convert_EX_to_diet.m

function [model] = convert_EX_to_diet(model)
%
% convert_EX_to_diet takes a model with typical exchange reactions 
% (i.e., 'EX_') and translates them to uptake (diet) and secretion (exit) 
% reactions.
%
% USAGE:
%
% [model] = convert_EX_to_diet(model)
%
% INPUTS:
%    model:          COBRA model structure with minimal fields:
%                      * .S
%                      * .c
%                      * .ub
%                      * .lb
%                      * .mets  
%                      * .rxns   
%
%Outputs
%   model: Returns the input model with new uptake ('Diet_') and secretion 
%          ('Exit_') reactions
%
%Authors: Bronson R. Weston 2022


exRxns=[];
for i=1:length(model.rxns)
    tmp=strsplit(model.rxns{i},'EX_');
    if isempty(tmp{1})
        exRxns=[exRxns,i];
    end
end

% Use S matrix to determine which metabolites are in exchange reactions
S=model.S;
exMets=[];
for i=1:length(exRxns)
    x=find(S(:,exRxns(i))~=0).';
    exMets=[exMets,x];
end
exMets=unique(exMets);

% Create environmental metabolites for each exchange metabolite
if any(contains(model.rxns,'Diet_')) 
    error('Model already contains dietary reactions')
end
envMets=model.mets(exMets);

% Create new exchange reactions importing metabolites in dietary reactions.
% Set ub and lb according to lb of original exchange reaction
ubs=model.ub(exRxns);
lbs=model.lb(exRxns);
model.rxns(exRxns)=strcat('Diet_',model.rxns(exRxns));
model.lb(exRxns)=model.lb(exRxns);
model.ub(exRxns)=model.lb(exRxns);

%Add an exit reaction from the environment for each exchange metabolite,
%with a lb of 0 and a ub of 100000
exitRxns= strcat('Exit_EX_',envMets);
model=addMultipleReactions(model,exitRxns,envMets,-1*eye(length(envMets)),'lb',zeros(1,length(exMets)), 'ub', 100000*ones(1,length(exMets)));