/** An atom pair list consists of 2 values for each entry, a beginning
* index and ending index. For molecules and residues this is the first
* and just beyond the last atom; for atoms it is just the atom itself
* twice.
*/
Image::PairType Image::CreatePairList(Topology const& Parm, Mode modeIn,
std::string const& maskExpression)
{
PairType atomPairs;
// Set up mask based on desired imaging mode.
if ( modeIn == BYMOL || modeIn == BYRES ) {
CharMask cmask( maskExpression );
if ( Parm.SetupCharMask( cmask ) ) return atomPairs;
cmask.MaskInfo();
if (cmask.None()) return atomPairs;
// Set up atom range for each entity to be imaged.
if (modeIn == BYMOL) {
atomPairs.reserve( Parm.Nmol()*2 );
for (Topology::mol_iterator mol = Parm.MolStart();
mol != Parm.MolEnd(); ++mol)
CheckRange( atomPairs, cmask, mol->BeginAtom(), mol->EndAtom());
} else { // BYRES
atomPairs.reserve( Parm.Nres()*2 );
for (Topology::res_iterator residue = Parm.ResStart();
residue != Parm.ResEnd(); ++residue)
CheckRange( atomPairs, cmask, residue->FirstAtom(), residue->LastAtom() );
}
} else { // BYATOM
AtomMask imask( maskExpression );
if ( Parm.SetupIntegerMask( imask ) ) return atomPairs;
imask.MaskInfo();
if (imask.None()) return atomPairs;
atomPairs.reserve( Parm.Natom()*2 );
for (AtomMask::const_iterator atom = imask.begin(); atom != imask.end(); ++atom) {
atomPairs.push_back( *atom );
atomPairs.push_back( (*atom)+1 );
}
}
// mprintf("\tNumber of %ss to be imaged is %zu based on mask '%s'\n",
// ModeString[modeIn], atomPairs.size()/2, maskIn.MaskString());
return atomPairs;
}
// Parm_CIF::ReadParm()
int Parm_CIF::ReadParm(FileName const& fname, Topology &TopIn) {
CIFfile infile;
CIFfile::DataBlock::data_it line;
if (infile.Read( fname, debug_ )) return 1;
CIFfile::DataBlock const& block = infile.GetDataBlock("_atom_site");
if (block.empty()) {
mprinterr("Error: CIF data block '_atom_site' not found.\n");
return 1;
}
// Does this CIF contain multiple models?
int Nmodels = 0;
int model_col = block.ColumnIndex("pdbx_PDB_model_num");
if (model_col != -1) {
line = block.end();
--line;
Nmodels = convertToInteger( (*line)[model_col] );
if (Nmodels > 1)
mprintf("Warning: CIF '%s' contains %i models. Using first model for topology.\n",
fname.full(), Nmodels);
}
// Get essential columns
int COL[NENTRY];
for (int i = 0; i < (int)NENTRY; i++) {
COL[i] = block.ColumnIndex(Entries[i]);
if (COL[i] == -1) {
mprinterr("Error: In CIF file '%s' could not find entry '%s' in block '%s'\n",
fname.full(), Entries[i], block.Header().c_str());
return 1;
}
if (debug_>0) mprintf("DEBUG: '%s' column = %i\n", Entries[i], COL[i]);
}
// Get optional columns
int occ_col = block.ColumnIndex("occupancy");
int bfac_col = block.ColumnIndex("B_iso_or_equiv");
int icode_col = block.ColumnIndex("pdbx_PDB_ins_code");
int altloc_col = block.ColumnIndex("label_alt_id");
std::vector<AtomExtra> extra;
// Loop over all atom sites
int current_res = 0;
double XYZ[3];
double occupancy = 1.0;
double bfactor = 0.0;
char altloc = ' ';
char icode;
icode = ' ';
Frame Coords;
for (line = block.begin(); line != block.end(); ++line) {
// If more than 1 model check if we are done.
if (Nmodels > 1) {
if ( convertToInteger( (*line)[model_col] ) > 1 )
break;
}
if (occ_col != -1) occupancy = convertToDouble( (*line)[ occ_col ] );
if (bfac_col != -1) bfactor = convertToDouble( (*line)[ bfac_col ] );
if (altloc_col != -1) altloc = (*line)[ altloc_col ][0];
// '.' altloc means blank?
if (altloc == '.') altloc = ' ';
extra.push_back( AtomExtra(occupancy, bfactor, altloc) );
if (icode_col != -1) {
icode = (*line)[ icode_col ][0];
// '?' icode means blank
if (icode == '?') icode = ' ';
}
XYZ[0] = convertToDouble( (*line)[ COL[X] ] );
XYZ[1] = convertToDouble( (*line)[ COL[Y] ] );
XYZ[2] = convertToDouble( (*line)[ COL[Z] ] );
NameType currentResName( (*line)[ COL[RNAME] ] );
// It seems that in some CIF files, there doesnt have to be a residue
// number. Check if residue name has changed.
if ( (*line)[ COL[RNUM] ][0] == '.' ) {
Topology::res_iterator lastResidue = TopIn.ResEnd();
--lastResidue;
if ( currentResName != (*lastResidue).Name() )
current_res = TopIn.Nres() + 1;
} else
current_res = convertToInteger( (*line)[ COL[RNUM] ] );
TopIn.AddTopAtom( Atom((*line)[ COL[ANAME] ], " "),
Residue(currentResName, current_res, icode,
(*line)[ COL[CHAINID] ][0]) );
Coords.AddXYZ( XYZ );
}
if (TopIn.SetExtraAtomInfo( 0, extra )) return 1;
// Search for bonds // FIXME nobondsearch?
BondSearch( TopIn, Coords, Offset_, debug_ );
// Get title.
CIFfile::DataBlock const& entryblock = infile.GetDataBlock("_entry");
std::string ciftitle;
if (!entryblock.empty())
ciftitle = entryblock.Data("id");
TopIn.SetParmName( ciftitle, infile.CIFname() );
// Get unit cell parameters if present.
CIFfile::DataBlock const& cellblock = infile.GetDataBlock("_cell");
if (!cellblock.empty()) {
double cif_box[6];
cif_box[0] = convertToDouble( cellblock.Data("length_a") );
cif_box[1] = convertToDouble( cellblock.Data("length_b") );
cif_box[2] = convertToDouble( cellblock.Data("length_c") );
cif_box[3] = convertToDouble( cellblock.Data("angle_alpha") );
cif_box[4] = convertToDouble( cellblock.Data("angle_beta" ) );
cif_box[5] = convertToDouble( cellblock.Data("angle_gamma") );
mprintf("\tRead cell info from CIF: a=%g b=%g c=%g alpha=%g beta=%g gamma=%g\n",
cif_box[0], cif_box[1], cif_box[2], cif_box[3], cif_box[4], cif_box[5]);
TopIn.SetParmBox( Box(cif_box) );
}
return 0;
}
/** Search for bonds between atoms in residues and atoms in adjacent residues
* using distance-based criterion that depends on atomic elements.
* \param top Topology to add bonds to.
* \param frameIn Frame containing atomic coordinates.
* \param offset Offset to add when determining if a bond is present.
* \param debug If > 0 print extra info.
*/
int BondSearch( Topology& top, Frame const& frameIn, double offset, int debug) {
mprintf("\tDetermining bond info from distances.\n");
if (frameIn.empty()) {
mprinterr("Internal Error: No coordinates set; cannot search for bonds.\n");
return 1;
}
# ifdef TIMER
Timer time_total, time_within, time_between;
time_total.Start();
time_within.Start();
# endif
// ----- STEP 1: Determine bonds within residues
for (Topology::res_iterator res = top.ResStart(); res != top.ResEnd(); ++res)
{
int stopatom = res->LastAtom();
// Check for bonds between each atom in the residue.
for (int atom1 = res->FirstAtom(); atom1 != stopatom; ++atom1) {
Atom::AtomicElementType a1Elt = top[atom1].Element();
// If this is a hydrogen and it already has a bond, move on.
if (a1Elt==Atom::HYDROGEN && top[atom1].Nbonds() > 0 )
continue;
for (int atom2 = atom1 + 1; atom2 != stopatom; ++atom2) {
Atom::AtomicElementType a2Elt = top[atom2].Element();
double D2 = DIST2_NoImage(frameIn.XYZ(atom1), frameIn.XYZ(atom2) );
double cutoff2 = Atom::GetBondLength(a1Elt, a2Elt) + offset;
cutoff2 *= cutoff2;
if (D2 < cutoff2) {
top.AddBond(atom1, atom2);
// Once a bond has been made to hydrogen move on.
if (a1Elt==Atom::HYDROGEN) break;
}
}
}
}
# ifdef TIMER
time_within.Stop();
time_between.Start();
# endif
// ----- STEP 2: Determine bonds between adjacent residues
Topology::mol_iterator nextmol = top.MolStart();
if (top.Nmol() > 0)
++nextmol;
for (Topology::res_iterator res = top.ResStart() + 1; res != top.ResEnd(); ++res)
{
// If molecule information is already present, check if first atom of
// this residue >= first atom of next molecule, which indicates this
// residue and the previous residue are in different molecules.
if ( (nextmol != top.MolEnd()) &&
(res->FirstAtom() >= nextmol->BeginAtom()) )
{
++nextmol;
continue;
}
// If this residue is recognized as solvent, no need to check previous or
// next residue
if ( res->NameIsSolvent() ) {
++res;
if (res == top.ResEnd()) break;
continue;
}
// Get previous residue
Topology::res_iterator previous_res = res - 1;
// If previous residue is recognized as solvent, no need to check previous.
if ( previous_res->NameIsSolvent() ) continue;
// Get previous residue start atom
int startatom = previous_res->FirstAtom();
// Previous residue stop atom, this residue start atom
int midatom = res->FirstAtom();
// This residue stop atom
int stopatom = res->LastAtom();
// Check for bonds between adjacent residues
for (int atom1 = startatom; atom1 != midatom; atom1++) {
Atom::AtomicElementType a1Elt = top[atom1].Element();
if (a1Elt==Atom::HYDROGEN) continue;
for (int atom2 = midatom; atom2 != stopatom; atom2++) {
Atom::AtomicElementType a2Elt = top[atom2].Element();
if (a2Elt==Atom::HYDROGEN) continue;
double D2 = DIST2_NoImage(frameIn.XYZ(atom1), frameIn.XYZ(atom2) );
double cutoff2 = Atom::GetBondLength(a1Elt, a2Elt) + offset;
cutoff2 *= cutoff2;
if (D2 < cutoff2)
top.AddBond(atom1, atom2);
}
}
}
# ifdef TIMER
time_between.Stop();
time_total.Stop();
time_within.WriteTiming(2, "Distances within residues", time_total.Total());
time_between.WriteTiming(2, "Distances between residues", time_total.Total());
time_total.WriteTiming(1, "Total for determining bonds via distances");
# endif
if (debug > 0)
mprintf("\t%s: %zu bonds to hydrogen, %zu other bonds.\n", top.c_str(),
top.BondsH().size(), top.Bonds().size());
return 0;
}
