//constructor (originally included in CADDSuite)
UCK::UCK(const Molecule& mol, bool ignore_hydrogens, Size d)
: depth_(d),
weight_(0.0)
{
id_ = mol.getName();
id_.trim();
depth_ = d;
ignore_hydrogens_ = ignore_hydrogens;
makeUCK(mol);
}
//constructor
UCK::UCK(const Molecule& mol, Size d)
: depth_(d),
weight_(0.0),
ignore_hydrogens_(false)
{
id_ = mol.getName();
id_.trim();
depth_ = d;
makeUCK(mol);
}
int main(int argc, char* argv[])
{
CommandlineParser parpars("RMSDCalculator", "calculate RMSD between ligand poses", VERSION, String(__DATE__), "Analysis");
parpars.registerMandatoryInputFile("i", "input molecule file");
parpars.registerMandatoryInputFile("org", "molecule file containing the original ('true') poses");
parpars.registerOptionalOutputFile("o", "output molecule file");
parpars.registerFlag("quiet", "by quiet, i.e. do not print progress information");
String man = "This tool calculates the RMSD between different conformations of the same molecule.\n\nThis tool can be used to evaluate the differences between ligand poses taken from co-crystal structures, e.g. generated by a docking run.\nNote:Molecules may be sorted differently in the two input files; a topology hashkey will be used to match molecules to each other.\n\nOutput of this tool is a molecule file which will for each molecule contain a property-tag 'RMSD' holding the calculated RMSD value.";
parpars.setToolManual(man);
parpars.setSupportedFormats("i",MolFileFactory::getSupportedFormats());
parpars.setSupportedFormats("org",MolFileFactory::getSupportedFormats());
parpars.setSupportedFormats("o","mol2,sdf,drf");
parpars.parse(argc, argv);
// Retrieve coordinates of original poses
GenericMolFile* original = MolFileFactory::open(parpars.get("org"));
HashMap<String, list<Vector3> > original_poses;
for (Molecule* mol = original->read(); mol; delete mol, mol = original->read())
{
String topology_hash;
FlexibleMolecule::generateTopologyHash(mol, topology_hash, true);
if (original_poses.find(topology_hash) != original_poses.end())
{
Log<<"[Warning:] more than one 'original' conformation for a molecule detected. Will use only the first conformation and ignore all other."<<endl;
}
else
{
list<Vector3> l;
HashMap<String, list<Vector3> >::iterator map_it = original_poses.insert(make_pair(topology_hash, l)).first;
for (AtomConstIterator it = mol->beginAtom(); +it; it++)
{
if (it->getElement().getSymbol() != "H")
{
map_it->second.push_back(it->getPosition());
}
}
}
}
delete original;
// Retrieve coordinates of input poses and calculate RMSDs
GenericMolFile* input = MolFileFactory::open(parpars.get("i"));
GenericMolFile* output = 0;
String filename = parpars.get("o");
if (filename != CommandlineParser::NOT_FOUND)
{
output = MolFileFactory::open(filename, ios::out, input);
}
double average_RMSD = 0;
int no_mols = 0;
int no_valid_rmsds = 0;
bool quiet = (parpars.get("quiet")!=CommandlineParser::NOT_FOUND);
for (Molecule* mol = input->read(); mol; delete mol, mol = input->read())
{
no_mols++;
String topology_hash;
FlexibleMolecule::generateTopologyHash(mol, topology_hash, true);
HashMap<String, list<Vector3> >::iterator map_it = original_poses.find(topology_hash);
if (map_it == original_poses.end())
{
Log<<"[Warning:] no original pose for molecule '"<<mol->getName()<<"' found, its RMSD can thus not be computed."<<endl;
mol->setProperty("RMSD", "N/A");
}
else
{
double RMSD = 0;
list<Vector3>::iterator list_it = map_it->second.begin();
int no_heavy_atoms = 0;
AtomConstIterator it = mol->beginAtom();
for (; +it ; it++)
{
if (it->getElement().getSymbol() != "H" && list_it != map_it->second.end())
{
RMSD += pow(it->getPosition().getDistance(*list_it), 2);
no_heavy_atoms++;
list_it++;
}
}
if (it != mol->endAtom() || list_it != map_it->second.end())
{
Log.error()<<"[Error:] Number of heavy atoms of input pose do not match number of heavy atoms of original pose!!"<<endl;
return 1;
}
RMSD = sqrt(RMSD/no_heavy_atoms);
mol->setProperty("RMSD", RMSD);
average_RMSD += RMSD;
no_valid_rmsds++;
if (!quiet) Log << "RMSD for molecule "<<no_mols<<", '"<<mol->getName()<<"' = "<<RMSD<<endl;
}
if (output) *output << *mol;
}
average_RMSD /= no_valid_rmsds;
Log <<endl<<"average RMSD = "<<average_RMSD<<endl<<endl;
delete input;
delete output;
return 0;
}
void mergeDRFiles(vector<String>& names, string& output_file, Size& best_k, string& e_property, double& score_cutoff, double& score_cuton)
{
DockResultFile* output = new DockResultFile(output_file, ios::out);
bool sort_by_scores = 1;
if (e_property == "") sort_by_scores = 0;
vector<Result*> new_results;
/// First of all, copy Result data
map<Result::Method, Result*> result_map;
for (Size file = 0; file < names.size(); file++)
{
DockResultFile* input = new DockResultFile(names[file]);
const vector<Result*>* results = input->getResults();
for (Size i = 0; i < results->size(); i++)
{
map<Result::Method, Result*>::iterator it = result_map.find((*results)[i]->getMethod());
if (it == result_map.end())
{
Result* result_copy = new Result(*(*results)[i]);
if (!sort_by_scores) output->addResult(result_copy);
else new_results.push_back(result_copy);
result_map.insert(make_pair(result_copy->getMethod(), result_copy));
}
else
{
*it->second += *(*results)[i];
}
}
input->close();
delete input;
}
if (e_property != "")
{
e_property = "score_"+new_results.back()->getMethodString();
}
/// If no sorting is desired, iterate over all input-files and write each input-molecules to output-file
if (!sort_by_scores)
{
output->disableAutomaticResultCreation();
for (Size file = 0; file < names.size(); file++)
{
GenericMolFile* input = MolFileFactory::open(names[file]);
int mol_no = 0;
for (Molecule* mol = input->read(); mol; mol = input->read(), mol_no++)
{
*output << *mol;
delete mol;
Log.level(20) << "\r" << names[file] << " : " << mol_no+1;
Log.flush();
}
Log.level(20)<<endl;
Log.flush();
input->close();
delete input;
}
}
/// If sorting is desired, iterate over all input-files and save each input-molecules to a map.
/// Then write all FlexibleMolecules in this map to the output file and adapt the Result objects.
else
{
multimap < double, FlexibleMolecule* > compounds; // map containing score and conformation-ID
set < String > IDs; // IDs of the base-conformations
for (Size file = 0; file < names.size(); file++)
{
DockResultFile* input = new DockResultFile(names[file]);
int mol_no = 0;
for (Molecule* mol = input->read(); mol; mol = input->read(), mol_no++)
{
if (!mol->hasProperty(e_property))
{
Log.level(10) << "Compound " << mol->getName() << " in file " << names[file] << " has no score property. Skipping this compound." << endl;
for (Size i = 0; i < new_results.size(); i++)
{
new_results[i]->erase(input->getCurrentLigand());
}
delete mol;
continue;
}
double score = ((String)mol->getProperty(e_property).toString()).toFloat();
if (score > score_cutoff || score < score_cuton)
{
for (Size i = 0; i < new_results.size(); i++)
{
new_results[i]->erase(input->getCurrentLigand());
}
delete mol;
continue;
}
if ((compounds.size() < best_k || score < compounds.rbegin()->first))
{
FlexibleMolecule* flexmol_copy = new FlexibleMolecule(*input->getCurrentLigand());
compounds.insert(make_pair(score, flexmol_copy));
IDs.insert(flexmol_copy->getId());
if (compounds.size() > best_k)
{
for (Size i = 0; i < new_results.size(); i++)
{
new_results[i]->erase(compounds.rbegin()->second);
}
IDs.erase(compounds.rbegin()->second->getId());
delete compounds.rbegin()->second;
multimap<double, FlexibleMolecule*>::iterator it = compounds.end();
it--;
compounds.erase(it);
}
}
else
{
for (Size i = 0; i < new_results.size(); i++)
{
new_results[i]->erase(input->getCurrentLigand());
}
delete mol;
}
Log.level(20) <<"\r"<<names[file]<<" : "<<mol_no+1<<flush;
}
Log.level(20)<<endl;
input->close();
delete input;
}
if (compounds.size() < best_k)
{
Log.level(20)<<"found "<<compounds.size()<<" compounds matching the given criteria."<<endl;
}
list<String> score_list;
for (multimap < double, FlexibleMolecule* > ::iterator it = compounds.begin();
it!=compounds.end(); it++)
{
output->writeLigand(it->second);
score_list.push_back(it->second->getId());
delete it->second;
}
// Remove those ligands from results for which no final result is available (e.g. due to missing atom parameters)
vector < String > import_IDs = *new_results[0]->getInputConformations();
for (Size i = 0; i < import_IDs.size(); i++)
{
if (IDs.find(import_IDs[i]) == IDs.end())
{
vector<vector<Result::ResultData> > data_list;
for (Size k = 0; k < new_results.size(); k++)
{
if (k == 0)
{
data_list.push_back(*new_results[k]->get(import_IDs[i]));
}
vector<vector<Result::ResultData> > new_data_list;
for (Size j = 0; j < data_list.size(); j++)
{
for (Size l = 0; l < data_list[j].size(); l++)
{
String ID = data_list[j][l].getLigandConformationId();
new_data_list.push_back(*new_results[k]->get(ID));
new_results[k]->erase(ID);
}
}
data_list = new_data_list;
}
}
}
for (Size i = 0; i < new_results.size(); i++)
{
list<String> new_list;
new_results[i]->sort(score_list, new_list);
score_list = new_list;
}
output->writeResults(new_results);
}
output->close();
delete output;
}
void sortMolecules(vector<String>& names, string& output_file, Size& best_k, string& e_property, double& score_cutoff, double& score_cuton)
{
multimap<double, Molecule*> compounds;
for (Size file = 0; file < names.size(); file++)
{
GenericMolFile* input = MolFileFactory::open(names[file]);
int mol_no = 0;
for (Molecule* mol = input->read(); mol; mol = input->read(), mol_no++)
{
if (!mol->hasProperty(e_property))
{
Log.level(10) << "Compound " << mol->getName() << " in file " << names[file] << " has no score property. Skipping this compound." << endl;
delete mol;
continue;
}
double score = ((String)mol->getProperty(e_property).toString()).toFloat();
if (score > score_cutoff || score < score_cuton)
{
delete mol;
continue;
}
if ((compounds.size() < best_k || score < compounds.rbegin()->first))
{
compounds.insert(make_pair(score, mol));
if (compounds.size() > best_k)
{
delete compounds.rbegin()->second;
multimap<double, Molecule*>::iterator it = compounds.end();
it--;
compounds.erase(it);
}
}
else
{
delete mol;
}
Log.level(20) << "\r" << names[file] << " : " << mol_no+1 << flush;
}
Log.level(20) << endl;
Log.flush();
input->close();
delete input;
}
if (compounds.size() < best_k)
{
Log.level(20) << "found " << compounds.size() << " compounds matching the given criteria." << endl;
}
GenericMolFile* output = MolFileFactory::open(output_file, ios::out, "mol2.gz");
for (multimap < double, Molecule* > ::iterator it = compounds.begin();
it!=compounds.end(); it++)
{
*output << *it->second;
delete it->second;
}
output->close();
delete output;
}
CHECK(Molecule() throw())
b = new Molecule;
TEST_NOT_EQUAL(b, 0)
RESULT
CHECK(~Molecule() throw())
delete b;
RESULT
CHECK(Molecule(const Molecule& molecule, bool deep = true) throw())
Atom a1;
Molecule m("a"), m2;
m.append(a1);
m2 = Molecule(m);
TEST_EQUAL(m2.getName(), "a")
TEST_EQUAL(m2.countAtoms(), 1)
RESULT
CHECK(Molecule(const String& name) throw())
Molecule m("a");
TEST_EQUAL(m.getName(), "a")
Molecule m2("");
TEST_EQUAL(m2.getName(), "")
RESULT
CHECK([EXTRA] clear())
System s;
Molecule m("a");
Atom a1;
m.append(a1);
bool KCFFile::write(const Molecule& molecule)
throw(File::CannotWrite)
{
if (!isOpen() || getOpenMode() != std::ios::out)
{
throw File::CannotWrite(__FILE__, __LINE__, name_);
}
// An alias for simplicity's sake...
std::ostream& os(getFileStream());
// Write ENTRY block
// number of blanks???? properties are not read, written??? Which ones are there?
os << ENTRY_TAG << " " << molecule.getName() << std::endl;
static char buffer[BALL_MAX_LINE_LENGTH];
// Write NODE block
// How to create the KEGG atom types? How many blanks?
// This is not specified in the KCF format description, so we use what we can
// deduce from example files.
// First line gets the NODE tag
os << NODE_TAG << " " << molecule.countAtoms() << "\n";
Size count = 1;
AtomConstIterator ai(molecule.beginAtom());
std::map<const Atom*, Position> atom_to_index;
for (; +ai; ++ai, ++count)
{
// Write the atom line.
// Blanks????
String type = ai->getTypeName();
String comment;
// Make sure the type is in the set of KEGG types????
// Blanks?
sprintf(buffer, " %d %s %s %6.4f %6.4f %s\n",
count, type.c_str(), ai->getElement().getSymbol().c_str(),
ai->getPosition().x, ai->getPosition().y, comment.c_str());
os << buffer;
// Remember the index of the current atom to map atom
// pointers back to indices for the EDGE section.
atom_to_index[&*ai] = count;
}
// Write EDGE block. Walk over all bonds to do so.
// Blanks????
os << "EDGE " << molecule.countBonds() << "\n";
count = 1;
for (ai = molecule.beginAtom(); +ai; ++ai)
{
for (Atom::BondConstIterator bi(ai->beginBond()); +bi; ++bi)
{
Position index1 = atom_to_index[bi->getFirstAtom()];
Position index2 = atom_to_index[bi->getSecondAtom()];
String comment;
// Write every bond just once
if (bi->getFirstAtom() == &*ai)
{
sprintf(buffer, " %4d %4d %4d %1d%s\n",
count, index1, index2, bi->getOrder(), comment.c_str());
os << buffer;
++count;
}
}
}
// Write the DELIMITER block
os << DELIMITER_TAG << std::endl;
return true;
}
void processMultiMoleculeFile(ScoringFunction* scoring_function, StructurePreparer* sp, String par_file, Rescoring* rescoring, bool train, double min_dock_score, String dock_score_label, GenericMolFile* input, GenericMolFile* output, double output_score_threshold, bool ignore_top)
{
list<pair<double, bool> > rescore_list;
double min_rescore = 1e12;
int i = 1;
for (Molecule* mol = input->read(); mol; mol = input->read(), i++)
{
String name = mol->getName();
Log<<"====== Ligand "<<i;
if (name != "") Log<<", "<<name;
Log<<" ============"<<endl;
Log.flush();
if (mol->hasProperty("score_ligcheck"))
{
double score_ligcheck = ((String)mol->getProperty("score_ligcheck").toString()).toDouble();
if (score_ligcheck < 0.95) // 0 = error, 1 = check passed
{
cout<<"Skipping compound because it has been marked as containing errors by LigCheck."<<endl;
delete mol;
continue;
}
}
double score = 0;
try
{
sp->prepare(mol, par_file);
if (!rescoring || train)
{
scoring_function->setLigand(mol);
scoring_function->update();
score = scoring_function->updateScore();
scoring_function->printResult();
}
if (rescoring)
{
if (train)
{
if (score > 1000)
{
Log.level(10)<<"score>1000, thus current compound is not added to training data set."<<endl;
delete mol;
continue;
}
rescoring->addScoreContributions(mol);
}
else score = rescoring->rescore(mol);
}
if (score < output_score_threshold)
{
if (ignore_top)
{
bool keep_dock_score = 0;
if (mol->hasProperty("score"))
{
double docking_score = mol->getProperty(dock_score_label).toString().toDouble();
if (docking_score < min_dock_score)
{
rescore_list.push_back(make_pair(docking_score, false));
keep_dock_score = 1;
}
}
if (!keep_dock_score)
{
rescore_list.push_back(make_pair(score, true));
if (score < min_rescore)
{
min_rescore = score;
}
}
}
else
{
mol->setProperty("re-score", score);
}
}
}
catch (BALL::Exception::GeneralException e)
{
cout<<e.getMessage()<<endl;
cout<<"Error! Skipping this molecule!"<<endl;
}
if (score < output_score_threshold && (!rescoring || !train) && !ignore_top)
{
*output << *mol;
}
delete mol;
}
/** If ignoring (i.e. not rescoring) the top fraction of docking-results, we need to make sure
that all rescored compounds get a rescore-value larger than that of the former. */
if (ignore_top && !train)
{
input->reopen();
int i = 1;
double offset = min_dock_score-min_rescore+0.25;
list<pair<double, bool> >::iterator rescore_it = rescore_list.begin();
for (Molecule* mol = input->read(); mol; mol = input->read(), i++)
{
String name = mol->getName();
if (mol->hasProperty("score_ligcheck"))
{
double score_ligcheck = ((String)mol->getProperty("score_ligcheck").toString()).toDouble();
if (score_ligcheck < 0.95) // 0 = error, 1 = check passed
{
delete mol;
continue;
}
}
double rescore = rescore_it->first;
if (rescore_it->second) // if compound was rescored
{
rescore += offset;
}
if (rescore < output_score_threshold)
{
mol->setProperty("re-score", rescore);
*output << *mol;
}
delete mol;
rescore_it++;
}
}
delete input;
if (output) delete output;
}
CHECK(const Fragment* getFragment() const throw())
TEST_EQUAL((atomx.getFragment() == &fragment), true)
RESULT
CHECK(Fragment* getFragment() throw())
atomx.getFragment()->setName("1234");
TEST_EQUAL(fragment.getName(), "1234")
RESULT
CHECK(const Molecule* getMolecule() const throw())
TEST_EQUAL(acx.getMolecule(), &molecule)
RESULT
CHECK(Molecule* getMolecule() throw())
atomx.getMolecule()->setName("asdff");
TEST_EQUAL(molecule.getName(), "asdff")
RESULT
Bond* bond;
CHECK(Bond* createBond(Atom& atom) throw())
atom->createBond(*atom3);
atom3->getBond(*atom);
TEST_EQUAL(atom->countBonds(), 1)
TEST_EQUAL(atom3->countBonds(), 1)
TEST_EQUAL(atom->getBond(*atom3), atom3->getBond(*atom))
atom->createBond(*atom4);
atom->createBond(*atom4);
bond = atom->createBond(*atom);
TEST_EQUAL(bond, 0);
TEST_EQUAL(atom4->countBonds(), 1)
