ecosystem

Supplementary material "Unveiling abundance-dependent metabolic phenotypes of microbial communities", consisting on an framework that allows to explore the abundance-growth space, defined by both the composition of a community as well as its growth rate. By performing a cualitative characterization of each point of these space, this method allows to characterize metabolic phenotypes in a community without the imposition of a biomass maximization objective function.

This framework is implemented in Python and uses the individual genome-scale models as an input to construct a community model in an Ecosystem object as described in the available tutorials on this repository.

Install instructions

Currently the raw code can be directly downloaded to be used from this repository

full_ecosystem.py: class of the ecosystem object
multiple_ecosystem.py: class of the ecosystem object optimized for multiple analysis (no model is loaded for each scenario for faster computation, only grid and results computed using full_ecosystem.py)

Basic usage

A basic set of command that allow to perform these analysis is described bellow, for a more detailed description see sup_toy_example.ipynb which includes additional explanation and plots.

Initiation and simulation conditions

To generate an 'Ecosystem' object for a community of two organisms with models model_1 and model_2, which contains a constrained-based model of the community:

eco2 = Ecosystem(models=[model_1, model_2], prefixes=['org1', 'org2], community_name = 'my first community', community_id = 'com_2', solver=gurobi)

To constrain inputs of nutrients (in this case metabolite A_e) to the common space shared by all members of the community:

eco2.set_pool_bounds({'A_e':(-lb,1000)}, bioCons=-bc)

This function also accepts an additional parameter where biological constraints (bioCons) can be set up for the organisms of the community.

To generate a grid of 10x10 for analysis:

eco2.set_cluster_reactions()
eco2.build_grid(numPoints = 10, expand = True, drop_zero=True)
eco2.get_member_reactions()
eco2.get_points_distribution()

Grid analysis and clustering

To analyze the grid different tests can be performed (feasibility, qualitative Flux Variability analysis (FVA)), as follows:

eco2.analyze_grid(analysis = 'feasibility', update_bounds=True)

checks the feasibility of each point of the pre-computed grid of abundance and community growth rates or

eco2.analyze_grid(analysis = 'qual_fva', update_bounds=True)

which performs qualitative FVA on each point of the grid. The output of this analysis can be analyzed by clustering a defined number of clusters (num_clusters):

eco2.clusterPoints('hierarchical', k = num_clusters)

which can be plot using:

eco2.plot_2D_slice(prefixes=[], fixed_values=[], parent_cmap='tab20c',s=70, figsize=(12,12), 
                         to_plot = 'cluster', show_edge=False,frac_prefix= None,
                            xlabel = '$f_{org1}$',
                            ylabel ='$Biomass_{community}$', saveFile='clusters_toy')

or analyzed on a table format by using

eco2.get_cluster_reaction_values(thr=0.8, changing= True)

and

eco2.compare_clusters(df, 'c3','c4')

to compare two clusters (in this case c3 and c4).

Quantitative Flux Coupling Analysis

A quantitative version of Flux Coupling Analysis (FCA) can be performed on different points of the grid defined by their coordinates, and for two reactions of choice to study how distribution of fluxes changes on different points of the grid, as follows:

eco2.quan_FCA(grid_x, grid_y, rxns_analysis)
plot_qFCA(col_wrap=4)

Supplementary files

S1: Model of a synthetic Escherichia coli model
S2: Qualitative states of reactions of an E. coli synthetic community
S3: Additional reactions for model curation of A. ferrooxidans and S. thermosulfidooxidans
S4: Models A. ferrooxidans Wenelen (iML510), S. thermosulfidooxidans Cutipay (iSM517) and the community model generated for the bioleaching community in the base case scenario, in both sbml and xls format.

Tutorials and case examples

sup_toy_example.ipynb includes an analysis that can be performed locally and its associated plots
sup_coli_synthetic_com.ipynb includes an analysis that can be performed locally and its associated plots for a synthetic community of two E. coli strains
sup_mat_multiple_bioleaching.ipynb includes analysis of multiple scenarios simultaneously from previously computed results stored on the folder result_files. It requires both models included in Supplementary File 2, simulations_pareto4.csv to filter the different simulated scenarios and the files available on the folder result_files

How to cite

Jiménez, NE., Acuña V., Cortés, MP., Eveillard, D., Maass A., (2023) Unveiling abundance-dependent metabolic phenotypes of microbial communities

mathomics/ecosystem