New

createGroupIncidenceMatrix(model, trainingModel, param)

USAGE

trainingModel = createGroupIncidenceMatrix (model, trainingModel)

INPUTS: model: model.mets m x 1 metabolite ids model.inchi.nonstandard m x 1 cell array of nonstandard InChI

trainingModel: trainingModel.S: p x n stoichiometric matrix of training data trainingModel.mets p x 1 metabolite abbreviations trainingModel.rxns n x 1 reaction abbreviations trainingModel.metKEGGID: p x 1 cell array of metabolite KEGGID trainingModel.inchi.nonstandard: p x 1 cell array of nonstandard InChI trainingModel.mappingScore

OPTIONAL INPUT: trainingModel.cids_that_dont_decompose cid of kegg compounds that are not decomposable with param.fragmentationMethod=’manual’;

OUTPUT: combinedModel: combinedModel.S: k x n stoichiometric matrix of training padded with zero rows for metabolites exclusive to test data combinedModel.drG0: n x 1 experimental standard reaction Gibbs energy combinedModel.drG0_prime: n x 1 experimental standard transformed reaction Gibbs energy combinedModel.T: n x 1 temperature combinedModel.I: n x 1 ionic strength combinedModel.pH: n x 1 pH combinedModel.pMg: n x 1 pMg combinedModel.G: k x g group incidence matrix combinedModel.groups: g x 1 cell array of group definitions combinedModel.trainingMetBool k x 1 boolean indicating training metabolites in G combinedModel.testMetBool k x 1 boolean indicating test metabolites in G combinedModel.groupDecomposableBool: k x 1 boolean indicating metabolites with group decomposition combinedModel.inchiBool k x 1 boolean indicating metabolites with inchi combinedModel.test2CombinedModelMap: m x 1 mapping of model.mets to combinedModel.mets

getGroupVectorFromInchi(inchi, printLevel)

USAGE

group_def = getGroupVectorFromInchi (inchi, silent, debug)

INPUTS

• inchi

• silent

• debug – 0: No verbose output, 1: Progress information only (no warnings), 2: Progress and warnings

OUTPUT

group_def

getMappingScores(model, trainingModel)

Finds the best mapping between the model metabolites and the training model metabolites, the higher the confidence score, the more reliable the mapping

USAGE

mappingScore = getMappingScores (model, trainingModel)

INPUTS

• model – model in a COBRA structure *.mets *.metKEGGID *.model.inchi.standard *.model.inchi.standardWithStereo *.model.inchi.standardWithStereoAndCharge

• trainingModel – training model in a COBRA structure *.mets *.metKEGGID *.inchi.standard *.inchi.standardWithStereo *.inchi.standardWithStereoAndCharge

OUTPUT

mappingScore – nMet x nTrainingMet sparse matrix giving best mapping

regulariseGroupIncidenceMatrix(combinedModel, printLevel)

within the combinedModel analyse the similar metabolites (having the same group decomposition vector) and the duplicates (also the same InChI)

INPUT combinedModel printLevel

OUTPUT groupM nMet xnTrainingMet x nTrainingMet logical matrix, true if metabolite is a duplicate inchiM nTrainingMet x nModelMet logical matrix, true if metabolite is a duplicate