Geometry optimisation fails with MMFF94

Hello everyone,

I’ve created a KNIME workflow that generates conformers, runs a geometry optimisation and keeps the lowest energy conformer:
MarvinSketch => RDKit from Molecule => RDKit Add Conformer => RDKIT Optimize Geometry

However, when I run the geometry optimisation on Bodipys I run into problems with the MMFF94 and MMFF94s force fields:

Force field creation failed. Creating empty output.

UFF works but it doesn’t always give the optimal conformer. My feeling is that the charges or the non-standard bond order (quaternary borate or nitrogen cation?) don’t go along with the MMFF force fields.

I’ve tried neutral representation as well as many different sanitization steps - but no luck.
Has anyone had a similar experience and, ideally, come across a solution?

Many thanks!

sample SMILES: F[B-]1(F)N2C(\C=C\C3=CC=CC=C3)=CC=C2C(=C2C=CC(\C=C\C3=CC=CC=C3)=[N+]12)C1=CC=CC=C1

Well, I figured out that Boron is not supported by MMFF94. That at least explains that part.
I’ll look for a work-around, maybe by either atom replacement or by setting dihedral angles. Thanks for reading.

1 Like