From Genomes to Watersheds: Functional Abstractions of Microbial Life

[Grigory Bronevetsky]

Modeling Talks

From Genomes to Watersheds: Functional Abstractions of Microbial Life Eoin Brodie, Berkeley Lab Tues, January 27, 2025 | 9am PT Meet | Youtube Stream Hi all, The presentation will be via Meet and all questions will be addressed there. If you cannot attend live, the event will be recorded and can be found afterward at https://sites.google.com/modelingtalks.org/entry/from-genomes-to-watersheds-functional-abstractions-of-microbial-life More information on previous and future talks: https://sites.google.com/modelingtalks.org/entry/home Abstract: This presentation will outline a trait-based approach for exploring and simulating soil microbiomes, crucial biological components that govern key ecological processes like climate regulation and nutrient cycling. Given that the vast majority of microbial species remain uncultivated, genomic data is the primary window into their function. Soil microbiomes are complex, and to handle this complexity, our framework distills genomes into "Functional Traits" - attributes that mechanistically link a microbe's genes to its fitness and ecosystem-level impact. This approach achieves essential dimensionality reduction and is designed to unify genomic, phenotypic, and ecological data for cross-system generalizability. The core concept is the Genomes-to-Ecosystem (G2E) framework, which integrates these derived traits into complex ecosystem models such as ecosys/EcoSIM. This integration is critical for accurate parameterization of microbial functional groups and has demonstrated that predictions of biogeochemical fluxes, like methane emissions, are highly sensitive to the trait parameters used. Research leveraging this framework explores trade-offs in carbon and energy allocation, finding, for example, that Carbon Use Efficiency (CUE) can be manipulated by selecting for efficient organisms via specific substrate additions. Ultimately, the process needs to be improved through more consistent measurements, model-guided field sensing, and the introduction of AI-Agents to automate experimentation and make complex models accessible. Bio: Eoin Brodie, Ph.D. is a Senior Scientist at Lawrence Berkeley National Laboratory and Deputy Director of the Climate and Ecosystem Sciences Division, as well as an Adjunct Professor at University of California, Berkeley and Co-Director of the Joint Berkeley Initiative in Microbiome Sciences. A microbiologist and biogeochemist, he studies how microscopic life shapes the health and resilience of soils, ecosystems, and the planet. His research integrates molecular biology, advanced sensing, and trait-based computational modeling to predict microbial function from genes to entire watersheds, with the goal of harnessing microbiomes to advance sustainable agriculture, ecosystem resilience, and environmental restoration.

[Grigory Bronevetsky]

Video Recording: https://youtube.com/live/QS40gscmhTc

Slides: https://drive.google.com/file/d/1_-vH2IIRSxKdx-FadgUkfMW-O-2qen1x/view?usp=sharing

Summary:

• Focus: Dynamics of organisms across gradient of energy

  • Energy

  • Water

  • Carbon

  • Nutrients

• Soil microbes: very valuable for soils, ecology and human use

  • Poorly known:

    • 400k species / 60k genera observed (mostly using genetic tools)

    • Predicted 1012 species total

    • 14k species have been cultivated, 97% from 4 phyla

    • We mostly know these organisms based on their genomes

  • Soils have a vast organism diversity

  • Approach: analyze genomes via their Functional Traits linked to their fitness

    • E.g. morphology, physiology, behavior

    • Numerical models are key for this analysis

      • Synthesis of trait data

      • Definition of functional traits

      • Traits are direct parameters or inform model structure

      • Extend existing trait databases

      • Infer organism niches from genomes

• Microbe dynamics

  • Entropy production machines

    • Structured via DNA

    • Burning exergy

    • Releasing heat

    • Dumping wase

  • Traits: growth rate, biomass stoichiometry, biosynthesis potential, optimum temp, substrate uptake kinetics, etc. 

  • Tradeoffs in carbon-energy allocation: whole-organism perspective

    • Resource acquisition

    • Design constraint

    • Resource allocation

    • Information processing

  • Dynamic Energy Budget Model, constrained by genome-informed traits

• Major challenge: representing constraints across scales

  • Global->landscape->soil column->particle pores/niches

  • Genomes-to-Ecosystems (G2E) Framework

  • microTrait: distills genomes to traits that impact fitness

    • https://github.com/ukaraoz/microtrait

    • Genomes -> trait matrices

    • Hidden Markov Models

    • Predict proteins -> metabolic pathways -> traits

    • Traits: life history, biophysical, thermodynamic, metabolic

    • Collapse genomes with similar traits into “guilds”

    • Need distribution of parameters within a guild

  • DEBmicroTrait: microTrait + biophysical, metabolic, thermodynamic traits to model behavior

    • Infers other traits: cell volume, stoichiometry, substrate uptake kinetics

    • Then predict how cells behave behave overall

• Soils are a globally important carbon reservoir

  • 3X more than atmosphere, 4x more than vegetation

  • Store carbon from vegetation for long amount of time

  • Microbes are a major modulator of this process

    • Low efficiency: eat and emit carbon

    • High efficiency: consume lots of carbon, die and store it in their bodies

  • Soil degradation has resulted in carbon emissions (133 GT total)

  • We need to get that carbon back into the soil

  • We know

    • Microbial Carbon Use Efficiency (CUE) positively correlated to soil carbon formation

    • Carbon oxidation of substrate strongly affects CUE

    • Phylogenetic origin of organism affects CUE 

    • Studied the chemical signaling between plants and soil bacteria

      • Succession of chemicals in the interaction

      • Observation: slow-growing bacteria are surprisingly competitive in the rhizosphere (root network)

      • Different population of microbes

        • Fast growers that consume simple sugars

        • Slow growers that are more efficient (CUE) and consume more complex molecules

        • Slow growers are able to effectively compete for other resources due to their higher CUE

      • Prediction: can manipulate CUE through substrate addition

      • Requires balance of other nutrients (K, P, N) to maximize CUE)

• Given microbial traits, can we predict biogeochemical flux?

  • Permafrost melting is releasing methane

  • Modulated by microbes growing in this environment

  • Experimental data used to parameterize ecosys model: https://github.com/jinyun1tang/ECOSYS

  • Chemical flows modulated by traits at the ecosystem scale

  • Strong sensitivity of CH4 emission to microbial trait parameterization across different categories of microbes (fermenters, methanotrophes, etc.)

  • 
    

• When analyzing the behavior of a soil ecosystem, which genomes do we focus on?

  • All present genomes (too diverse, includes irrelevant species)

  • Dominant genomes (misses to much detail)

  • Community-aggregated approach (best balance)

• Next generation: Ecosim

  • https://ecosim.lbl.gov/

  • https://github.com/jinyun1tang/EcoSIM

• Modeling direction: Watershed Modeling focus area: https://watershed.lbl.gov

  • Combination of traits of entire watersheds into an entire simulation framework

  • Observational data

    • LiDAR

    • Hyperspectral imagery of plants (estimates of Carbon, Water, Nitrogen)

    • Geophysical measurements of soils

    • Genome-resolved metagenomics of microbial traits