Add molecules collapses structure as shown

[shyamramesh]

I keep attaching molecules, and as shown the molecule structure gets collapsed. How do i fix this.

I am a beginner, need help

[greglandrum]

Hi @shyamramesh, we can't help if you don't tell us exactly what you did. Please provide example code which demonstrates the problem.

[shyamramesh]

@greglandrum my colleague has shared the code that indicates what we do. Adding hexane to another hexane to create honeycomb like structure , i get this collapsed

code JSMol::attach_molecule( const JSMol &other){ ..} as shown below

[shyamramesh]

Any suggestions on what we are missing to correct?

[Suganesh-repo]

Hi @greglandrum

hereby, i am sharing the code ..

void JSMol::attach_molecule( const JSMol &other){
//d_mol is molecule1, other is molecule2 --comments...
RWMol *rwm_1 = new RWMol(*d_mol);
RWMol *rwm = new RWMol(*combineMols(*d_mol, *other.d_mol));
vector sites_1 = [4,5] //for example, attachment points(selected atoms - 4,5)
vector sites_2 = [0.1 //for example, attachment points(selected atoms - 0,1)

rwm->beginBatchEdit();

int i = 0;
for( int idx : sites_1){
Atom *a = rwm->getAtomWithIdx(idx);

ROMol::ADJ_ITER_PAIR ngbr = rwm->getAtomNeighbors(a);
for (const auto &nbri : boost::make_iterator_range(ngbr)){
  const auto &neighbor = (*rwm)[nbri];
  if (list_contains(sites_1, neighbor->getIdx() )) 
    continue;
  Bond *b = rwm->getBondBetweenAtoms( a->getIdx(), neighbor->getIdx() );
  int bondNum = rwm->addBond( neighbor->getIdx(), sites_2.at(i), b->getBondType());
}
i++;
rwm->removeAtom(idx);

}

rwm->commitBatchEdit();
normalize_structure(*rwm,*rwm_1);
d_mol = std::unique_ptr(rwm);
}

void normalize_structure(ROMol &mol, ROMol &temp){

RDDepict::ConstrainedDepictionParams ps;
ps.adjustMolBlockWedging = false;
ps.acceptFailure = true;
ps.forceRDKit = true;
RDDepict::generateDepictionMatching2DStructure(mol,temp, -1, NULL, ps);
MatchVectType atomMap;
if (RDKit::SubstructMatch(mol, temp, atomMap)) {
RDKit::Conformer &conf = temp.getConformer();
RDGeom::INT_POINT2D_MAP coordMap;
for (auto mv : atomMap) {
RDGeom::Point3D pt3 = conf.getAtomPos(mv.first);
RDGeom::Point2D pt2(pt3.x, pt3.y);
coordMap[mv.second] = pt2;
}
RDDepict::compute2DCoords(mol, &coordMap);
}
}

[Suganesh-repo]

Can someone help us fix this issue?
@greglandrum @DavidACosgrove

[Suganesh-repo]

Team, can anyone help us fix this?

@greglandrum @cowsandmilk @DavidACosgrove

[cowsandmilk]

Tagging random people who haven't worked in computational chemistry for over half a decade is not how to get your issue solved.

[bp-kelley]

@Suganesh-repo you are hitting some limitations in the rdkit depiction code, you can try using coordgen:

from rdkit.Chem import rdDepictor
rdDepictor.SetPreferCoordGen(True)

But also your code is flawed, you are making a new template but not actually using it. Try using this:

def _bondnormalize(mol):
    temp = template_mol #this is mol1 before attachment..

    params = Chem.AdjustQueryParameters()
    params.makeBondsGeneric=True
    params.makeAtomsGeneric=True
    tq = Chem.AdjustQueryProperties(temp, params)
    AllChem.GenerateDepictionMatching2DStructure(mol, tq, acceptFailure=False)

    mcs = rdFMCS.FindMCS([temp, mol])
    patt = Chem.MolFromSmarts(mcs.smartsString)
    query_match = mol.GetSubstructMatch(patt)
    template_match = temp.GetSubstructMatch(patt)
    AllChem.AlignMol(mol, temp, atomMap=list(zip(query_match,template_match)))
    return mol

[DavidACosgrove]

Your first molecule makes no sense at all, chemically, having as it does several hexavalent carbons. I don't think it's reasonable to expect a good picture of it.
The second molecule is at least chemically correct. However, adding another cyclopentyl ring would normally put atoms on top of those already there (29 and 30). The drawing code is doing its best to avoid that, and that's why you're getting the odd layout for the last ring. You might do better using coordgen instead of the default RDKit layout routines. (rdDepictor.SetPreferCoordGen(True), off the top of my head). But it is always going to be difficult, and there is no obvious solution.
Incidentally, the second situation with a circle of rings gives rise to an interesting kind of chirality. The molecule in real life has exactly the same problem as your drawing - when enough rings are added, atoms come very close to each other, but in the physical case it can use the 3rd dimension and the last ring can go above or below the first ring. When the rings are benzene rather than your cyclopentanes it's called a helicene.