Meeko ligand preparation fails on certain molecules with terminal triple bonds

[Chryzl]

Bug Description

Meeko ligand preparation crashes with ValueError: list.remove(x): x not in list when processing molecules containing certain terminal triple bonds. The failure occurs during the MoleculePreparation.prepare() step and is caused by a bond canonicalization bug in the BondTyperLegacy class.

Root Cause

The error originates in meeko/bondtyper.py at line 45 in the BondTyperLegacy.__call__() method during the single_triple_single processing loop:

triple_bonds = [(x[0], x[1]) for x in setup.find_pattern("[*]#[*]")]
single_triple_single = [
    (x[0], x[1], x[2], x[3]) for x in setup.find_pattern("[*]-[*]#[*]-[*]")
]
single_to_rigidify = []
for i, j, k, m in single_triple_single:
    triple_bonds.remove((j, k))  # ❌ BUG: Bond not canonicalized
    single_to_rigidify.append((i, j))
    single_to_rigidify.append((k, m))

The problem occurs because:

1. triple_bonds: [(0, 1)]
2. single_triple_single: pattern matching returns bonds in their original order from the SMARTS pattern [*]-[*]#[*]-[*] --> [(2, 1, 0, 21)]
3. When trying to remove (1, 0) from triple_bonds, it's not found in the list, causing the ValueError

Minimal Reproduction

from rdkit import Chem
from scrubber import Scrub
from meeko import MoleculePreparation

# SMILES with terminal triple bond that triggers the bug
smiles = "C#CCCOC(=O)N1C[C@@H](c2nc(C)n[nH]2)[C@H](C2CC2)C1"

# Ligand preparation pipeline
mol = Chem.MolFromSmiles(smiles)
scrub = Scrub(ph_low=7.4, ph_high=7.4, skip_tautomers=True)
mol_states = list(scrub(mol))

# This fails during bond typing
mk_prep = MoleculePreparation()
for mol_state in mol_states:
    molsetup_list = mk_prep.prepare(mol_state)  # ❌ Crashes here

Error:

File "minimal_triple_bond_bug.py", line 14, in <module>
    molsetup_list = mk_prep.prepare(mol_state)
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/chem2017/Meeko/meeko/preparation.py", line 565, in prepare
    self._bond_typer(
  File "/home/chem2017/Meeko/meeko/bondtyper.py", line 45, in __call__
    triple_bonds.remove((j, k))
ValueError: list.remove(x): x not in list