|
Recently protein structure bioinformatics has seen a big upsurge in ab initio structure prediction from multiple sequence alignment analyses through correlated mutation analyses. We believe that the definition of what constitutes a contact is one of the (many) problems still to be addressed in this field. Therefore a large number of contact definitions has been used here. These included Cα-Cα distances, actual atomic contacts of either side chain atoms or any atoms; center of mass distances, salt-bridges, etc. |
Be aware that we (deliberately) do not do much filtering here. For example, covalently bound neighbours tend to be listed as contacts too. We have shown (see refs at the bottom) that filtering by a whole series of terms (including covalent distance along the chain) can significantly improve the quality of contact predictions, but we will leave it up to you to make these decisions.
| Code | Type of data |
| cc1 | Cα-Cα distance < 12.5 Ångström |
| cc2 | Residue spheres closer than 0.25 Ångström |
| cc3 | Residue spheres closer than 2.5 Ångström |
| cc4 | Residues with an atom pair closer than 0.25 Ångström |
| cc5 | Residues with an atom pair closer than 2.5 Ångström |
| cc6 | Residues with a side chain atom pair closer than 0.25 Ångström |
| cc7 | Residues with a side chain atom pair closer than 2.5 Ångström |
| cc8 | Residues in different chains with a side chain atom pair closer than 2.5 Ångström |
| cc9 | Cβ-Cβ distance < 12.5 Ångström |
For each residue the distance of its Cα is given to the Cαs of all other residues, but only when this distance is shorter that 12.5 Ångström. We call every pair of residues fulfilling this criterion a ′contact′.
From left to right each line holds the sequential number of the residue in the PDB file, the residue type, in brackets the residue number and insertion code found in the PDB file, the chain identifier of the chain in which this residue is found; and after the minus sign this information is repeated for the contact partner residue. The last column holds the Cα - Cα distance in Ångströms. Be aware that each contact is only listed once.
Both residues must at least have an intact Cα atom. If the residue in the left-hand residue-column doesn't have an intact Cα then that gets indicated; if the residue in the right-hand residue column doesn′t have an intact Cα then it gets rejected and is not listed in the table.
For each residue the smallest sphere is determined that encompasses all atoms in that residue. If a pair of residues have less than 0.25 Ångström distance between the surfaces of these pseudo spheres, we call this pair of residues a contact.
From left to right each line holds the sequential number of the residue in the PDB file, the residue type, in brackets the residue number and insertion code found in the PDB file, the chain identifier of the chain in which this residue is found; and after the minus sign this information is repeated for the contact partner residue. Be aware that each contact is only listed once.
Both residues must be completely intact, i.e. all atoms must be present and no other pathologies should be present. If the residue in the left-hand residue column isn′t completely intact then that gets indicated; if the residue in the right-hand residue column isn′t completely intact; then it gets rejected and is not listed in the table.
The ′cc3′-Lists entries are similar to the ′cc2′ entries except that the ′cc2′ Lists entries hold contact pairs when the space between the pseudo spheres is less than 0.25 Ångström while the ′cc3′ entries use a cut-off of 2.5 Ångström.
The ′cc4′-Lists entries are similar to the previous few Lists entries except that a pair of residues in a ′cc4′ table is called in contact if they share a pair of atoms (so one atom in the one residue and the other atom in the other residue) that are so close that there is less than 0.25 Ångström space between their Van de Waals′ surfaces.
The ′cc5′-Lists entries are similar to the ′cc4′-Lists entries except that ′cc5′-Lists entries were made with a cut-off of 2.5 Ångström.
The ′cc6′-Lists entries are similar to the ′cc4′-Lists entries except that ′cc6′-Lists entries were made looking only at side chain atoms. So contacts with or between backbone atoms are neglected.
The ′cc7′-Lists entries are similar to the ′cc5′-Lists entries except that ′cc7′-Lists entries were made looking only at side chain atoms. So contacts with or between backbone atoms are neglected.
The ′cc8′-Lists entries are similar to the ′cc5′-Lists entries except that ′cc8′-Lists entries only hold contacts between protein chains.
For each residue the distance of its Cβ is given to the Cβs of all other residues, but only when this distance is shorter that 12.5 Ångström. We call every pair of residues fulfilling this criterion a ′contact′. The output format is similar to the ′cc1′-Lists.