Merge pull request #2158 from opencobra/develop

Develop
opencobra · Jun 29, 2023 · 79e5ecc · 79e5ecc
2 parents cb64245 + 047a9fa
commit 79e5ecc
Show file tree

Hide file tree

Showing 10 changed files with 456 additions and 9 deletions.
diff --git a/external/analysis/PolytopeSamplerMatlab b/external/analysis/PolytopeSamplerMatlab
diff --git a/external/analysis/octave-networks-toolbox b/external/analysis/octave-networks-toolbox
diff --git a/src/base/io/json/model2JSON.m b/src/base/io/json/model2JSON.m
@@ -0,0 +1,264 @@
+function model2JSON(model,fileName)
+% This function writes a json file from matlab structure.
+% I validated json format with https://jsonlint.com/.
+%
+% INPUT
+% model     model structure
+% fileName  name of file including extension .json
+%
+%
+% Ines Thiele, May 2023
+
+if isempty(strfind(fileName,'.json'));
+    fileName = strcat(fileName,'.json');
+end
+
+fid = fopen(fileName, 'w');
+fprintf(fid, '{\n');
+fprintf(fid, '"metabolites":[\n');
+cnt = 1;
+% write metabolites
+for i = 1 : length(model.mets)
+    fprintf(fid, '{\n');
+    met = regexprep(model.mets{i},'\[','_');
+    met = regexprep(met,'\]','');
+    fprintf(fid,strcat( '"id":"',met,'",\n'));
+    fprintf(fid,strcat( '"name":"',model.metNames{i},'",\n'));
+    x = split(model.mets{i},'[');
+    comp = regexprep(x{2},'\]','');
+    fprintf(fid,strcat( '"compartment":"',comp,'",\n'));
+    fprintf(fid,strcat( '"charge":',num2str(model.metCharges(i)),',\n'));
+    fprintf(fid,strcat( '"formula":"',(model.metFormulas{i}),'",\n'));
+    fprintf(fid,'"notes":{\n');
+    fprintf(fid,'"original_vmh_ids":[\n');
+    fprintf(fid,strcat('"',model.mets{i},'"\n'));
+    fprintf(fid,']\n');
+    fprintf(fid,'},\n');
+    fprintf(fid,'"annotation":{\n');
+    if isfield(model, 'metBiGGID')
+        fprintf(fid,'"bigg.metabolite":[\n');
+        fprintf(fid,strcat('"',model.metBiGGID{i},'"\n'));
+        fprintf(fid,'],\n');
+    end
+    if isfield(model, 'metBioCycID')
+        fprintf(fid,'"biocyc":[\n');
+        model.metBioCycID{i} = regexprep(model.metBioCycID{i},'%','');
+        fprintf(fid,strcat('"',model.metBioCycID{i},'"\n'));
+        fprintf(fid,'],\n');
+    end
+    if isfield(model, 'metChEBIID')
+        fprintf(fid,'"chebi":[\n');
+        fprintf(fid,strcat('"',model.metChEBIID{i},'"\n'));
+        fprintf(fid,'],\n');
+    end
+    if isfield(model, 'metHMDBID')
+        fprintf(fid,'"hmdb":[\n');
+        fprintf(fid,strcat('"',model.metHMDBID{i},'"\n'));
+        fprintf(fid,'],\n');
+    end
+    if isfield(model, 'metInchiKey')
+        fprintf(fid,'"inchi_key":[\n');
+        fprintf(fid,strcat('"',model.metInchiKey{i},'"'));
+        fprintf(fid,'],');
+    end
+    if isfield(model, 'metKEGGID')
+        fprintf(fid,'"kegg.compound":[\n');
+        fprintf(fid,strcat('"',model.metKEGGID{i},'"\n'));
+        fprintf(fid,'],\n');
+    end
+    if isfield(model, 'metMetaNetXID')
+        fprintf(fid,'"metanetx.chemical":[\n');
+        fprintf(fid,strcat('"',model.metMetaNetXID{i},'"\n'));
+        fprintf(fid,'],\n');
+    end
+    if isfield(model, 'metReactomeID')
+        fprintf(fid,'"reactome.compound":[\n');
+        fprintf(fid,strcat('"',model.metReactomeID{i},'"\n'));
+        fprintf(fid,'],\n');
+    end
+    if isfield(model, 'metSabiork')
+        fprintf(fid,'"sabiork":[\n');
+        fprintf(fid,strcat('"',model.metSabiork{i},'"\n'));
+        fprintf(fid,'],\n');
+
+    end
+    if isfield(model, 'metSBOTerms')
+        fprintf(fid,strcat('"sbo":"',model.metSBOTerms{i},'",\n'));
+    end
+    if isfield(model, 'metSEEDID')
+        fprintf(fid,'"seed.compound":[\n');
+        fprintf(fid,strcat('"',model.metSEEDID{i},'"'));
+    end
+    fprintf(fid,']\n');
+    fprintf(fid,'}\n'); % close annotation
+    if i < length(model.mets)
+        fprintf(fid,'},\n'); % close metabolite
+    else
+        fprintf(fid,'}\n'); % close metabolite
+    end
+end
+% close list of metabolites
+fprintf(fid,'],\n');
+
+% write reactions
+fprintf(fid, '"reactions":[\n');
+for i = 1 : length(model.rxns)
+    fprintf(fid, '{\n');
+    rxn = model.rxns{i};
+    fprintf(fid,strcat( '"id":"',rxn,'",\n'));
+    fprintf(fid,strcat( '"name":"',model.rxnNames{i},'",\n'));
+    fprintf(fid,'"metabolites":{\n');
+    [metList, stoichiometries] = findMetsFromRxns(model,i);
+    metList= metList{1};
+    stoichiometries = stoichiometries{1};
+    for j = 1 : length(metList)
+        met = regexprep(metList{j},'\[','_');
+        met = regexprep(met,'\]','');
+        if isempty(strfind(num2str(stoichiometries(j,1)),'.'))
+            fprintf(fid,strcat('"',met,'":',num2str(stoichiometries(j,1)),'.0'));
+        else
+            fprintf(fid,strcat('"',met,'":',num2str(stoichiometries(j,1)),''));
+        end
+        if j <  length(metList)
+            fprintf(fid, ',\n');
+        else
+            fprintf(fid, '\n');
+        end
+    end
+    fprintf(fid, '},\n');
+    fprintf(fid,strcat('"lower_bound":',num2str(model.lb(i)),',\n'));
+    fprintf(fid,strcat('"upper_bound":',num2str(model.ub(i)),',\n'));
+    fprintf(fid,strcat('"gene_reaction_rule":"',model.grRules{i},'",\n'));
+    try
+        fprintf(fid,strcat('"subsystem":"',model.subSystems{i},'",\n'));
+    catch % there seems to be a cell array in some instances
+        a=model.subSystems{i};
+        fprintf(fid,strcat('"subsystem":"',a{1},'",\n'));
+    end
+    fprintf(fid,strcat('"notes":','{\n'));
+    fprintf(fid,strcat('"original_vmh_ids":','[\n'));
+    fprintf(fid,strcat('"',model.rxns{i},'"\n'));
+    fprintf(fid,']\n');
+    fprintf(fid,'},\n');
+    fprintf(fid,'"annotation":{\n');
+    fprintf(fid,'"metanetx.reaction":[\n');
+    fprintf(fid,strcat('"',model.rxnMetaNetXID{i},'"\n'));
+    fprintf(fid,'],\n');
+    fprintf(fid,strcat('"sbo":"',model.rxnSBOTerms{i},'"\n'));
+
+    fprintf(fid,'}'); % close annotation
+    if i < length(model.rxns)
+        fprintf(fid,'},\n'); % close reaction
+    else
+        fprintf(fid,'}\n'); % close reaction
+    end
+end
+% close list of reactions
+fprintf(fid,'],\n');
+
+% write genes
+% here is a more comprehensive annotation version from Recon3D
+%
+% {
+% "id":"26_AT1",
+% "name":"AOC1",
+% "notes":{
+% "original_bigg_ids":[
+% "26.1"
+% ]
+% },
+% "annotation":{
+% "ccds":[
+% "CCDS43679.1",
+% "CCDS64797.1"
+% ],
+% "ncbigene":[
+% "26"
+% ],
+% "ncbigi":[
+% "73486661",
+% "1034654825",
+% "1034654831",
+% "440918691",
+% "1034654829",
+% "1034654827"
+% ],
+% "omim":[
+% "104610"
+% ],
+% "refseq_name":[
+% "AOC1"
+% ],
+% "refseq_synonym":[
+% "KAO",
+% "DAO",
+% "DAO1",
+% "ABP",
+% "ABP1"
+% ],
+% "sbo":"SBO:0000243"
+% }
+% },
+fprintf(fid, '"genes":[\n');
+for i = 1 : length(model.genes)
+    fprintf(fid, '{\n');
+    fprintf(fid,strcat( '"id":"',model.genes{i},'",\n'));
+    fprintf(fid,strcat( '"name":"','",\n'));
+    fprintf(fid,strcat( '"notes":{','\n'));
+    fprintf(fid,strcat( '"original_vmh_ids":[','\n'));
+    fprintf(fid,strcat( '"',model.genes{i},'"\n'));
+    fprintf(fid,strcat( ']\n'));
+    fprintf(fid,strcat( '},\n'));
+    fprintf(fid,strcat( '"annotation":{','\n'));
+    fprintf(fid,strcat('"sbo":"',model.geneSBOTerms{i},'"\n'));
+
+    fprintf(fid,strcat( '}\n')); % close annotation
+
+    if i < length(model.genes)
+        fprintf(fid, '},\n');
+    else
+        fprintf(fid, '}\n');
+    end
+end
+% close list of genes
+fprintf(fid,'],\n');
+% model ID
+fprintf(fid, strcat('"id":"',model.modelID,'",\n'));
+% compartments
+[~, uniqueCompartments, ~, ~] = getCompartment(model.mets);
+fprintf(fid, strcat('"compartments":{','\n'));
+for i = 1 : length(uniqueCompartments)
+    if strcmp('c',uniqueCompartments{i})
+        fprintf(fid, '"c":"cytosol"');
+    elseif strcmp('e',uniqueCompartments{i})
+        fprintf(fid, '"e":"extracellular space"');
+    elseif strcmp('g',uniqueCompartments{i})
+        fprintf(fid, '"g":"golgi apparatus",');
+    elseif strcmp('i',uniqueCompartments{i})
+        fprintf(fid, '"i":"inner mitochondrial compartment"');
+    elseif strcmp('l',uniqueCompartments{i})
+        fprintf(fid, '"l":"lysosome"');
+    elseif strcmp('m',uniqueCompartments{i})
+        fprintf(fid, '"m":"mitochondria"');
+    elseif strcmp('n',uniqueCompartments{i})
+        fprintf(fid, '"n":"nucleus"');
+    elseif strcmp('r',uniqueCompartments{i})
+        fprintf(fid, '"r":"endoplasmic reticulum"');
+    elseif strcmp('p',uniqueCompartments{i})
+        fprintf(fid, '"p":"periplasm');
+    elseif strcmp('x',uniqueCompartments{i})
+        fprintf(fid, '"x":"peroxisome/glyoxysome"');
+    end
+    if i < length(uniqueCompartments)
+        fprintf(fid, ',\n');
+    else
+        fprintf(fid, '\n');
+    end
+end
+fprintf(fid, '},\n'); % close compartments
+if isfield(model,'modelAnnotation')
+    fprintf(fid, strcat('"version":"',model.modelAnnotation{2},'"\n'));
+end
+
+fprintf(fid, '}\n'); % close file
+fclose(fid);
diff --git a/src/base/io/writeCbModel.m b/src/base/io/writeCbModel.m
@@ -11,7 +11,7 @@
 % OPTIONAL INPUTS:
 %    varargin:          Optional parameters in 'Parametername',value
 %                       format. Available parameterNames are:
-%                       * format:   File format to be used ('text', 'xls', 'mat'(default) 'expa' or 'sbml')
+%                       * format:   File format to be used ('text', 'xls', 'mat'(default), 'expa', 'json' or 'sbml')
 %                         text will only output data from required fields (with GPR rules converted to string representation)
 %                         xls is restricted to the fields defined in the xls io documentation.
 %                         expa will print all reactions with Exchangers being detected by findExcRxns
@@ -29,6 +29,7 @@
 %       - Richard Que 3/17/10 -  Added ability to specify compartment names and symbols
 %       - Longfei Mao 26/04/2016 -  Added support for the FBCv2 format
 %       - Thomas Pfau May 2017  - Changed To Parameter/Value pairs and added flexibility
+%       - Ines Thiele May 2023 - Added option to write json files
 %
 % EXAMPLES:
 %
@@ -161,6 +162,8 @@
             [fileNameFull, filePath] = uiputfile({'*.mat'});
         case 'expa'
             [fileNameFull, filePath] = uiputfile({'*.expa'});
+        case 'json'
+            [fileNameFull, filePath] = uiputfile({'*.json'});
         case 'toselect'
             [fileNameFull, filePath] = uiputfile({'*.mat', 'Matlab File'; '*.xml' 'SBML Model'; '*.txt' 'Text Export'; '*.xls;*.xlsx' 'Excel Export'; '*.MPS' 'MPS Export'});
         otherwise
@@ -231,6 +234,9 @@
     case 'expa'
         convertModelToEX(model,fileName,0,model.rxns(findExcRxns(model)));
         %% Uknown
+    case 'json'
+        model2JSON(model,fileName);
+        %% Uknown
     otherwise
         error('Unknown file format');
 end

diff --git a/src/dataIntegration/XomicsToModel/XomicsToModel.m b/src/dataIntegration/XomicsToModel/XomicsToModel.m
@@ -43,7 +43,6 @@
 %
 %   * .activeReactions -cell array of reaction identifiers know to be active based on bibliomic data (Default: empty).
 %   * .inactiveReactions - cell array of reaction identifiers know to be inactive based on bibliomic data (Default: empty).
-i
 %   * .coupledRxns -Table containing information about the coupled reactions. This includes the coupled reaction identifier, the
 %                    list of coupled reactions, the coefficients of those reactions, the constraint, the sense or the directionality of the constraint,
 %                    and the reference (Default: empty).
@@ -222,7 +221,7 @@
 %
 % Requires The COBRA Toolbox and a linear optimisation solver (e.g. Gurobi) to be installed
 %
-% 2022 German Preciat, Agnieszka Wegrzyn, Ronan Fleming
+% 2023 German Preciat, Agnieszka Wegrzyn, Xi Luo, Ronan Fleming
 
 model = genericModel;
 
@@ -1030,7 +1029,11 @@
 if ~any(activeModelGeneBool)
     activeEntrezGeneID = [];
 else
-    activeEntrezGeneID = model.genes(activeModelGeneBool);
+    try
+        activeEntrezGeneID = model.genes(activeModelGeneBool);
+    catch
+        activeEntrezGeneID = model.genes(find(activeModelGeneBool));
+    end
 end
 
 % Active genes from manual curation
@@ -1963,6 +1966,7 @@
     end
 
     if 0 && param.debug && param.findThermoConsistentFluxSubset
+
         % Identify the flux consistent set
         paramFluxConsistency.epsilon = param.fluxEpsilon;
         paramFluxConsistency.method = param.fluxCCmethod;
@@ -2214,4 +2218,4 @@
         fprintf('%s', 'XomicsToModel run is complete at:')
         fprintf('%s\n', datetime)
     end
-end
+end
+1 −1		code/prepare/Polytope.m
+1 −1		code/solver/compile_solver.m
+1 −1		coverage/coverage_test.m