When atomic contacts are displayed in IMGT/3Dstructure-DB between peptide and MHC or between TR and pMHC, are they as stated in the original papers or are they derived data from an IMGT algorithmic analysis on the structural information?
All the atom contacts information in IMGT/3Dstructure-DB are directly
measured in the 3D structures with an algorithm developed at
IMGT. Using the same algorithm is a great advantage because it
allows to compare the atom contact analysis from different
structures.
How are the contacts differentiated between what is actually described in the original papers and the ones that are results of the IMGT analysis?
The atom contacts that are described in the original papers do not appear
on the IMGT website because they are an interpretation of the real
data, the 3D structure, and because they are often incomplete and not
comparable (manual identification or use of different software). The
IMGT algorithm is very simple, two atoms (and so corresponding
positions) are considered in contact when they are less than 5
angstroms apart(considering the center of the atoms). When the two
atoms are oxygen and/or nitrogen an atom contact is labelled as
"polar". A "polar" contact can be ultimately
identified as an hydrogen bond based on a 3.2 angstroms distance
cut-off and a couple of angular constraints.
What is the difference between the peptide interactions (with MHC) displayed via the peptide amino acid "IMGT Residue@Position cards" and through the IMGT Collier de Perles with "IMGT pMHC contact sites" (i.e the list of C1 to C11 contact sites)?
The C1 to C11 are standardized sites that emulates the pockets concept. But that shouldn't make the contact sites different for lets say P1 in these two different displays.
There are quite a few differences between the "IMGT pMHC contact sites"
and the classical contact analysis. The peptide amino acid "IMGT
Residue@Position" cards present all the contacts of the
positions. The "IMGT pMHC contact sites" card displays the
main interactions at the interface and is therefore based on an
analysis of all the contacts. With the "IMGT pMHC contact
sites", one MHC contacting position is assigned to one peptide
position. The "IMGT pMHC contact sites" displays the MHC
positions the most in contact with a given peptide position at the
pMHC interface. The selection of a peptide position partner is based
on a scoring scheme that gives a different score to hydrogen bonds,
polar contacts and van der Waals contacts. A score cut-off allows to
remove the MHC positions that are poorly involved at the interface.
The "pMHC contact sites" analysis was perfomed on a great
number of pMHC 3D structures to define the MHC positions the most
frequently involved at the interface. The comparison of the resulting
analysis.
(Kaas and Lefranc, In Silico Biology, 2005, table 3,
http://www.bioinfo.de/isb/2005/05/0046/)
with a peculiar 3D structure is very useful to spot the contacts that are unusual.
When there are two crystallographic complexes then is the contacts analysis based on one set of chains (i.e one single complex)? How did you handle these situations and is there a caution of some kind to be aware of?
The coordinate files found in the PDB contains the crystallographic unit
cell which can contain one (or several) crystallographic complex(es).
The unit cell is the smallest cell of the crystal that cannot be reproduced
by rotation and/or translation of a smaller cell. Its a purely
physical description that has nothing to do with the biological
significance of the protein complexes found in the unit cell. The
distinction between biological and non-biological is not trivial.
Most of the time we used the number of atom contacts to identify the
crystallographic complexes (noted CC on the IMGT/3Dstructure-DB
website), annotators are also using their biological knowledge. The
atom contact analysis presents all the interfaces, biologically
relevant or not. The table on the top of the page helps to choose the
relevant domain interactions.
Are only V-region alleles that are "F", "(F) " or "[F]" assigned to proteins in IMGT/3Dstructure-DB?
Yes, only alleles which are F, (F) or [F] are assigned to proteins in IMGT/3Dstructure-DB. Alleles which are ORF are not assigned.
How to display the genes and alleles names of the individual chains?
In the "Display results" section (at the top of the IMGT/3Dstructure-DB query page), select "Overview" radio button and "with IMGT allele(s) name".
