Description of KEGG2BiGGMets

0001 function [Metkegg,Metbigg] = KEGG2BiGGMets(D,CbModel)
0002 % KEGG2BiGGMets
0003 % is a subfunction of MapAdjuster for displaying compound's
0004 % data on created customized metabolic map by NetDraw. The function
0005 % extracts all metabolite correspondences between KEGG and BiGG based on
0006 % the current COBRA model and data in BiGG API generated by
0007 % AllKEGG2BiGGmet. The structure is very similar to that of used in
0008 % Bigg2Kegg with minor adjustments.
0009 %
0010 % Inputs:
0011 % D: XML data of current COBRA model from BiGG
0012 % CbModel: current COBRA model structure
0013 %
0014 % Outputs:
0015 % Metkegg: All KEGG compounds' IDs
0016 % Metbigg: All BiGG compounds' IDs
0017 
0018 % O. Jamialahmadi
0019 % TMU, Chem. Eng. Dept., Biotech. Group
0020 % July 2016
0021 % -------------------------------------------------------------------------
0022 
0023 Pth1 = which ('Bigg2Kegg.m');
0024 tind = find(Pth1=='\',1,'last');
0025 Pth = Pth1(1:tind-1);
0026 % Reading xml
0027 % Find listOfSpecies which contains metabolite annotations
0028 SpeciesLoci1 = strfind(D{1},'listOfSpecies');
0029 SpeciesLoci = find(~cellfun('isempty', SpeciesLoci1));
0030 % Discard other data in the model:
0031 SpeciesData = D{1}(SpeciesLoci(1):SpeciesLoci(2));
0032 clear D
0033 % Find each metabolite-------------------------------------------------
0034 SpeciesStart1 = strfind(SpeciesData,'<species');
0035 SpeciesStart = find(~cellfun('isempty', SpeciesStart1));
0036 SpeciesEnd1 = strfind(SpeciesData,'</species>');
0037 SpeciesEnd = find(~cellfun('isempty', SpeciesEnd1));
0038 if numel(SpeciesStart) ~= numel(SpeciesEnd)
0039     msgbox('SBML format error:Species have wrong format!','Error',...
0040         'error');
0041     return
0042 end
0043 % ---------------------------------------------------------------------
0044 % Search for KEGG compounds for each BiGG metabolite. Note that each BiGG
0045 % metabolite may have more than one equivalent KEGG compund, and although
0046 % HMR and RECON2 (Thiele lab.) do not provide more than one KEGG compound
0047 % for each metabolite, here, we followed the format of BiGG models (i.e.
0048 % several equivalent KEGG compounds for a specific metabolite).
0049 Metkegg = {''}; % KEGG compounds will fill this cell!
0050 for count = 1:numel(SpeciesStart)
0051     SpeciesTempStr = SpeciesData(SpeciesStart(count):SpeciesEnd(count));
0052     SpeciesTempStr = [SpeciesTempStr{:}];
0053     KeggCpdLoci = regexp(SpeciesTempStr,...
0054         '(?<=http://identifiers.org/kegg.compound/)[CG]\d{5}','match');
0055     if isempty(KeggCpdLoci)
0056         Metkegg{count} = '';
0057         continue
0058     end
0059     for count1 = 1:numel(KeggCpdLoci)
0060         if strcmp(KeggCpdLoci{count1},'C00001') % H2O
0061             Metkegg{count}= {'C00001'};
0062             break
0063         elseif strcmp(KeggCpdLoci{count1},'C00027') % H2O2
0064             Metkegg{count}= {'C00027'};
0065             break
0066         else
0067             Metkegg{count}{count1}= KeggCpdLoci{count1};
0068         end
0069     end
0070 end
0071 % Model modification: some metabolites have wrong/absent KEGG annotations.
0072 % Therefore, on the basis of modified KEGG compound IDs for RECON1.xml,
0073 % these metabolites will be added to Metkegg separately.
0074 D = CbModel;
0075 ModelName = D.description;
0076 tind1 = find(ModelName=='\',1,'last');
0077 ModelName1 = ModelName(tind1+1:end-4);
0078 if ~strcmp(ModelName1,'RECON1')
0079     MetAbr = CbModel.mets;
0080     Metkegg = ModelMoidfy (Metkegg,MetAbr);
0081 end
0082 
0083 load(fullfile(Pth,'AllKEGG2BiGGmet.mat'))
0084 Metkegg1 = AllKEGG2BiGGmet.Metkegg; Metkegg1 = [Metkegg1{:}];
0085 Metbigg1 = AllKEGG2BiGGmet.Metbigg; Metbigg1 = [Metbigg1{:}];
0086 MetAbr = strrep(MetAbr,'[','_'); MetAbr = strrep(MetAbr,']','');
0087 Metkegg2 = ({}); ct = 1; MetAbr1 = ({});
0088 for i1 = 1:numel(Metkegg)
0089     if isempty(Metkegg{i1})
0090         continue
0091     end
0092     for i2 = 1:numel(Metkegg{i1})
0093         Metkegg2{ct} = Metkegg{i1}{i2};
0094         MetAbr1{ct} = MetAbr{i1};
0095         ct = ct + 1;
0096     end
0097 end
0098 clear Metkegg 
0099 Metbigg = [Metbigg1,MetAbr1]; Metkegg = [Metkegg1,Metkegg2];
0100 Metbigg = regexp(Metbigg,'\w*(?=_\w{1}\>)','match'); % Remove compartments
0101 Metbigg = [Metbigg{:}];
0102 Mixdat = strcat(Metbigg,Metkegg);
0103 [~, N1] = unique(Mixdat);
0104 Metbigg = Metbigg(N1);
0105 Metkegg = Metkegg(N1);
0106 
0107 % Consensus Metkegg: Metkegg1 from BiGG API and Metkegg from COBRA models
0108 % are mixed to generate the most complete set of KEGG metabolites for a
0109 % SBML model.
0110 % for cp1 = 1:numel(Metkegg1)
0111 %     if isempty(Metkegg1{cp1}) || (numel(Metkegg1{cp1}) ~= numel(Metkegg{cp1}))
0112 %         for cp2 = 1:numel(Metkegg{cp1})
0113 %             Metkegg1{cp1}{cp2} = Metkegg{cp1}{cp2};
0114 %         end
0115 %     end
0116 % end
0117 
0118 
0119 function Metkegg = ModelMoidfy (Metkegg,MetAbr)
0120 % Read Metkegg data for RECON1.xml
0121 Mdat = load('RECON1Metkegg.mat');
0122 RECON1k = Mdat.RECON1meta.K;
0123 RECON1m = Mdat.RECON1meta.B;
0124 [Fx1,Fx2] = ismember(RECON1m,MetAbr);
0125 Fx2(Fx2==0) = [];
0126 % Metkegg(Fx2) = RECON1k(Fx1);
0127 Fx3 = find(Fx1);
0128 for i1 = 1:numel(Fx2)
0129     if numel(Metkegg{Fx2(i1)}) > numel(RECON1k{Fx3(i1)})
0130        Tempmet = Metkegg{Fx2(i1)};
0131        whrc = ismember(Tempmet,RECON1k{Fx3(i1)});
0132        if isempty(whrc)
0133         Metkegg{Fx2(i1)} = RECON1k(Fx3(i1));
0134         end
0135     else
0136         Metkegg{Fx2(i1)} = RECON1k{Fx3(i1)};
0137     end
0138 end
0139
KEGG2BiGGMets

PURPOSE

SYNOPSIS

DESCRIPTION

CROSS-REFERENCE INFORMATION

SUBFUNCTIONS

SOURCE CODE
