void CChemEq::addElement(CCopasiVector < CChemEqElement > & structure,
const CChemEqElement & element,
CChemEq::MetaboliteRole role)
{
unsigned C_INT32 i;
std::string key = element.getMetaboliteKey();
if (key == "")
return; // don�t add empty element
for (i = 0; i < structure.size(); i++)
if (key == structure[i]->getMetaboliteKey())
break;
if (i >= structure.size())
{
CChemEqElement * Element = new CChemEqElement(element);
if (role == CChemEq::SUBSTRATE)
Element->setMultiplicity(- Element->getMultiplicity());
structure.add(Element, true);
}
else if (role == CChemEq::SUBSTRATE)
structure[i]->addToMultiplicity(- element.getMultiplicity());
else
structure[i]->addToMultiplicity(element.getMultiplicity());
}
// calculates the clustering coefficient of gene g
C_FLOAT64 cluster(CCopasiVector < CGene > &gene, C_INT32 g)
{
C_INT32 i, j, k, m, n, edges;
CCopasiVector <CGene> neighbors;
CGene *ptG;
// create the neighborhood
n = gene[g]->getModifierNumber();
if (n <= 1)
return 0.0;
for (i = 0; i < n; i++)
neighbors.add(gene[g]->getModifier(i));
// count the number of links within the neighborhood
for (edges = i = 0; i < n; i++)
{// one neighbor at a time
m = neighbors[i]->getModifierNumber();
for (j = 0; j < m; j++)
{// one modifier at a time
ptG = neighbors[i]->getModifier(j);
// increment if this modifier is in the neighborhood
// but it is not itself
if (ptG != neighbors[i])
for (k = 0; k < n; k++)
if (ptG == neighbors[k])
edges++;
}
}
return (C_FLOAT64) edges / (C_FLOAT64) (n*(n - 1));
}
void WriteDot(char *Title, CCopasiVector < CGene > &gene)
{
unsigned C_INT32 i;
C_INT32 j;
char strtmp[1024];
ofstream fout;
sprintf(strtmp, "%s.dot", Title);
fout.open(strtmp, ios::out);
// dot file header
fout << "digraph \"" << Title << "\" {\n";
fout << "\tgraph\n\t[\n";
fout << "\t\tcenter=\"true\"\n";
fout << "\t\toverlap=\"false\"\n";
fout << "\t\tDamping=0.999\n";
fout << "\t\tfontname=\"Helvetica\"\n";
fout << "\t\tmaxiter=1000000\n";
fout << "\t\tsplines=\"true\"\n";
fout << "\t\tsep=0.8\n";
fout << "\t\tepsilon=0.0000001\n";
fout << "\t\tlabel=\"" << Title << "\"\n";
fout << "\t\tratio=\"auto\"\n";
fout << "\t]\n\n";
fout << "\tnode\n\t[\n";
fout << "\t\tfontsize=9\n";
fout << "\t\tfontname=\"Helvetica-bold\"\n";
fout << "\t\tshape=\"circle\"\n";
fout << "\t\tstyle=\"bold\"\n";
fout << "\t]\n\n";
fout << "\tedge\n\t[\n";
fout << "\t\tfontsize=9\n";
fout << "\t\tfontname=\"Helvetica\"\n";
fout << "\t\tcolor=\"blue\"\n";
fout << "\t\tarrowhead=\"normal\"\n";
if (gene.size() < 15)
fout << "\t\tstyle=\"bold\"\n";
fout << "\t\tlen=2.5\n";
fout << "\t]\n\n";
// graph description
for (i = 0; i < gene.size(); i++)
for (j = 0; j < gene[i]->getModifierNumber(); j++)
{
fout << "\t" << gene[i]->getModifier(j)->getName() << " -> " << gene[i]->getName() << endl;
if (gene[i]->getModifierType(j) == 0)
fout << "\t\t[arrowhead=\"tee\"\n\t\tcolor=\"red\"]" << endl;
}
fout << "\n}";
fout.close();
}
void MakeGeneNetwork(C_INT32 n,
C_INT32 k,
C_FLOAT64 p,
C_FLOAT64 r,
C_FLOAT64 coopval,
C_FLOAT64 rateval,
C_FLOAT64 constval,
CCopasiVector < CGene > &gene,
char *comments)
{
C_INT32 i, j, l, l2, modf, links, links2;
char gn[1024];
// links accounts for the number of links per gene
links = k / n;
links2 = links / 2;
// create and name genes
gene.resize(n);
for (i = 0; i < n; i++)
{
sprintf(gn, "G%ld", i + 1);
gene[i]->setName(gn);
}
// create a regular 1-dimensional grid
for (i = 0; i < n; i++) // each gene
{
for (j = 1; j <= links2; j++) // each link (one link each side)
{
l = i + j;
if (l >= n)
l %= n;
l2 = i - j;
if (l2 < 0)
l2 = n + l2;
// add the two links, i->l i->l2
if (dr250() < p)
modf = 1;
else
modf = 0;
gene[i]->addModifier(gene[l], l, modf, constval, coopval);
if (dr250() < p)
modf = 1;
else
modf = 0;
gene[i]->addModifier(gene[l2], l2, modf, constval, coopval);
}
gene[i]->setRate(rateval);
gene[i]->setDegradationRate(rateval);
}
sprintf(comments, "Model of a gene network on a regular 1D grid (ring)\nwith %ld genes and %ld input connections each.\n\nCreated automatically by the A-Biochem system", n, 2*links2);
}
void SliderDialog::removeSlider()
{
if (mpCurrSlider)
{
assert(CCopasiRootContainer::getDatamodelList()->size() > 0);
CCopasiVector<CSlider>* pSliderList = (*CCopasiRootContainer::getDatamodelList())[0]->getGUI()->getSliderList();
size_t i, maxCount = pSliderList->size();
for (i = 0; i < maxCount; ++i)
{
CSlider* pTmpSlider = (*pSliderList)[i];
if (pTmpSlider == mpCurrSlider->getCSlider())
{
pSliderList->remove(i);
break;
}
}
deleteSlider(mpCurrSlider);
mpCurrSlider = NULL;
}
}
void WriteDistri(char *Title, CCopasiVector < CGene > &gene)
{
C_INT32 i, size;
char strtmp[1024];
ofstream fout;
vector <C_INT32> idegree, odegree, tdegree;
size = gene.size();
// create a vector to count all the degrees
// there are N or less degrees
// (this could be more efficient on memory done in a different way...)
idegree.resize(size + 1);
odegree.resize(size + 1);
tdegree.resize(2*(size + 1));
for (i = 0; i <= size; i++)
{
idegree[i] = 0;
odegree[i] = 0;
tdegree[i] = 0;
}
for (i = size; i < 2*(size + 1); i++)
tdegree[i] = 0;
// count over all genes
for (i = 0; i < size; i++)
{
idegree[gene[i]->getInDegree()]++;
odegree[gene[i]->getOutDegree()]++;
tdegree[gene[i]->getTotalDegree()]++;
}
// now write the data in a gnuplot file
sprintf(strtmp, "%s.distri.plt", Title);
fout.open(strtmp, ios::out);
// fout << "set data style linespoints" << endl;
fout << "set linestyle 1 lw 2\nset linestyle 2 lw 2\nset linestyle 3 lw 2" << endl;
fout << "set title \'degree distribution for " << Title << "'" << endl;
fout << "set xlabel \'# of links\'" << endl;
fout << "set ylabel \'comulative distribution\'" << endl;
fout << "plot \'-\' t \'in-degree\' ls 1, \'-\' t \'out-degree\' ls 2, \'-\' t \'all-degree\' ls 3" << endl;
for (i = 0; i <= size; i++)
fout << i << "\t" << idegree[i] << endl;
fout << "e" << endl;
for (i = 0; i <= size; i++)
fout << i << "\t" << odegree[i] << endl;
fout << "e" << endl;
for (i = 0; i <= size; i++)
fout << i << "\t" << tdegree[i] << endl;
fout << "e" << endl;
fout.close();
}
void SliderDialog::removeSlider()
{
if (mpCurrSlider)
{
CCopasiDataModel * pDataModel = mpCurrSlider->getCSlider()->getObjectDataModel();
assert(pDataModel != NULL);
CCopasiVector<CSlider>* pSliderList = pDataModel->getGUI()->getSliderList();
size_t i, maxCount = pSliderList->size();
for (i = 0; i < maxCount; ++i)
{
CSlider* pTmpSlider = &pSliderList->operator[](i);
if (pTmpSlider == mpCurrSlider->getCSlider())
{
pSliderList->remove(i);
break;
}
}
deleteSlider(mpCurrSlider);
mpCurrSlider = NULL;
}
}
// calculates the single source shortest-path for gene g
// it also stores all the shortest distances in the vector dist
C_INT32 SSSP(CCopasiVector < CGene > &gene, C_INT32 g, C_INT32 *dist)
{
C_INT32 i, j, n, u, cnt, depth;
vector < C_INT32 > *distList, *nextList, nbrs;
n = gene.size();
// set all distances to the maximum
for (i = 0; i < n; i++)
dist[i] = n;
// check if valid gene index
if (g >= n)
return n;
// distance to self is zero
dist[g] = 0;
cnt = 0;
distList = new vector < C_INT32 >;
distList->push_back(g);
for (depth = 1; distList->size() > 0; depth++)
{
nextList = new vector < C_INT32 >;
for (i = 0; i < distList->size(); i++)
{
// fill nbrs with the neighbors
nbrs.clear();
for (j = 0; j < gene[(*distList)[i]]->getModifierNumber(); j++)
nbrs.push_back(gene[(*distList)[i]]->getModifierIndex(j));
for (j = 0; j < nbrs.size(); j++)
{
u = nbrs[j];
if (dist[u] == n)
{
cnt++;
dist[u] = depth;
nextList->push_back(u);
}
}
}
delete distList;
distList = nextList;
}
delete nextList;
return cnt == n - 1 ? depth - 2 : n;
}
void WritePajek(char *Title, CCopasiVector < CGene > &gene)
{
C_INT32 i, j, l, size;
char strtmp[1024];
ofstream fout;
size = gene.size();
sprintf(strtmp, "%s.net", Title);
fout.open(strtmp, ios::out & ~ios::binary);
fout << "*Network " << Title << endl;
fout << "*Vertices " << size << endl;
for (i = 0; i < size; i++)
fout << i + 1 << " \"" << gene[i]->getName() << "\"" << endl;
fout << "*Arcs" << endl;
for (i = 0; i < size; i++)
for (j = 0; j < gene[i]->getModifierNumber(); j++)
for (l = 0; l < size; l++)
if (gene[i]->getModifier(j) == gene[l])
fout << l + 1 << " " << i + 1 << " 1 c " << (gene[i]->getModifierType(j) == 0 ? "Red" : "Blue") << endl;
fout.close();
}
bool CCopasiSimpleSelectionTree::isMetaboliteNameUnique(const std::string & name, const CCopasiVector<CMetab> & metabolites)
{
bool unique = true;
bool found = false;
unsigned int counter;
for (counter = 0; counter < metabolites.size(); ++counter)
{
const std::string& thisName = metabolites[counter]->getObjectName();
if (name == thisName)
{
if (found)
{
unique = false;
break;
}
found = true;
}
}
return unique;
}
C_INT main(C_INT argc, char *argv[])
{
C_INT32 n, k, i, j, tot, seed, pos, specific;
CCopasiVector < CGene > GeneList;
string prefix, fdb;
CModel model;
C_FLOAT64 positive, interact, rewiring, coopval, rateval, constval;
char NetTitle[512], comments[2048];
CGeneModifier *test;
test = new CGeneModifier;
// parse command line
try
{
clo::parser cmdl;
cmdl.parse(argc, argv);
// get options from command line parser
const clo::options &options = cmdl.get_options();
if (options.version)
{
std::cout << argv[0] << versionString << endl;
std::cout << "Written by Pedro Mendes" << endl;
std::cout << "Copyright (C) 2003 Virginia Polytechnic Institute and State University" << endl;
return 0;
}
if (options.specific_activation) specific = 1; else specific = 0;
seed = options.seed;
tot = options.total;
n = options.genes;
k = options.inputs;
prefix = options.prefix;
positive = options.positive;
interact = options.interaction;
rewiring = options.rewire;
coopval = options.coop;
rateval = options.rates;
constval = options.constants;
}
catch (clo::autoexcept &e)
{
switch (e.get_autothrow_id())
{
case clo::autothrow_help:
std::cout << "Usage: " << argv[0] << " [options]" << endl;
std::cout << e.what();
return 0;
}
}
catch (std::exception &e)
{
std::cout << argv[0] << ": " << e.what() << endl;
return 1;
}
Copasi = new CGlobals;
// set the Function DB filename to be in the directory of the executable
fdb = argv[0];
pos = fdb.find_last_of('/');
if (pos == string::npos)
pos = fdb.find_last_of('\\');
if (pos != string::npos)
{
fdb = fdb.substr(0, pos + 1) + "FunctionDB.gps";
Copasi->pFunctionDB->setFilename(fdb);
}
Copasi->setArguments(argc, argv);
if (seed)
r250_init(seed);
else
r250_init(((int) time(NULL)));
// else r250_init(((int) time(NULL)) | 12783579);
// generate tot networks of n genes with k random outputs each
for (i = 0; i < tot; i++)
{
// build the gene network
MakeGeneNetwork(n, k, positive, rewiring, coopval, rateval, constval, GeneList, comments);
// sort out the modifiers
for (j = 0; j < n; j++) GeneList[j]->sortModifiers();
sprintf(NetTitle, "%s%03ld", prefix.data(), i + 1);
// write the command line used to generate this model
WriteCmdLine(Copasi, NetTitle);
// create appropriate kinetic types, only those
// that are really needed for this model
for (j = 0; j < n; j++)
MakeKinType(*Copasi->pFunctionDB, GeneList[j]->getModifierNumber(), GeneList[j]->getPositiveModifiers(), specific);
// calculate several metrics
GraphMetrics(GeneList, k, NetTitle);
// create graph files
WriteDot(NetTitle, GeneList);
WritePajek(NetTitle, GeneList);
// create distribution graph file
WriteDistri(NetTitle, GeneList);
// create COPASI model
MakeModel(NetTitle, comments, GeneList, n, k, model, specific);
// save Gepasi model
WriteGepasi(NetTitle, model);
// save SBML file
WriteSBML(NetTitle, model);
// cleanup model and vectors
model.cleanup();
GeneList.cleanup();
}
delete Copasi;
return 0;
}
void GraphMetrics(CCopasiVector < CGene > &gene, C_INT32 tedges, char *Title)
{
C_INT32 i, j, k, l, n;
CMatrix<C_FLOAT64> Adj;
CEigen *eigen;
C_FLOAT64 ssRes, *eigen_r, *eigen_i;
C_INT32 **dist;
ofstream fout;
char fname[1024];
vector <C_FLOAT64> mpl, tpl, b;
C_FLOAT64 cpl, cl, cll, cbl, acc, density, netcl, tmp, tmp2, netbc;
C_INT32 idcenter, odcenter, clcenter, diameter, id, od, bcenter;
// open the file for statistics
sprintf(fname, "%s.netstat", Title);
fout.open(fname, ios::out & ~ios::binary);
n = gene.size();
density = (C_FLOAT64) tedges / (C_FLOAT64) (n * n);
fout << "number of vertices\t" << n << endl;
fout << "number of arcs\t" << tedges << endl;
fout << "density\t" << density << endl;
#ifdef XXXX
// create the adjacency matrix
Adj.newsize(n, n);
for (i = 0; i < n; i++)
{
for (j = 0; j < n; j++)
Adj[j][i] = 0.0;
for (l = 0; l < gene[i]->getModifierNumber(); l++)
{
for (j = 0; j < n; j++)
if (gene[i]->getModifier(l) == gene[j])
{
Adj[j][i] = 1.0;
// if(gene[i]->getModifierType(l) == 0) Adj[j][i] = -1.0;
// else Adj[j][i] = 1.0;
break;
}
}
}
// calculate its eigenvalues
eigen = new CEigen();
ssRes = 1e-9;
eigen->CalcEigenvalues(ssRes, Adj);
eigen_r = eigen->getEigen_r();
eigen_i = eigen->getEigen_i();
// TODO: linear regression on positive real parts
delete eigen;
#endif
if (n < LARGE_GRAPH)
{
// find the diameter
diameter = 0;
dist = new C_INT32 * [n];
for (i = 0; i < n; i++)
{
dist[i] = new C_INT32[n];
l = SSSP(gene, i, dist[i]);
if (l > diameter)
diameter = l;
}
fout << "diameter\t" << diameter << endl;
// calculate total and mean path lengths
mpl.reserve(n);
for (i = 0; i < n; i++)
{
cpl = 0.0;
for (j = 0; j < n; j++)
cpl += dist[i][j];
cll = (C_FLOAT64) (n - 1) / cpl;
cpl /= n - 1;
mpl.push_back(cpl);
tpl.push_back(cll);
}
}
// find the degree centers
for (id = od = idcenter = odcenter = i = 0; i < n; i++)
{
k = gene[i]->getOutDegree();
if (k > od)
{
odcenter = i;
od = k;
}
k = gene[i]->getInDegree();
if (k > id)
{
idcenter = i;
id = k;
}
}
cl = (C_FLOAT64) id / (C_FLOAT64)(n - 1);
fout << "in-degree center\t" << gene[idcenter]->getName() << endl;
fout << "center's in-degree\t" << id << endl;
fout << "center's relative in-degree\t" << cl << endl;
cl = (C_FLOAT64) od / (C_FLOAT64)(n - 1);
fout << "out-degree center\t" << gene[odcenter]->getName() << endl;
fout << "center's out-degree\t" << od << endl;
fout << "center's relative out-degree\t" << cl << endl;
if (n < LARGE_GRAPH)
{
// find the closeness center, degree centralization,
// and closeness centralization
for (cll = 0.0, clcenter = -1, i = 0; i < n; i++)
{
if (tpl[i] > cll)
{
cll = tpl[i];
clcenter = i;
}
}
for (netcl = 0.0, i = 0; i < n; i++)
netcl += cll - tpl[i];
netcl /= ((C_FLOAT64)n - 1.0) * ((C_FLOAT64)n - 2.0) / (2.0 * (C_FLOAT64)n - 3.0);
fout << "closeness center\t" << gene[clcenter]->getName() << endl;
fout << "center's relative closeness\t" << cll << endl;
fout << "network closeness centralization\t" << netcl << endl;
// calculate betweeness centralities
b.reserve(n);
Betweeness(gene, b);
//
for (cbl = 0.0, bcenter = 0, i = 0; i < n; i++)
{
if (b[i] > cbl)
{
cbl = b[i];
bcenter = i;
}
}
for (netbc = 0.0, i = 0; i < n; i++)
netbc += cbl - b[i];
fout << "betweenness center\t" << gene[bcenter]->getName() << endl;
fout << "center's relative betweenness\t" << cbl << endl;
fout << "network betweenness centralization\t" << netbc << endl;
// calculate characteristic path length
sort(mpl.begin(), mpl.end());
if (n / 2 == 0)
cpl = (mpl[n / 2] + mpl[1 + n / 2]) * 0.5;
else
cpl = mpl[n / 2];
fout << "characteristic path length\t" << cpl << endl;
// calculate clustering coefficients
for (acc = 0.0, i = 0; i < n; i++)
{
tmp = cluster(gene, i);
acc += tmp;
}
acc /= n;
fout << "average clustering coefficient\t" << acc << endl;
}
fout.close();
}
// calculates the betweeness centrality for each gene
// uses the algorithm of Brandes (J Math Sociol 25, 163-177, 2001)
void Betweeness(CCopasiVector < CGene > &gene, vector <C_FLOAT64> &b)
{
C_INT32 i, n, s, tot, v, w;
vector < C_INT32 > sigma, d;
vector < C_FLOAT64 > delta;
stack < C_INT32 > S;
queue < C_INT32 > Q;
typedef list < C_INT32 > INTLIST;
vector < INTLIST* > P;
n = gene.size();
sigma.reserve(n);
delta.reserve(n);
d.reserve(n);
P.reserve(n);
for (i = 0; i < n; i++)
P[i] = new INTLIST;
// reset variables
for (i = 0; i < n; i++)
b[i] = 0.0;
// iterate over all genes
for (s = 0; s < n; s++)
{
// reset variables
while (!S.empty())
S.pop();
while (!Q.empty())
Q.pop();
for (i = 0; i < n; i++)
{
sigma[i] = 0;
d[i] = -1;
if (!P[i]->empty())
P[i]->clear();
}
sigma[s] = 1;
d[s] = 1;
// add s to queue
Q.push(s);
while (! Q.empty())
{
v = Q.front();
Q.pop();
S.push(v);
tot = gene[v]->getModifierNumber();
for (i = 0; i < tot; i++)
{
w = gene[v]->getModifierIndex(i);
// w found for the first time?
if (d[w] < 0)
{
Q.push(w);
d[w] = d[v] + 1;
}
// shortest path to w via v?
if (d[w] == d[v] + 1)
{
sigma[w] += sigma[v];
P[w]->push_back(v);
}
}
}
for (i = 0; i < n; i++)
delta[i] = 0.0;
// S returns vertices in order of non-increasing distance from s
while (!S.empty())
{
w = S.top();
S.pop();
INTLIST::iterator it;
for (it = P[w]->begin(); it != P[w]->end(); it++)
delta[*it] += sigma[*it] * (1 + delta[w]) / sigma[w];
if (w != s)
{
C_FLOAT64 tmp = delta[w];
b[w] += delta[w];
}
}
}
for (i = 0; i < n; i++)
{
// make the measures relative
b[i] /= ((C_FLOAT64)n - 1.0) * ((C_FLOAT64)n - 2.0);
// delete the lists allocated
delete P[i];
}
}
