We are trying to finish up the refinement on two structures, but we have run into a strange problem that I have never seen before.  So we need to ask the Phenix experts.  As background, we recently solved the structure of an 808 residue enzyme, which crystallizes as a tetramer.  In the asymmetric unit there are two tetramers (>6400 residues, ~52,000 atoms, but 67% solvent).  After SAD phasing and autobuild (many kudos to the Phenix team), we were able to trace easily the chain in two different ligand complexes, one at 2.85 Å and the other at 2.8 Å.  Although the C-terminal domains are quite well resolved (and obey 222 NCS), the first domain (~110 residues out of 808) is a bit disordered and shows some conformational differences between each monomer.  While the main chain in this domain can be traced, side chain densities are mediocre to missing in spots.  In order to ensure a consistent tracing and sequence assignment between all 8 N-terminal domain, we built in the "correct" sequence, but truncated the side chains to CB if the side chain density was poor to non-existent.  After several rounds of refinement and model building, the finished structures look quite good, have a consistent sequence assignment, and have Rwork in the high teens and Rfree in the low 20s.

THE PROBLEM: It is the 2.8 Å complex that has the problem.  We used Moleman2 to evaluate the B-factors for the 4 domains in the 8 monomers in the ASU.  When we compiled the numbers, we noticed that Moleman2 and phenix.refine gave different numbers for the protein atoms for chain B in the 2.8 Å complex (6152 versus 6289).  We checked the PDB and confirmed that there are 6152 protein atoms in chain B; the other 7 chains show no differences between the programs.  So where did phenix.refine get 6289 protein atoms when only 6152 protein atoms are input for chain B?  The number of residues are correct (780).

Adding to the confusion, there are truncated residues in each of the 8 chains.  In the phenix.refine output attached below, the PDB file is being read in. For chain A, the number of atoms and the number of residues are correct, as well as the number of incomplete residues ({'truncation_to_alanine': 14}).  However, for chain B, the number of incomplete residues ({'truncation_to_alanine': 21}) should be 37.

Any ideas?

Thanks,

Michael

phenix.refine output:

  Monomer Library directory:

    "/usr/local/phenix-1.7-650/chem_data/mon_lib"

  Total number of atoms: 51463

  Number of models: 1

  Model: ""

    Number of chains: 57

    Chain: "A"

      Number of atoms: 6289

      Number of conformers: 1

      Conformer: ""

        Number of residues, atoms: 781, 6289  <==== matches PDB file

          Classifications: {'peptide': 781}

          Incomplete info: {'truncation_to_alanine': 14}   <==== MATCHES PDB file

          Link IDs: {'PTRANS': 34, 'TRANS': 746}

          Unresolved non-hydrogen bonds: 47

          Unresolved non-hydrogen angles: 60

          Unresolved non-hydrogen dihedrals: 36

          Unresolved non-hydrogen chiralities: 7

          Planarities with less than four sites: {'PHE:plan': 1, 'GLU:plan': 3, 'ASN:plan2': 1, 'ASN:plan1': 1}

          Unresolved non-hydrogen planarities: 23

                  .

                  .

                  .

                  .

    Chain: "B"

      Number of atoms: 6236

      Number of conformers: 1

      Conformer: ""

        Number of residues, atoms: 780, 6236 <==== DOES NOT MATCH PDB file 780 residues with 6152 atoms

          Classifications: {'peptide': 780}

          Incomplete info: {'truncation_to_alanine': 21}   <====  37 truncation_to_alanine

          Link IDs: {'PTRANS': 34, 'TRANS': 745}

          Unresolved chain links: 1

          Chain breaks: 1

          Unresolved chain link angles: 3

          Unresolved chain link dihedrals: 4

          Unresolved chain link planarities: 2

          Unresolved non-hydrogen bonds: 80

          Unresolved non-hydrogen angles: 101

          Unresolved non-hydrogen dihedrals: 62

          Unresolved non-hydrogen chiralities: 9

          Planarities with less than four sites: {'GLN:plan1': 1, 'GLN:plan2': 1, 'ASP:plan': 1, 'PHE:plan': 1, 'GLU:plan': 6, 'ARG:plan': 2}

          Unresolved non-hydrogen planarities: 49