Try to re run the command by first starting meeko and rdkit . to convert PDB files to PDQT files. It can help vina recognize the atom . assuming B for Boron atom, autodock vina cannot handle boron, due to the scoring function.

The existing answer is pointing at the key issue: this is a PDBQT atom-type problem, not a docking-box or command-line problem. Vina is stopping because the last column of one of the ligand PDBQT atom records is B, and standard AutoDock/Vina atom types do not include boron.

A quick check is:

grep -n " B$" ligand.pdbqt
grep -n "^ATOM\|^HETATM" ligand.pdbqt | head

Look at the final field of the ATOM / HETATM lines. That is the AutoDock atom type Vina reads. If the atom is actually boron, this is not just a formatting error: standard Vina does not have a normal boron parameter/scoring term, so the result would not be chemically reliable.

I would avoid "fixing" this by renaming B to C, A, Br, etc. That may make the file parse, but then the score and pose no longer represent the ligand you think you are docking. First confirm whether the atom is genuinely boron or whether it came from a bad conversion, for example a bromine atom or another element being written incorrectly during ligand preparation.

If the ligand really contains boron, the safer options are:

1. Use a docking/scoring setup that explicitly supports the chemistry you need.
2. Parameterize and validate a custom AutoDock/Vina-compatible setup before trusting the results.
3. If this is only exploratory, document the limitation clearly and do not interpret the Vina affinity as evidence of activity.

Regenerating the ligand PDBQT with Meeko/RDKit or another chemistry-aware preparation workflow is still a good step, but it only helps if the original structure, bond orders, protonation state, and element identities are correct. It will not make standard Vina support boron in a scientifically meaningful way by itself.

I work on ProteinIQ's AutoDock Vina tool, where we try to surface this kind of file-preparation issue before a run. But the main point here is independent of the platform: do not force the PDBQT to pass by assigning an arbitrary atom type unless you are comfortable with the chemistry limitation that introduces.