensional landscape connecting community composition to ecosystem function (the production and secretion of biotechnologically relevant molecules, such as enzymes or peptides). This theoretical map will guide the search for optimal communities. Genetic, environmental, and modeling techniques will be employed to mechanistically explain the discovered patterns and link them to specific species traits.
The anticipated impact will be large, both practically (by enabling rational design of high-functioning microbial consortia for industrial or environmental use) and theoretical—laying the groundwork for a unifying framework that bridges evolutionary and ecological theory, with wide-reaching implications for both fields. We envision three important axis of impact
Expected Impact
• Fundamental scientific knowledge: By determining how species interactions shape community function, we’ll help push microbial ecology to a new level of quantitative understanding.
• Industrial and clinical applications: Our methods can benefit areas like biotechnology, where microbial consortia are key (e.g. fermentation, food production, antibiotic resistance research).
• Innovation: If we can reliably predict community behavior, this will pave the way to new collaborations, industrial partnerships, and potentially spin-off applications (like improved microbial production platforms).
Work Performed
• High-throughput community assembly protocol
We have developed a simple yet powerful liquid-handling method that can generate all possible sub-communities from a chosen set of up to 10–12 species. Published in eLife (2024), this approach drastically cuts down on labor and time and overcomes the typical bottlenecks of constructing hundreds (or even thousands) of microbial combinations.
• Testing the “global epistasis” concept
In a pilot study with eight different strains, we found that each strain’s effect on community function can be modeled using what we call “Functional Effect Equations” (FEEs). These equations show how a species’ contribution depends on the baseline function of the community. These results were featured in a recent article in Cell (2024).
• Mechanistic insights into species interactions
Beyond the descriptive models, we dug into how pairwise and higher-order interactions shape community function. A core discovery was that a species’ net effect is driven by the sum of its interactions with all other species. This confirms the project’s main hypothesis that epistasis-like patterns apply to entire microbial ecosystems.
• Context dependence
We also tested how environmental conditions—like nutrient availability or drug concentrations—shift these interaction patterns. Studies such as our work in Nature Communications (2023) highlight how interactions can flip from positive to negative under different conditions, underscoring the importance of context in microbial consortia.
Main Achievements
• Full factorial community assembly technique: A key achievement allowing labs worldwide to build large sets of combinations quickly, published in eLife (2024).
• Empirical demonstration of global epistasis in microbial consortia: Shown in Cell (2024), providing a robust predictive framework for understanding community-wide behavior.
• High-impact publications: The project’s core findings have been published in journals like Cell, eLife, Science, Nature Communications, and Molecular Systems Biology. To highlight the importance of our work so far, Cell commissioned an Editorial article about it.
• High-throughput community assembly protocol
We have developed a simple yet powerful liquid-handling method that can generate all possible sub-communities from a chosen set of up to 10–12 species. Published in eLife (2024), this approach drastically cuts down on labor and time and overcomes the typical bottlenecks of constructing hundreds (or even thousands) of microbial combinations.
• Testing the “global epistasis” concept
In a pilot study with eight different strains, we found that each strain’s effect on community function can be modeled using what we call “Functional Effect Equations” (FEEs). These equations show how a species’ contribution depends on the baseline function of the community. These results were featured in a recent article in Cell (2024).
• Mechanistic insights into species interactions
Beyond the descriptive models, we dug into how pairwise and higher-order interactions shape community function. A core discovery was that a species’ net effect is driven by the sum of its interactions with all other species. This confirms the project’s main hypothesis that epistasis-like patterns apply to entire microbial ecosystems.
• Context dependence
We also tested how environmental conditions—like nutrient availability or drug concentrations—shift these interaction patterns. Studies such as our work in Nature Communications (2023) highlight how interactions can flip from positive to negative under different conditions, underscoring the importance of context in microbial consortia.
Main Achievements
• Full factorial community assembly technique: A key achievement allowing labs worldwide to build large sets of combinations quickly, published in eLife (2024).
• Empirical demonstration of global epistasis in microbial consortia: Shown in Cell (2024), providing a robust predictive framework for understanding community-wide behavior.
• High-impact publications: The project’s core findings have been published in journals like Cell, eLife, Science, Nature Communications, and Molecular Systems Biology. To highlight the importance of our work so far, Cell commissioned an Editorial article about it.
• Applying global epistasis to ecology: We have shown that ideas and mathematical formalisms that were developed to understand genetic mutations in single organisms (epistasis) can be applied to entire microbial communities. This bridges two research fields that had historically developed in parallel.
• Predictive statistical modelling: We have demonstrated that the statistical regression models reflecting the global epistasis patterns we have found, can be concatenated to infer the entire relationship between community composition and function, improving upon the state-of-the-art models as demonstrated in our Cell paper.
• Mechanistic modeling: The statistical models have been explained from mechanistic interactions. By showing that a species’ effect on community function can be explained by the sum of its pairwise “effective interactions”, we provide a layer of interpretation to our models. They go beyond simply describing the patterns, providing a direct explanation for them. This paves the way to engineer traits that will improve community functions.
• High-throughput factorial approach: Our assembly method for consortia moves the field past a major experimental hurdle, making it feasible to test large numbers of species combinations in a short time, at little expense. This methodology can be widely adopted, as it requires basic laboratory equipment.
• Predictive statistical modelling: We have demonstrated that the statistical regression models reflecting the global epistasis patterns we have found, can be concatenated to infer the entire relationship between community composition and function, improving upon the state-of-the-art models as demonstrated in our Cell paper.
• Mechanistic modeling: The statistical models have been explained from mechanistic interactions. By showing that a species’ effect on community function can be explained by the sum of its pairwise “effective interactions”, we provide a layer of interpretation to our models. They go beyond simply describing the patterns, providing a direct explanation for them. This paves the way to engineer traits that will improve community functions.
• High-throughput factorial approach: Our assembly method for consortia moves the field past a major experimental hurdle, making it feasible to test large numbers of species combinations in a short time, at little expense. This methodology can be widely adopted, as it requires basic laboratory equipment.