Welcome to Danish!

Danish (Database for machine learning & Simulation of Heme proteins) is a database of heme protein structures prepared for machine learning and molecular dynamics simulations. Biological assemblies of heme proteins in the PDB are split into protein chains, hemes and other molecules, checked for missing residues and non-standard amino acids, and provided as ready-to-use structure files together with heme geometry and sequence annotations. Missing residues are modeled with MODELLER; both the original PDB structures and the modeled structures are available.

Current data

Last update: 2026-10-15

• 7,652 PDB entries / 10,348 biological assemblies
• 45,540 protein chains (7,851 with modeled missing residues)
• 24,511 hemes (heme b 16,410 / heme c 7,318 / heme a 696 / heme o 20)

What you can do

• Browse & search: find entries by PDB ID, UniProt ID, protein name or mutation (e.g. H65V), and filter by experimental method, resolution, missing residues, non-standard amino acids and mutations.
• Entry pages: view each assembly in 3D and download structure files per chain or per assembly (PDB format), including the modeled structures.
• Heme pages: axial ligands, Fe displacement and the out-of-plane / in-plane distortion of the porphyrin by normal-coordinate structural decomposition (NSD), with the heme environment in 3D.
• UniProt pages: all chains of a protein, with mutations, insertions and deletions relative to the UniProt sequence.
• Force field: generate input files for MCPB.py and Gaussian to parameterize a heme and its coordinating residues.

Quick access

• X-ray structures at 2.0 Å or better (3,896 entries)
• Structures without mutations (3,514 entries)
• UniProt P14779 (Bifunctional cytochrome P450/NADPH--P450 reductase) (372 chains)

• A UniProt page can be opened directly at https://danish.rd.kitami-it.ac.jp/browse/entry/uniprot/P14779 (replace P14779 with the UniProt ID).
• An entry page can be opened directly at https://danish.rd.kitami-it.ac.jp/browse/entry/101m (replace 101m with the PDB ID; the extended form such as pdb_0000101m is also accepted).

Sister site

Danish is a sister site of PyDISH. For structural analysis of hemes, please use PyDISH.