int SequenceAlign(CpptrajState& State, ArgList& argIn) {
std::string blastfile = argIn.GetStringKey("blastfile");
if (blastfile.empty()) {
mprinterr("Error: 'blastfile' must be specified.\n");
return 1;
}
ReferenceFrame qref = State.DSL()->GetReferenceFrame(argIn);
if (qref.error() || qref.empty()) {
mprinterr("Error: Must specify reference structure for query.\n");
return 1;
}
std::string outfilename = argIn.GetStringKey("out");
if (outfilename.empty()) {
mprinterr("Error: Must specify output file.\n");
return 1;
}
TrajectoryFile::TrajFormatType fmt = TrajectoryFile::GetFormatFromArg(argIn);
if (fmt != TrajectoryFile::PDBFILE && fmt != TrajectoryFile::MOL2FILE)
fmt = TrajectoryFile::PDBFILE; // Default to PDB
int smaskoffset = argIn.getKeyInt("smaskoffset", 0) + 1;
int qmaskoffset = argIn.getKeyInt("qmaskoffset", 0) + 1;
// Load blast file
mprintf("\tReading BLAST alignment from '%s'\n", blastfile.c_str());
BufferedLine infile;
if (infile.OpenFileRead( blastfile )) return 1;
// Seek down to first Query line.
const char* ptr = infile.Line();
bool atFirstQuery = false;
while (ptr != 0) {
if (*ptr == 'Q') {
if ( strncmp(ptr, "Query", 5) == 0 ) {
atFirstQuery = true;
break;
}
}
ptr = infile.Line();
}
if (!atFirstQuery) {
mprinterr("Error: 'Query' not found.\n");
return 1;
}
// Read alignment. Replacing query with subject.
typedef std::vector<char> Carray;
typedef std::vector<int> Iarray;
Carray Query; // Query residues
Carray Sbjct; // Sbjct residues
Iarray Smap; // Smap[Sbjct index] = Query index
while (ptr != 0) {
const char* qline = ptr; // query line
const char* aline = infile.Line(); // alignment line
const char* sline = infile.Line(); // subject line
if (aline == 0 || sline == 0) {
mprinterr("Error: Missing alignment line or subject line after Query:\n");
mprinterr("Error: %s", qline);
return 1;
}
for (int idx = 12; qline[idx] != ' '; idx++) {
if (qline[idx] == '-') {
// Sbjct does not have corresponding res in Query
Smap.push_back(-1);
Sbjct.push_back( sline[idx] );
} else if (sline[idx] == '-') {
// Query does not have a corresponding res in Sbjct
Query.push_back( qline[idx] );
} else {
// Direct Query to Sbjct map
Smap.push_back( Query.size() );
Sbjct.push_back( sline[idx] );
Query.push_back( qline[idx] );
}
}
// Scan to next Query
ptr = infile.Line();
while (ptr != 0) {
if (*ptr == 'Q') {
if ( strncmp(ptr, "Query", 5) == 0 ) break;
}
ptr = infile.Line();
}
}
// DEBUG
std::string SmaskExp, QmaskExp;
if (State.Debug() > 0) mprintf(" Map of Sbjct to Query:\n");
for (int sres = 0; sres != (int)Sbjct.size(); sres++) {
if (State.Debug() > 0)
mprintf("%-i %3s %i", sres+smaskoffset, Residue::ConvertResName(Sbjct[sres]),
Smap[sres]+qmaskoffset);
const char* qres = "";
if (Smap[sres] != -1) {
qres = Residue::ConvertResName(Query[Smap[sres]]);
if (SmaskExp.empty())
SmaskExp.assign( integerToString(sres+smaskoffset) );
else
SmaskExp.append( "," + integerToString(sres+smaskoffset) );
if (QmaskExp.empty())
QmaskExp.assign( integerToString(Smap[sres]+qmaskoffset) );
else
QmaskExp.append( "," + integerToString(Smap[sres]+qmaskoffset) );
}
if (State.Debug() > 0) mprintf(" %3s\n", qres);
}
mprintf("Smask: %s\n", SmaskExp.c_str());
mprintf("Qmask: %s\n", QmaskExp.c_str());
// Check that query residues match reference.
for (unsigned int sres = 0; sres != Sbjct.size(); sres++) {
int qres = Smap[sres];
if (qres != -1) {
if (Query[qres] != qref.Parm().Res(qres).SingleCharName()) {
mprintf("Warning: Potential residue mismatch: Query %s reference %s\n",
Residue::ConvertResName(Query[qres]), qref.Parm().Res(qres).c_str());
}
}
}
// Build subject using coordinate from reference.
//AtomMask sMask; // Contain atoms that should be in sTop
Topology sTop;
Frame sFrame;
Iarray placeHolder; // Atom indices of placeholder residues.
for (unsigned int sres = 0; sres != Sbjct.size(); sres++) {
int qres = Smap[sres];
NameType SresName( Residue::ConvertResName(Sbjct[sres]) );
if (qres != -1) {
Residue const& QR = qref.Parm().Res(qres);
Residue SR(SresName, sres+1, ' ', QR.ChainID());
if (Query[qres] == Sbjct[sres]) { // Exact match. All non-H atoms.
for (int qat = QR.FirstAtom(); qat != QR.LastAtom(); qat++)
{
if (qref.Parm()[qat].Element() != Atom::HYDROGEN)
sTop.AddTopAtom( qref.Parm()[qat], SR );
sFrame.AddXYZ( qref.Coord().XYZ(qat) );
//sMask.AddAtom(qat);
}
} else { // Partial match. Copy only backbone and CB.
for (int qat = QR.FirstAtom(); qat != QR.LastAtom(); qat++)
{
if ( qref.Parm()[qat].Name().Match("N" ) ||
qref.Parm()[qat].Name().Match("CA") ||
qref.Parm()[qat].Name().Match("CB") ||
qref.Parm()[qat].Name().Match("C" ) ||
qref.Parm()[qat].Name().Match("O" ) )
{
sTop.AddTopAtom( qref.Parm()[qat], SR );
sFrame.AddXYZ( qref.Coord().XYZ(qat) );
}
}
}
} else {
// Residue in query does not exist for subject. Just put placeholder CA for now.
Vec3 Zero(0.0);
placeHolder.push_back( sTop.Natom() );
sTop.AddTopAtom( Atom("CA", "C "), Residue(SresName, sres+1, ' ', ' ') );
sFrame.AddXYZ( Zero.Dptr() );
}
}
//sTop.PrintAtomInfo("*");
mprintf("\tPlaceholder residue indices:");
for (Iarray::const_iterator p = placeHolder.begin(); p != placeHolder.end(); ++p)
mprintf(" %i", *p + 1);
mprintf("\n");
// Try to give placeholders more reasonable coordinates.
if (!placeHolder.empty()) {
Iarray current_indices;
unsigned int pidx = 0;
while (pidx < placeHolder.size()) {
if (current_indices.empty()) {
current_indices.push_back( placeHolder[pidx++] );
// Search for the end of this segment
for (; pidx != placeHolder.size(); pidx++) {
if (placeHolder[pidx] - current_indices.back() > 1) break;
current_indices.push_back( placeHolder[pidx] );
}
// DEBUG
mprintf("\tSegment:");
for (Iarray::const_iterator it = current_indices.begin();
it != current_indices.end(); ++it)
mprintf(" %i", *it + 1);
// Get coordinates of residues bordering segment.
int prev_res = sTop[current_indices.front()].ResNum() - 1;
int next_res = sTop[current_indices.back() ].ResNum() + 1;
mprintf(" (prev_res=%i, next_res=%i)\n", prev_res+1, next_res+1);
Vec3 prev_crd(sFrame.XYZ(current_indices.front() - 1));
Vec3 next_crd(sFrame.XYZ(current_indices.back() + 1));
prev_crd.Print("prev_crd");
next_crd.Print("next_crd");
Vec3 crd_step = (next_crd - prev_crd) / (double)(current_indices.size()+1);
crd_step.Print("crd_step");
double* xyz = sFrame.xAddress() + (current_indices.front() * 3);
for (unsigned int i = 0; i != current_indices.size(); i++, xyz += 3) {
prev_crd += crd_step;
xyz[0] = prev_crd[0];
xyz[1] = prev_crd[1];
xyz[2] = prev_crd[2];
}
current_indices.clear();
}
}
}
//Topology* sTop = qref.Parm().partialModifyStateByMask( sMask );
//if (sTop == 0) return 1;
//Frame sFrame(qref.Coord(), sMask);
// Write output traj
Trajout_Single trajout;
if (trajout.PrepareTrajWrite(outfilename, argIn, &sTop, CoordinateInfo(), 1, fmt)) return 1;
if (trajout.WriteSingle(0, sFrame)) return 1;
trajout.EndTraj();
return 0;
}
/** Determine what atoms each mask pertains to for the current parm file.
*/
Action::RetType Action_NMRrst::Setup(ActionSetup& setup) {
if (!viewrst_.empty() && rsttop_ == 0) rsttop_ = setup.TopAddress();
// ---------------------------------------------
// Set up NOEs from file.
for (noeDataArray::iterator noe = NOEs_.begin(); noe != NOEs_.end(); ++noe) {
if (setup.Top().SetupIntegerMask( noe->dMask1_ )) return Action::ERR;
if (setup.Top().SetupIntegerMask( noe->dMask2_ )) return Action::ERR;
if (noe->dMask1_.None() || noe->dMask2_.None()) {
mprintf("Warning: One or both masks for NOE '%s' have no atoms (%i and %i).\n",
noe->dist_->legend(), noe->dMask1_.Nselected(),
noe->dMask2_.Nselected());
noe->active_ = false;
} else
noe->active_ = true;
}
// ---------------------------------------------
// Set up potential NOE sites.
if (findNOEs_) {
if (setup.Top().SetupCharMask( Mask_ )) return Action::ERR;
Mask_.MaskInfo();
if (Mask_.None()) return Action::SKIP;
SiteArray potentialSites; // .clear();
AtomMap resMap;
resMap.SetDebug( debug_ );
std::vector<bool> selected;
Range soluteRes = setup.Top().SoluteResidues();
for (Range::const_iterator res = soluteRes.begin(); res != soluteRes.end(); ++res)
{
int res_first_atom = setup.Top().Res(*res).FirstAtom();
selected.assign( setup.Top().Res(*res).NumAtoms(), false );
// Find symmetric atom groups.
AtomMap::AtomIndexArray symmGroups;
if (resMap.SymmetricAtoms(setup.Top(), symmGroups, *res)) return Action::ERR;
// DEBUG
if (debug_ > 0) {
mprintf("DEBUG: Residue %i: symmetric atom groups:\n", *res + 1);
for (AtomMap::AtomIndexArray::const_iterator grp = symmGroups.begin();
grp != symmGroups.end(); ++grp)
{
mprintf("\t\t");
for (AtomMap::Iarray::const_iterator at = grp->begin();
at != grp->end(); ++at)
mprintf(" %s", setup.Top().TruncAtomNameNum( *at ).c_str());
mprintf("\n");
}
}
// Each symmetric hydrogen atom group is a site.
for (AtomMap::AtomIndexArray::const_iterator grp = symmGroups.begin();
grp != symmGroups.end(); ++grp)
{ // NOTE: If first atom is H all should be H.
if ( setup.Top()[ grp->front() ].Element() == Atom::HYDROGEN )
{
Iarray symmAtomGroup;
for (Iarray::const_iterator at = grp->begin();
at != grp->end(); ++at)
if (Mask_.AtomInCharMask( *at ))
symmAtomGroup.push_back( *at );
if (!symmAtomGroup.empty()) {
potentialSites.push_back( Site(*res, symmAtomGroup) );
// Mark symmetric atoms as selected.
for (AtomMap::Iarray::const_iterator at = grp->begin();
at != grp->end(); ++at)
selected[ *at - res_first_atom ] = true;
}
}
}
// All other non-selected hydrogens bonded to same heavy atom are sites.
for (int ratom = res_first_atom; ratom != setup.Top().Res(*res).LastAtom(); ++ratom)
{
if ( setup.Top()[ratom].Element() != Atom::HYDROGEN ) {
Iarray heavyAtomGroup;
for (Atom::bond_iterator ba = setup.Top()[ratom].bondbegin();
ba != setup.Top()[ratom].bondend(); ++ba)
if ( Mask_.AtomInCharMask(*ba) &&
*ba >= res_first_atom &&
*ba < setup.Top().Res(*res).LastAtom() )
{
if ( !selected[ *ba - res_first_atom ] &&
setup.Top()[ *ba ].Element() == Atom::HYDROGEN )
heavyAtomGroup.push_back( *ba );
}
if (!heavyAtomGroup.empty())
potentialSites.push_back( Site(*res, heavyAtomGroup) );
}
}
}
mprintf("\t%zu potential NOE sites:\n", potentialSites.size());
for (SiteArray::const_iterator site = potentialSites.begin();
site != potentialSites.end(); ++site)
{
mprintf(" %u\tRes %i:", site - potentialSites.begin(), site->ResNum()+1);
for (unsigned int idx = 0; idx != site->Nindices(); ++idx)
mprintf(" %s", setup.Top().TruncAtomNameNum( site->Idx(idx) ).c_str());
mprintf("\n");
}
if (noeArray_.empty()) {
size_t siteArraySize = 0;
// Set up all potential NOE pairs. Keep track of size.
for (SiteArray::const_iterator site1 = potentialSites.begin();
site1 != potentialSites.end(); ++site1)
{
for (SiteArray::const_iterator site2 = site1 + 1;
site2 != potentialSites.end(); ++site2)
{
if (site1->ResNum() != site2->ResNum()) {
std::string legend = site1->SiteLegend(setup.Top()) + "--" +
site2->SiteLegend(setup.Top());
DataSet* ds = 0;
if (series_) {
ds = masterDSL_->AddSet(DataSet::FLOAT,
MetaData(setname_, "foundNOE", noeArray_.size()));
if (ds == 0) return Action::ERR;
// Construct a data set name.
ds->SetLegend(legend);
}
noeArray_.push_back( NOEtype(*site1, *site2, ds, legend) );
siteArraySize += (2 * sizeof(int) * site1->Nindices()) +
(2 * sizeof(int) * site2->Nindices());
}
}
}
numNoePairs_ = noeArray_.size();
size_t siteSize = sizeof(int) + (2 * sizeof(Iarray)) + sizeof(Site);
size_t noeSize = (2 * siteSize) + sizeof(DataSet*) + sizeof(double)
+ sizeof(NOEtype);
if (series_) noeSize += sizeof(std::vector<float>);
size_t noeArraySize = (noeSize * numNoePairs_) + siteArraySize;
if (series_)
noeArraySize += (setup.Nframes() * numNoePairs_ * sizeof(float));
mprintf("\t%zu potential NOE pairs. Estimated memory usage is %s\n",
numNoePairs_, ByteString(noeArraySize, BYTE_DECIMAL).c_str());
} else if (numNoePairs_ != potentialSites.size()) {
mprinterr("Warning: Found NOE matrix has already been set up for %zu potential\n"
"Warning: NOEs, but %zu NOEs currently found.\n", numNoePairs_,
potentialSites.size());
return Action::SKIP;
}
}
// ---------------------------------------------
// Set up NOEs specified on the command line
if (!Pairs_.empty()) {
if (!specifiedNOEs_.empty()) {
mprintf("Warning: Specifying NOEs currently only works with first topology used.\n");
return Action::SKIP;
}
for (MaskPairArray::iterator mp = Pairs_.begin(); mp != Pairs_.end(); mp++) {
if (setup.Top().SetupIntegerMask( mp->first )) return Action::ERR;
int res1 = CheckSameResidue(setup.Top(), mp->first);
if (res1 < 0) continue;
if (setup.Top().SetupIntegerMask( mp->second )) return Action::ERR;
int res2 = CheckSameResidue(setup.Top(), mp->second);
if (res2 < 0) continue;
Site site1( res1, mp->first.Selected() );
Site site2( res2, mp->second.Selected() );
std::string legend = site1.SiteLegend(setup.Top()) + "--" +
site2.SiteLegend(setup.Top());
DataSet* ds = 0;
if (series_) {
ds = masterDSL_->AddSet(DataSet::FLOAT,
MetaData(setname_, "specNOE", specifiedNOEs_.size()));
if (ds == 0) return Action::ERR;
ds->SetLegend(legend);
}
specifiedNOEs_.push_back( NOEtype(site1, site2, ds, legend) );
}
}
// Set up imaging info for this parm
Image_.SetupImaging( setup.CoordInfo().TrajBox().Type() );
if (Image_.ImagingEnabled())
mprintf("\tImaged.\n");
else
mprintf("\tImaging off.\n");
return Action::OK;
}
/** Find potential symmetric atoms. All residues up to the last selected
* residue are considered.
*/
int SymmetricRmsdCalc::SetupSymmRMSD(Topology const& topIn, AtomMask const& tgtMask, bool remapIn)
{
// Allocate space for remapping selected atoms in target frame. This will
// also put the correct masses in based on the mask.
tgtRemap_.SetupFrameFromMask(tgtMask, topIn.Atoms());
// Create map of original atom numbers to selected indices
Iarray SelectedIdx( topIn.Natom(), -1 );
int tgtIdx = 0;
for (int originalAtom = 0; originalAtom != topIn.Natom(); ++originalAtom)
if ( originalAtom == tgtMask[tgtIdx] )
SelectedIdx[originalAtom] = tgtIdx++;
if (debug_ > 0) {
mprintf("DEBUG: Original atom -> Selected Index mapping:\n");
for (int originalAtom = 0; originalAtom != topIn.Natom(); ++originalAtom)
mprintf("\t%8i -> %8i\n", originalAtom + 1, SelectedIdx[originalAtom] + 1);
}
// Create initial 1 to 1 atom map for all selected atoms; indices in
// SymmetricAtomIndices will correspond to positions in AMap.
AMap_.resize( tgtRemap_.Natom() );
// Determine last selected residue.
int last_res = topIn[tgtMask.back()].ResNum() + 1;
mprintf("\tResidues up to %s will be considered for symmetry correction.\n",
topIn.TruncResNameNum(last_res-1).c_str());
// In each residue, determine which selected atoms are symmetric.
SymmetricAtomIndices_.clear();
AtomMap resmap;
if (debug_ > 1) resmap.SetDebug(1);
for (int res = 0; res < last_res; ++res) {
AtomMap::AtomIndexArray residue_SymmetricGroups;
if (resmap.SymmetricAtoms(topIn, residue_SymmetricGroups, res)) {
mprinterr("Error: Finding symmetric atoms in residue '%s'\n",
topIn.TruncResNameNum(res).c_str());
return 1;
}
if (!residue_SymmetricGroups.empty()) {
// Which atoms in symmetric groups are selected?
bool resHasSelectedSymmAtoms = false;
for (AtomMap::AtomIndexArray::const_iterator symmGroup = residue_SymmetricGroups.begin();
symmGroup != residue_SymmetricGroups.end();
++symmGroup)
{
Iarray selectedAtomIndices;
for (Iarray::const_iterator atnum = symmGroup->begin();
atnum != symmGroup->end(); ++atnum)
{
if ( SelectedIdx[*atnum] != -1 )
selectedAtomIndices.push_back( SelectedIdx[*atnum] ); // Store tgtMask indices
}
if (!selectedAtomIndices.empty()) {
SymmetricAtomIndices_.push_back( selectedAtomIndices );
resHasSelectedSymmAtoms = true;
}
}
// If remapping and not all atoms in a residue are selected, warn user.
// TODO: Should they just be considered even if not selected?
if (remapIn && resHasSelectedSymmAtoms) {
for (int atom = topIn.Res(res).FirstAtom(); atom != topIn.Res(res).LastAtom(); ++atom)
if (SelectedIdx[atom] == -1) {
mprintf("Warning: Not all atoms selected in residue '%s'. Re-mapped\n"
"Warning: structures may appear distorted.\n",
topIn.TruncResNameNum(res).c_str());
break;
}
}
}
}
if (debug_ > 0) {
mprintf("DEBUG: Potential Symmetric Atom Groups:\n");
for (AtomIndexArray::const_iterator symmatoms = SymmetricAtomIndices_.begin();
symmatoms != SymmetricAtomIndices_.end();
++symmatoms)
{
mprintf("\t%8u) ", symmatoms - SymmetricAtomIndices_.begin());
for (Iarray::const_iterator atom = symmatoms->begin();
atom != symmatoms->end(); ++atom)
mprintf(" %s(%i)", topIn.AtomMaskName(tgtMask[*atom]).c_str(), tgtMask[*atom] + 1);
mprintf("\n");
}
}
return 0;
}
