/** Print the population fraction of the entire simulation for each state and
* each residue
*/
void ProtTraj::PrintProtPop(string const& fname) {
// Set up the table of populations, initializing everything to zero
vector<StateCount> pop_counts;
// Now loop through every state and every residue, building pop_counts
for (Cpin::ResIterator rit = cpin_->begin(); rit != cpin_->end(); rit++) {
StateCount newstate;
newstate.nstates = 0;
for (int i = 0; i < rit->numStates(); i++) {
newstate.state_cnt.push_back(0ll);
newstate.prot_cnt.push_back(rit->numProtons(i));
newstate.nstates++;
}
pop_counts.push_back(newstate);
}
// Now loop through the entire trajectory, building the population list
for (int i = 1; i < nframes_ + 1; i++) {
ProtVector curst = statelist_[i];
for (int j = 0; j < nres_; j++)
pop_counts[j].state_cnt[ statelist_[i][j] ] += 1ll;
}
// Now print out the population counts
ofstream fp(fname.c_str());
if (!fp.is_open())
throw FileIOError("Could not open " + fname + " for writing!");
// Get the max number of states
int maxst = cpin_->getResidues()[0].numStates();
for (Cpin::ResIterator rit = cpin_->begin(); rit != cpin_->end(); rit++)
maxst = MAX(maxst, rit->numStates());
// Write the header
fp << right << fixed << setw(17) << "Residue Number" << " ";
for (int i = 0; i < maxst; i++)
fp << setw(9) << "State" << setw(3) << i << ' ';
fp << endl << setfill('-') << setw(maxst*13+18) << '-' << endl;
// Now loop through every residue and print out the populations
for (int i = 0; i < nres_; i++) {
stringstream iss;
iss << "Residue: " << cpin_->getResidues()[i].getResname() << " "
<< cpin_->getResidues()[i].getResnum();
fp << setfill(' ') << left << setw(17) << iss.str() << " ";
for (int j = 0; j < pop_counts[i].nstates; j++) {
double pop = (double) pop_counts[i].state_cnt[j] / (double) nframes_;
fp << setprecision(6) << setw(8) << pop << " (" << pop_counts[i].prot_cnt[j]
<< ") ";
}
fp << endl;
}
fp.close();
}
/// Prints a cumulative running average time series of the desired property
void ProtTraj::PrintCumulative(string const& fname, const int interval,
const bool print_prot, const bool print_pka) {
// Open up the file and write a header
ofstream fp(fname.c_str());
if (!fp.is_open())
throw FileIOError("Could not open " + fname + " for writing!");
fp << fixed << "#Time step ";
for (Cpin::ResIterator rit = cpin_->begin(); rit != cpin_->end(); rit++) {
stringstream iss;
iss << rit->getResname() << " " << rit->getResnum();
fp << setw(8) << iss.str() << " ";
}
fp << " Total Avg. Prot." << endl;
// Now go through the trajectory
vector<long long int> nprot(nres_, 0ll);
long long int totprot = 0ll;
int c = 0; // simple counter
for (int i = 1; i < nframes_ + 1; i++) {
{ // scope this
int j = 0;
for (Cpin::ResIterator rit = cpin_->begin(); rit != cpin_->end(); rit++) {
long long int protadd = (long long int) (rit->isProtonated(statelist_[i][j]));
nprot[j] += protadd;
totprot += (long long int) rit->numProtons(statelist_[i][j]);
j++;
}
}
if (c * time_step_ >= interval) {
c = 0;
// Now we print a point
int j = 0;
fp << setw(10) << (i-1)*time_step_ << " ";
for (Cpin::ResIterator rit = cpin_->begin(); rit != cpin_->end(); rit++) {
double fracprot = (double)nprot[j] / (double)(i+1);
if (print_prot) {
// We want the fraction protonated
fp << setprecision(5) << setw(8) << fracprot << " ";
}else if (print_pka) {
// We want the pKa
double pKa = pH_ - log10( (1.0 - fracprot) / fracprot );
fp << setprecision(4) << setw(8) << pKa << " ";
}else {
// We want the fraction deprotonated
fp << setprecision(5) << setw(8) << 1.0-fracprot << " ";
}
j++;
}
double fracprot = (double)totprot / (double)i;
fp << setprecision(6) << setw(17) << fracprot << endl;
}
c++; // heh
}
fp.close();
return;
}
/// Print the rolling/running average time series with a given window
void ProtTraj::PrintRunningAvg(const int window, const int interval,
string const& fname, const bool print_prot,
const bool print_pka) {
const int halfwin = window / (2 * time_step_);
// Open the file for writing and print a header
ofstream fp(fname.c_str());
if (!fp.is_open())
throw FileIOError("Could not open " + fname + " for writing!");
fp << fixed << "#Time step ";
for (Cpin::ResIterator rit = cpin_->begin(); rit != cpin_->end(); rit++) {
stringstream iss;
iss << rit->getResname() << " " << rit->getResnum();
fp << setw(8) << iss.str() << " ";
}
fp << " Total Avg. Prot." << endl;
for (int i = 1; i < nframes_ + 1; i+= interval/time_step_) {
vector<long long int> nprot(nres_, 0ll);
long long int totprot = 0ll;
// Loop over all frames we should include here
for (int j = MAX(0, i-halfwin); j < MIN(nframes_, i+halfwin); j++) {
// Loop over every residue
int k = 0;
for (Cpin::ResIterator rit = cpin_->begin(); rit != cpin_->end(); rit++) {
long long int protadd = (long long int) (rit->isProtonated(statelist_[j][k]));
nprot[k] += protadd;
totprot += (long long int) rit->numProtons(statelist_[j][k]);
k++;
}
}
// Print this frame to the output file
int j = 0;
int n = MIN(nframes_, i+halfwin) - MAX(0, i-halfwin);
fp << setw(10) << time_step_*i << " ";
for (Cpin::ResIterator rit = cpin_->begin(); rit != cpin_->end(); rit++) {
double fracprot = (double)nprot[j] / (double)n;
if (print_prot) {
// We want the fraction protonated
fp << setprecision(5) << setw(8) << fracprot << " ";
}else if (print_pka) {
// We want the pKa
double pKa = pH_ - log10( (1.0 - fracprot) / fracprot );
fp << setprecision(4) << setw(8) << pKa << " ";
}else {
// We want the fraction deprotonated
fp << setprecision(5) << setw(8) << 1.0-fracprot << " ";
}
j++;
}
double fracprot = (double)totprot / (double)n;
fp << setprecision(6) << setw(17) << fracprot << endl;
}
fp.close();
return;
}
/// Print a time series of the desired properties in 'chunks' of the simulation
void ProtTraj::PrintChunks(const int window, string const& fname,
const bool print_prot, const bool print_pka) {
// Open up the file and write a header (set fixed precision)
ofstream fp(fname.c_str());
if (!fp.is_open())
throw FileIOError("Could not open " + fname + " for writing!");
fp << fixed << "#Time step ";
for (Cpin::ResIterator rit = cpin_->begin(); rit != cpin_->end(); rit++) {
stringstream iss;
iss << rit->getResname() << " " << rit->getResnum();
fp << setw(8) << iss.str() << " ";
}
fp << " Total Avg. Prot." << endl;
int interval = window / time_step_;
int first = 0;
while (first < nframes_+1) {
vector<long long int> nprot(nres_, 0ll);
long long int totprot = 0ll;
int last = MIN(nframes_+1, first + interval);
for (int i = first; i < last; i++) {
int j = 0;
for (Cpin::ResIterator rit = cpin_->begin(); rit != cpin_->end(); rit++) {
long long int protadd = (long long int) (rit->isProtonated(statelist_[i][j]));
nprot[j] += protadd;
totprot += (long long int) rit->numProtons(statelist_[i][j]);
j++;
}
}
// Now write the information
int j = 0;
fp << setw(10) << time_step_*(first + interval / 2) << " ";
for (Cpin::ResIterator rit = cpin_->begin(); rit != cpin_->end(); rit++) {
double fracprot = (double)nprot[j] / (double)(last - first);
if (print_prot) {
// We want the fraction protonated
fp << setprecision(5) << setw(8) << fracprot << " ";
}else if (print_pka) {
// We want the pKa
double pKa = pH_ - log10( (1.0 - fracprot) / fracprot );
fp << setprecision(4) << setw(8) << pKa << " ";
}else {
// We want the fraction deprotonated
fp << setprecision(5) << setw(8) << 1.0-fracprot << " ";
}
j++;
}
double fracprot = (double)totprot/(double)(last - first);
fp << setprecision(6) << setw(17) << fracprot << endl;
first += interval;
}
// Now that I'm here (and done), close the file
fp.close();
return;
}
/// Prints the calcpka-style output
void ProtTraj::PrintCalcpka(ostream& fd, const int start) {
// First calculate the statistics
vector<long long int> nprot(nres_, 0ll);
vector<int> transitions(nres_, 0);
long long int totprot = 0ll;
// Start with the first point, but skip it in the iteration
last_point_ = statelist_[start];
for (int i = start + 1; i < nframes_ + 1; i++) {
int j = 0;
for (Cpin::ResIterator rit = cpin_->begin(); rit != cpin_->end(); rit++) {
transitions[j] += (int) (rit->isProtonated(last_point_[j]) !=
rit->isProtonated(statelist_[i][j]));
long long int protadd = (long long int) (rit->isProtonated(statelist_[i][j]));
nprot[j] += protadd;
totprot += (long long int) rit->numProtons(statelist_[i][j]);
j++;
}
// Reassign the last point to the current one before incrementing
last_point_ = statelist_[i];
}
// Now do the printing
fd.precision(3);
fd << fixed;
fd << "Solvent pH is " << setw(8) << pH_ << endl;
int i = 0;
for (Cpin::ResIterator rit = cpin_->begin(); rit != cpin_->end(); rit++) {
double pKa = pH_ - log10( (double) (nframes_-nprot[i]) / (double) nprot[i] );
float offset = (float) pKa - pH_;
fd << setw(3) << rit->getResname() << " " << setw(4) << left << rit->getResnum()
<< ": Offset " << right << setw(6) << offset << " Pred " << setw(6) << pKa
<< " Frac Prot " << setw(5) << (double) nprot[i] / (double) nframes_
<< " Transitions " << setw(9) << transitions[i] << endl;
i++;
}
fd << endl << "Average total molecular protonation: " << setw(7)
<< (double)totprot / (double)nframes_ << endl;
return;
}
int main(int argc, char**argv) {
// Set up the command-line options and parse them
CLOptions clopt = CLOptions(argc, argv);
if (clopt.Parse())
return 1;
// Check the input
if (clopt.CheckInput()) {
cerr << "Error: Input errors detected! See messages above." << endl;
return 1;
}
// test_clopt(clopt); /* Uncomment to test command-line parsing */
int nres = 0; // number of residues (for consistency tests)
/* Set up the cpin and print some diagnostic information (only necessary if
* we're not fixing REMD files
*/
Cpin my_cpin;
if (clopt.REMDPrefix().empty()) {
if ( clopt.Cpin().empty() ) {
#ifdef REDOX
cerr << "Error: No cein file provided!" << endl;
#else
cerr << "Error: No cpin file provided!" << endl;
#endif
return 1;
}
if ( my_cpin.Parse(clopt.Cpin()) )
return 1;
nres = (int) my_cpin.getTrescnt();
if (nres <= 0) {
cerr << "Error: Did not detect any residues in " <<
clopt.Cpin() << endl;
return 1;
}
if (clopt.Debug()) {
cout << "There are " << nres << " titratable residues defined in " <<
my_cpin.getFilename() << ":" << endl;
cout << "They are:" << endl;
for (Cpin::ResIterator it = my_cpin.begin(); it != my_cpin.end(); it++) {
cout << "\t" << setw(3) << it->getResname() << left << " " <<
setw(3) << it->getResnum() << " (" << it->numStates()
<< " states) [ ";
for (int j = 0; j < it->numStates(); j++) {
#ifdef REDOX
if (it->isProtonated(j))
cout << "R ";
else
cout << "O ";
#else
if (it->isProtonated(j))
cout << "P ";
else
cout << "D ";
#endif
}
cout << "]" << endl;
}
}
} // if clopt.REMDPrefix().empty()
// Set up the cpouts
CpoutList cpouts;
for (CLOptions::cpout_iterator it = clopt.begin(); it != clopt.end(); it++) {
CpoutFile c = CpoutFile(&my_cpin, *it);
// Skip over invalid cpouts
if (!c.Valid()) {
#ifdef REDOX
cerr << "Error: Ceout file " << *it << " is invalid! Skipping." << endl;
#else
cerr << "Error: Cpout file " << *it << " is invalid! Skipping." << endl;
#endif
continue;
}
// For REMD fixing where a cpin is unnecessary, make sure all cpouts have
// the same number of residues, so set nres to the first cpout's Nres
if (nres <= 0) nres = c.Nres();
// Skip over cpouts with a residue mismatch
if (c.Nres() != nres) {
#ifdef REDOX
cerr << "Error: Ceout file " << *it << " has " << c.Nres() <<
#else
cerr << "Error: Cpout file " << *it << " has " << c.Nres() <<
#endif
" residues. I expected " << my_cpin.getTrescnt() <<
". Skipping." << endl;
continue;
}
if (clopt.Debug())
#ifdef REDOX
cout << "Added [[ " << *it << " ]] to ceout list." << endl;
#else
cout << "Added [[ " << *it << " ]] to cpout list." << endl;
#endif
cpouts.push_back(c);
}
if (clopt.Debug())
#ifdef REDOX
cout << "Analyzing " << clopt.Cpouts().size() << " ceouts." << endl;
if (cpouts.size() != clopt.Cpouts().size()) {
cerr << "Error: Number of Ceout files " << cpouts.size() <<
#else
cout << "Analyzing " << clopt.Cpouts().size() << " cpouts." << endl;
if (cpouts.size() != clopt.Cpouts().size()) {
cerr << "Error: Number of Cpout files " << cpouts.size() <<
#endif
" does not equal number specified: " << clopt.Cpouts().size() << endl;
return 1;
}
// Special-case REMD re-ordering
if (!clopt.REMDPrefix().empty())
return sort_remd_files(cpouts, clopt.REMDPrefix(), clopt.Overwrite());
#ifdef REDOX
ProtTraj stats = ProtTraj(&my_cpin, cpouts[0].pH(), cpouts[0].GetRecord(), cpouts[0].Temperature());
#else
ProtTraj stats = ProtTraj(&my_cpin, cpouts[0].pH(), cpouts[0].GetRecord());
#endif
for (cpout_iterator it = cpouts.begin(); it != cpouts.end(); it++) {
// If we are here, then warn if we are about to use a REMD file
if (!clopt.Expert())
it->WarnRemd();
stats.LoadCpout(*it);
}
// Do the normal calcpka-style output if requested
if (clopt.Calcpka()) {
if (clopt.Output().empty())
stats.PrintCalcpka(cout);
else {
ofstream fd; fd.open(clopt.Output().c_str());
stats.PrintCalcpka(fd);
fd.close();
}
}
// Do chunk analysis
if (clopt.ChunkWindow() > 0)
stats.PrintChunks(clopt.ChunkWindow(), clopt.ChunkOutput(),
clopt.PrintProtonated() && !clopt.pKa(), clopt.pKa());
// Do cumulative analysis
if (clopt.doCumulative())
stats.PrintCumulative(clopt.CumulativeOutput(), clopt.Interval(),
clopt.PrintProtonated() && !clopt.pKa(), clopt.pKa());
// Do running averages
if (clopt.RunningAvgWindow() > 0)
stats.PrintRunningAvg(clopt.RunningAvgWindow(), clopt.Interval(),
clopt.RunningAvgOutput(),
clopt.PrintProtonated() && !clopt.pKa(), clopt.pKa());
// Do protonation fraction dump
if (clopt.Population().size() > 0)
stats.PrintProtPop(clopt.Population());
// Do conditional probabilities
if (clopt.CondProbs().size() > 0) {
for (CLOptions::prob_iterator it = clopt.condbegin();
it != clopt.condend(); it++)
if (it->Set(my_cpin) == ConditionalProb::ERR) {
cerr << "Quitting due to errors above." << endl;
return 1;
}
stats.PrintCondProb(clopt.ConditionalOutput(), clopt.CondProbs());
}
// Do conditional probability time series (we already set the conditional
// probabilities above, so no need to do it again)
if (clopt.CondProbs().size() > 0 && !clopt.ConditionalChunkOut().empty())
stats.PrintCondTimeseries(clopt.ConditionalChunkOut(), clopt.Interval(),
clopt.CondProbs());
if (clopt.Debug()) cout << "All done!" << endl;
return 0;
}