Data in Danish

This page describes how the data in Danish are prepared: which PDB entries are included, how biological assemblies, protein chains and hemes are processed, how missing residues are modeled, and what each item on the entry, heme and UniProt pages means. The data are produced by an automated pipeline from the PDB archive (via PDBj) and UniProt.

Last update of the data: 2026-10-15

1. Selection of PDB entries

All PDB entries containing at least one of the heme components HEM, HEA, HEB, HEC or HEO are included. Other porphyrins and heme analogues are not included.
No filter is applied on experimental method or resolution: X-ray, NMR, electron microscopy and other methods are all included. Experimental method, resolution, deposition date and title are taken from the PDB entry.
The data are updated incrementally: new and modified entries are processed again, and entries removed from the PDB are removed from Danish.

2. Coordinates

Only the first model of each entry is used.
Hydrogen (and deuterium) atoms are removed.
For atoms with alternate locations, the location with the highest occupancy is kept (the first one in case of a tie).
When the iron of a heme is deposited as a separate residue, it is merged into the heme if it lies at the center of the porphyrin (within 0.8 Å of the centroid of the four pyrrole N atoms, with all Fe–N distances between 1.8 and 2.5 Å).

3. Biological assemblies

Biological assemblies are generated from the assembly definitions of the PDB entry (pdbx_struct_assembly, pdbx_struct_oper_list).
Author-defined assemblies (including author-and-software-defined ones) are used. Software-defined assemblies are added only for chains that are not part of any author-defined assembly.
Chains generated by symmetry operations keep the label_asym_id of the original chain with a suffix: the first copy keeps the original ID, and further copies are named A-1, A-2, … The auth_asym_id is not changed.
Assemblies with more than 30 chains are shown without a 3D viewer on the entry page; their hemes can be viewed on the heme pages.

Items on the entry page (assembly)
Structural data	PDB: no chain of the assembly needed modeling. modeled: the chains that needed modeling were modeled. including_not_modeled: some chains could not be modeled (see Modeling).
Oligomeric count	Number of chains in the assembly, as given by the PDB entry (with the type of assembly definition).
label_asym_ids of protein / heme / the others	Chains in the assembly. Protein chains are polymer chains with more than 10 residues; all other molecules (ligands, short peptides, nucleic acids, waters) are listed as "the others".

4. Protein chains

Each protein chain is checked for the following features (shown in "Each chain information" on the entry page and on the UniProt pages):

Missing residues	A chain has missing residues when its residue numbering (`auth_seq_id`) jumps, or when a residue lacks backbone atoms (N, CA or C). Residues missing at the N- or C-terminus are not counted. On the entry page, the residues that were rebuilt by modeling are listed.
Structural gaps	A chain break confirmed by geometry: the distance between C of one residue and N of the next exceeds 2.5 Å.
Nonstandard amino acids	Residues other than the 20 standard amino acids in the chain (e.g. MSE), listed by residue name.
Unknown residues	The chain contains unknown residues (UNK) or ambiguous ASX.
S-S bonds	Disulfide bonds within the chain (residue numbers in pairs).
Residue indices (UniProt)	First and last residue numbers of the chain in the structure (`auth_seq_id`), and the corresponding range of the UniProt sequence.
label_asym_ids of heme	Hemes coordinated by the chain (the chain provides an axial ligand). For hemes without an axial ligand, the neighboring protein chains are listed.
Available structure	PDB: only the deposited structure. PDB and Modeled: a modeled structure is also available (see below). PDB (not modeled): the chain needed modeling but could not be modeled.

On the browse page, "Missing residues / Structural gap" and "Non-standard amino acids" of an entry are Exists / Included only when every assembly of the entry has them. "Structural data" of an entry is available when its chains need no modeling or were modeled.

5. Modeling of missing residues

Chains with missing residues, structural gaps or nonstandard amino acids are rebuilt as follows. Both the deposited and the modeled structures are provided.

Completion of side chains. Missing heavy atoms of incomplete residues are added with pdb4amber (AmberTools) for all chains.
Modeling. Missing segments inside the chain are built with MODELLER (automodel, one model, fast refinement), using the sequence of the PDB entry (entity_poly, canonical sequence) as the target. Nonstandard amino acids are replaced by their standard parent residues. Hemes in the assembly are included as rigid bodies so that the model does not overlap them; neighboring chains are included for large assemblies.
- Only internal segments are modeled; missing N- and C-terminal residues are not added.
- Segments of 20 residues or longer are not modeled.
- Chains with unknown residues, or whose observed residues differ from the sequence of the entry, are not modeled.
Grafting. Only the modeled residues are taken from the MODELLER model (superposed on the flanking residues); all other residues keep their experimental coordinates.
Energy minimization. The chain is minimized with GROMACS (AMBER99SB-ILDN, steepest descent, in vacuum) with position restraints on heavy atoms (5000 kJ mol⁻¹ nm⁻² for experimental residues, 1000 for modeled residues). Residues that coordinate a heme keep their experimental coordinates.

RMSD of experimental residues after energy minimization	Heavy-atom RMSD of the residues that were not modeled, between the structure before and after energy minimization (after superposition). It shows how much the experimental part moved during minimization; it is not a measure of the accuracy of the modeled residues.

6. Hemes

Each heme in each assembly is analyzed. A heme must have the iron and all 24 atoms of the porphyrin core; incomplete hemes are not analyzed.

heme type	`HEA`, `HEC` and `HEO` are heme a, heme c and heme o. For `HEM` and `HEB`, the type is assigned from the substituents of the porphyrin (vinyl, formyl, hydroxyethylfarnesyl and thioether links to cysteine); for example, a `HEM` covalently bound to cysteines is classified as heme c. Hemes that do not match a known pattern are Unknown.
Axial ligand(s)	Residues with an atom within 3.1 Å of the iron. Protein residues are searched first; other molecules (water, CO, O₂ and other small molecules) are considered only when fewer than two protein residues are found. Each residue is counted once (symmetry copies of a chain are distinguished). Up to two ligands are recorded, each with the distance from the iron to its nearest atom (Coordinate distance). When there are two, an amino acid is listed first.
Fe displacement	Distance of the iron from the mean plane of the four pyrrole nitrogen atoms (Å). The value has no sign (it does not indicate the proximal or distal side).
Neighbor chains	Protein chains with a Cα atom within ±8.5 Å of the iron in each of x, y and z (a cube centered on the iron), and other molecules with any atom within ±5.0 Å. The viewer on the heme page shows the heme with these chains and molecules. For assemblies with more than 30 chains, the whole protein chains containing the heme are shown in a second viewer.

7. Heme distortion (NSD)

The distortion of the Fe-porphyrin is described by normal-coordinate structural decomposition (NSD), computed as in PyDISH (doi:10.1093/database/baaa066).

The iron and the 24 atoms of the porphyrin core (25 atoms; the iron is included, unlike the usual 24-atom NSD) are superposed on a planar reference heme (least-squares fit), and the displacement from the reference is projected onto 12 normal modes (minimal basis; the mode vectors are orthonormal). Values are in Å.
Out-of-plane modes: sad (saddling), ruf (ruffling), dom (doming), wvx, wvy (waving x, y) and pro (propellering).
In-plane modes: mst, nst (meso- and N-stretching), trx, try (translation x, y), bre (breathing) and rot (rotation). The in-plane "translation" and "rotation" are internal deformations of the porphyrin, not rigid-body motions.
The sign of each value depends on the orientation of the heme given by its atom names (rings A–D); it does not correspond to the proximal or distal side. Compare magnitudes when the sign is not meaningful.
On the heme page, the mean and standard deviation over all hemes in Danish are shown for comparison; values beyond one or two standard deviations are highlighted in the chart.

8. UniProt and sequence differences

The UniProt entry of each chain is taken from SIFTS. Protein name, organism, length and sequence are taken from UniProt (an obsolete UniProt entry is shown as inactive).
The sequence of the PDB entry (entity_poly, canonical sequence: expression tags included, modified residues given as their parent amino acids) is aligned to the UniProt sequence (global alignment; match +2, mismatch −1, gap opening −10, gap extension −0.5).

Mutations (UniProt numbering)	Substitutions, written as UniProt residue – UniProt number – residue in the PDB entry (e.g. `F47V`). On the entry page, each mutation can be highlighted in the 3D viewer.
Insertions	Residues of the PDB sequence not present in UniProt, written after the preceding UniProt number (e.g. `0insMGSSHHHHHH`: an N-terminal tag; `463insLEHHHHHH`).
Deletions	UniProt residues not present in the PDB sequence, including the termini (e.g. `M1-K20`). Residues that are only missing from the coordinates are not deletions.
Sequence identity to UniProt	Fraction of identical residues among the aligned residues.
coverage	Fraction of the UniProt sequence that is present in the PDB sequence.

For cytochrome P450 BM3 (UniProt P14779), positions follow the conventional numbering without the initial methionine (UniProt number − 1), e.g. F87A.

9. Structure files

Structure files can be downloaded per chain or per assembly from the entry page (PDB format). The files of an assembly are named {PDB ID}_{type}_{label_asym_id}_{auth_asym_id}.pdb:

_protein_	Protein chain (deposited structure, completed side chains). No hydrogen atoms.
_protein_…_modeled	Modeled and energy-minimized protein chain (see Modeling). No hydrogen atoms.
_heme_	Heme.
_axialwater_	Water coordinated to a heme iron (Fe–O ≤ 2.8 Å), separated from the other waters.
_water_	Waters.
_others_	Other molecules (ligands, ions, short peptides, nucleic acids, …).

Residue numbers are those of the PDB entry (auth_seq_id); modeled residues are numbered consecutively.
In the protein files, residues that coordinate a heme are renamed for use with AMBER force fields (histidine HID, cysteine CYM); modeled files may contain AMBER residue names.
The PDB format allows one-character chain IDs; the chain IDs in the files therefore do not always match auth_asym_id. Use the file names to identify chains.
Molecules that occupy the same position as another one (e.g. overlapping copies of a heme generated by symmetry) are left out of the files.

The 3D viewers use the coordinates of the whole assembly (mmCIF) and, on the heme pages, of the heme with its neighbor chains.

10. PDB IDs

Danish uses the extended PDB ID format (e.g. pdb_0000101m). Classic four-character IDs are also accepted in the search box and in URLs (e.g. /browse/entry/101m) and are redirected to the extended form.