Extra oxygen atom in glycan prediction.

[wyku]

sequences:
    - protein:
        id: A
        sequence: SPDPNWELLSSLGEYKDINLESSNASNITYDLEKYKNLDEGTIVVRFNSKDSKIQSLLGISNSKTKNGYFNFYVTNSRVGFELRNQKNEGNTQNGTENLVHMYKDVALNDGDNTVALKIEKNKGYKLFLNGKMIKEVKDTNTKFLNNIENLDSAFIGKTNRYGQSNEYNFKGNIGFMNIYNEPLGDDYLLSKTGETK
        msa: msaA.a3m
    - ligand:
        id: B
        ccd: ["GAL", "SIA"]
constraints:
    - bond:
        atom1: [B, 1, O6]
        atom2: [B, 2, C2]

When I run boltz2 to predict glycan-protein complex structure, the final structure have an extra O2 atom. Is there any solution to avoid this?

Thank you!