Additionalanalysis

analyseObjectiveShadowPrices(modelFolder, objectiveList, varargin)[source]

This function determines the shadow prices indicating metabolites that are relevant for the flux through one or multiple objective functions optimized in one or more COBRA model structures. The objective functions entered are optimized one by one. By default, all metabolites with nonzero shadow prices are extracted from the computed flux solutions. The function was written for the Microbiome Modeling Toolbox, with the aim of allowing the computation of metabolic dependencies for metabolites of interest secreted by the personalzied models, though it can also be used for any other COBRA model.

USAGE

[objectives,shadowPrices]=analyseObjectiveShadowPrices (modelFolder,objectiveList,varargin)

INPUTS

  • modelFolder Folder containing one or more COBRA model structures

  • objectiveList Cell array containing the names of one or more – objective functions of interest in vertical order Optional: second column with exchange reaction IDs for objective-specific precursors

OPTIONAL INPUTS

  • resultsFolder char with path of directory where results are saved

  • osenseStr String indicating whether objective function (s) – should be maximized or minimized. Allowed inputs: ‘min’,’max’, default:’max’.

  • SPDef String indicating whether positive, negative, or – all nonzero shadow prices should be collected. Allowed inputs: ‘Positive’,’Negative’,’Nonzero’, default: ‘Nonzero’.

  • numWorkers Number indicating number of workers in parallel pool – (default: 0).

OUTPUT

  • objectives Computed objectives values

  • shadowPrices Table with shadow prices for metabolites that are – relevant for each analyzed objective in each analyzed model

analyzeMgPipeResults(infoFilePath, resPath, varargin)[source]

This function takes simulation results generated by mgPipe as the input and determines which computed fluxes and reaction abundances are significantly different between groups. Also creates violin plots of the simulation results. Requires a file with sample information (e.g., disease group, age) for the microbiome models that were generated and interrogated through mgPipe.

USAGE

analyzeMgPipeResults (infoFilePath,resPath,varargin)

REQUIRED INPUTS: infoFilePath: Path to text file or spreadsheet with information

on analyzed samples including group classification with sample IDs as rows

resPath: char with path of directory where simulation

results are saved

OPTIONAL INPUTS statPath: char with path of directory where results of

statistical analysis are saved

violinPath: char with path of directory where violin plots are

saved

plotType Type of plot to be created: violin plot (default)

or boxplot. Allowed entries: “ViolinPlot”,”Boxplot”

sampleGroupHeaders list of one or more column headers in file with the

sample information that should be analyzed (e.g., disease status, age). If not provided, the second column will be used.

calculateReactionAbundance(abundancePath, modelPath, taxonomyPath, rxnsList, numWorkers)[source]

Part of the Microbiome Modeling Toolbox. This function calculates and plots the total abundance of reactions of interest in a given microbiome sample based on the strain-level composition. Reaction presence or absence in each strain is derived from the reaction content of the respective AGORA model. Two results are given: the total abundance, and the abundance on different taxonomical levels.

USAGE

[ReactionAbundance,TaxonomyInfo] = calculateReactionAbundance(abundancePath, modelPath, taxonomyPath, rxnsList, numWorkers)

INPUTS

  • abundancePath – Path to the .csv file with the abundance data. Example: ‘cobratoolbox/papers/018_microbiomeModelingToolbox/examples/normCoverage.csv’

  • modelPath – Folder containing the strain-specific AGORA models

OPTIONAL INPUTS

  • taxonomyPath – Path to the spreadsheet with the taxonomy information on organisms (default: AGORA_infoFile.xlsx)

  • rxnsList – List of reactions for which the abundance should be calculated (if left empty: all reactions in all models)

  • numWorkers – Number of workers used for parallel pool. If left empty, the parallel pool will not be started. Parallellization is recommended if all reactions are computed.

OUTPUT

ReactionAbundance Structure with abundance for each microbiome – and reaction in total and on taxon levels TaxonomyInfo: Taxonomical information on each taxon level

calculateSubsystemAbundance(reactionAbundancePath)[source]

Computes the subsystem abundance in microbiome models that were generated through the mgPipe pipeline from the calculated reaction abundances. The subsystem abundance is shown as the fraction of the highest possible subsystem abundance that could be achieved, which would be every reaction in the subsystem being in the microbiome model at a total abundance of 1.

USAGE: subsystemAbundance = calculateSubsystemAbundance(reactionAbundancePath)

INPUT reactionAbundancePath Path to spreadsheet or text file with

calculated reaction abundances for one or more microbiome models

OUTPUT subsystemAbundance Table with calculated subsystem abundances

AUTHOR

  • Almut Heinken, 08/2020

correlateFluxWithTaxonAbundance(abundancePath, fluxPath, infoFilePath, corrMethod)[source]

Part of the Microbiome Modeling Toolbox. This function calculates and plots the correlations between fluxes for one or more reactions of interest in a number of microbiome samples and the relative microbe abundance on different taxonomical levels in the same samples. The function should be used after running mgPipe to identify correlations between the computed metabolic profiles and specific taxa in the samples.

USAGE

[FluxCorrelations, PValues, TaxonomyInfo] = correlateFluxWithTaxonAbundance(abundancePath, fluxPath, taxonomy, corrMethod)

INPUTS

abundancePath –

Path to the .csv file with the abundance data.

Needs to be in same format as example file ‘cobratoolbox/papers/018_microbiomeModelingToolbox/examples/normCoverage.csv’

fluxPath: Path to the .csv file with the fluxes for reactions

of interest with sample IDs as rows and reaction IDs in microbiome community models as columns

OPTIONAL INPUTS

infoFilePath –

Path to the spreadsheet with the taxonomy information

on organisms (default: AGORA_infoFile.xlsx)

corrMethod: Method to compute the linear correlation

coefficient. Allowed inputs: ‘Pearson’ (default), ‘Kendall’, ‘Spearman’.

OUTPUTS

  • FluxCorrelations – Structure with correlations between fluxes for each reaction and abundances on taxon levels

  • PValues – p-values corresponding to each calculated correlation

  • TaxonomyInfo – Taxonomical information on each taxon level

drawSecPot(dietPath, rxnmax, startMet, levels, cutoff)[source]

Draws part of the community structure when maximized with the given reaction

INPUTS

dietPath char with path of the model

rxnmax char with name of the rxn that gets maximized with fastFVA startMet char with starting metabolite for visualizing levels number with levels that are connected to the starting

metabolite

cutoff number of fluxes that are considered for drawing,

only fluxes>abs(cutoff) are considered

fastCalculateReactionAbundance(abundancePath, modelPath, rxnsList, numWorkers)[source]

Part of the Microbiome Modeling Toolbox. This function calculates and plots the total abundance of reactions of interest in a given microbiome sample based on the strain-level composition. Reaction presence or absence in each strain is derived from the reaction content of the respective AGORA model.

USAGE

ReactionAbundance = fastCalculateReactionAbundance(abundancePath, modelPath, rxnsList, numWorkers)

INPUTS

  • abundancePath – Path to the .csv file with the abundance data. Example: ‘cobratoolbox/papers/018_microbiomeModelingToolbox/examples/normCoverage.csv’

  • modelPath – Folder containing the strain-specific AGORA models

OPTIONAL INPUTS

  • rxnsList – List of reactions for which the abundance should be calculated (if left empty: all reactions in all models)

  • numWorkers – Number of workers used for parallel pool. If left empty, the parallel pool will not be started. Parallellization is recommended if all reactions are computed.

OUTPUT

  • ReactionAbundance Table with total abundance for each microbiome – and reaction

  • ReactionPresence – Table with absolute reaction presence for all reactions in all microbiome samples

inorder(node, model, visitedmets, visitedrxns, levels)[source]

Concatenates all reactions and metabolites connected over levels

USAGE

[visitedmets, visitedrxns] = inorder (node, model, visitedmets, visitedrxns, levels)

INPUTS

  • node – char with the name of the starting node

  • model – COBRA model of a community/microbe one microbe model that can carry flux

  • visitedmets – list of metabolites that are excluded (already visited)

  • visitedrxns –

    list of reactions that are excluded

    (already visited)

    levels indicates how many levels from the starting node (metabolite) are

    considered

OUTPUTS

  • visitedmets – list of all metabolites that are connected over the levels

  • visitedrxns – list of all reactions that are connected over the levels

makeViolinPlots(sampleData, sampleInformation, varargin)[source]

This function creates violin plots of input data (e.g., community model fluxes) while separating the data into two or more violins based on available sample stratification information.

REQUIRED INPUTS sampleData Table with input data to analyze (e.g., fluxes) with

computed features as rows and sample IDs as columns

sampleInformation Table with information on analyzed samples including

group classification with sample IDs as rows

OPTIONAL INPUTS stratification Column header containing the desired group

classification in sampleInformation table. If not provided, the second column will be used.

plotType Type of plot to be created: violin plot (default) or

boxplot. Allowed entries: “ViolinPlot”,”Boxplot”

plottedFeature Name of the feature to plot that will be displayed

as the plot title (e.g., ‘Flux”es’)

unit Unit of the plotted data that will be displayed as

the y axis label (e.g., mmol/person/day)

AUTHOR

  • Almut Heinken, 12/2020

performStatisticalAnalysis(sampleData, sampleInformation, varargin)[source]

This function determines if there is a significant difference between features computed for two or more groups in a cohort of samples. If the cohort contains two groups, the Wilcoxon rank sum test is used. If the cohort contains three or more groups, the Kruskal Wallis test is used.

USAGE [Statistics,significantFeatures] = performStatisticalAnalysis(sampleData,sampleInformation,stratification)

INPUTS sampleData Table with input data to analyze (e.g., fluxes) with

computed features as rows and sample IDs as columns

sampleInformation Table with information on analyzed samples including

group classification with sample IDs as rows

OPTIONAL INPUT stratification Column header containing the desired group

classification in sampleInformation table. If not provided, the second column will be used.

groupTest Decides whether Kruskal-Wallis test(default) or

ANOVA should be used for group comparisons. Allowed inputs: “Kruskal-Wallis”,”ANOVA”

OUTPUTS Statistics Table with results of statistical tests for each

computed feature

significantFeatures Table with input data reduced to only features that

were statistically significant

AUTHOR

  • Almut Heinken, 12/2020

plotFluxesAgainstOrganismAbundances(abundancePath, fluxPath, metabolites)[source]

Part of the Microbiome Modeling Toolbox. This function plots the relationship between organism abundances and the computed flux profiles for metabolites of interest. The function should be used after running mgPipe to identify correlations between metabolites and specific taxa in the samples.

USAGE

plotFluxesAgainstOrganismAbundances(fluxPath,abundancePath,metabolites)

INPUTS

  • abundancePath – Path to the .csv file with the abundance data. Needs to be in same format as example file ‘cobratoolbox/papers/018_microbiomeModelingToolbox/examples/normCoverage.csv’

  • fluxPath – Path to the .csv file with the fluxes for reactions of interest with sample IDs as rows and reaction IDs in microbiome community models as columns

OPTIONAL INPUT

metabolites – Cell array with VMH metabolite IDs of metabolites to plot (default: all computed metabolites)

predictMicrobeContributions(modPath, varargin)[source]

Predicts the minimal and maximal fluxes through internal exchange reactions in microbes in a list of microbiome community models for a list of metabolites. This allows for the prediction of the individual contribution of each microbe to total metabolite uptake and secretion by the community.

[minFluxes,maxFluxes,fluxSpans] = predictMicrobeContributions(modPath, varargin)

INPUTS

modPath char with path of directory where models are stored

OPTIONAL INPUTS

  • metList List of VMH IDs for metabolites to analyze – (default: all exchanged metabolites)

  • resultsFolder Path where results will be saved

  • numWorkers integer indicating the number of cores to use – for parallelization

OUTPUTS

  • minFluxes – Minimal fluxes through analyzed exchange reactions, corresponding to secretion fluxes for each microbe

  • maxFluxes – Maximal fluxes through analyzed exchange reactions, corresponding to uptake fluxes for each microbe

  • fluxSpans – Range between min and max fluxes for analyzed exchange reactions

Define default input parameters if not specified