Naming for evaluation results follows the format:
Txx_Pyy.Mzz
, where
Txx is the target ID,
Pyy is the participant ID, and
Mzz is the MODEL ID of the submission
In addition, three further structures are evaluated:
capri_xx_xray
the target, with xx the target ID
unbound
the unbound ligand and receptor structures, fitting their invariant parts onto the target
interface
the interface ligand and receptor structures, fitting their interface parts onto the target
The summary displays a logfile for the evaluation. A lot of it is irrelevant information and you could even safely ignore this file. However, some useful information is:
using a sequence alignment, the prediction chains are fitted onto the target chains, this definition is then used throughout the rest of the evaluation
for each target chain, first ligand then receptor, 4 sequence strings are given:
lists the residues that are part of the interface
full sequence for the target
uppercase residues are conserved throughout all predictions, lowercase residues have backbone presence but are not conserved
receptor residues that are invariant to conformational changes between unbound and bound structures
The C and S sequence strings are used to filter out those residues that are not present or conserved throughout the predictions and the ones that show a large conformational change between target unbound and bound receptor molecule(s), see settings
f(nat) is the fraction of ligand-receptor contacts that is also found in the native (target) structure
f(non-nat) is the fraction of ligand-receptor contacts that is found, but that is not present in the native (target) structure
A list of ligand-receptor contacts and whether they are native (1) or non-native (0) is supplied
f(IR) is the fraction of interface residues for the prediction that match the interface of the target, based on solvent-accessibility
f(OP) is the overprediction, or fraction of interface residues for the prediction that do not match the interface of the target, based on solvent-accessibility
A list of interface residues and whether they are native (1) or non-native (0) is supplied. Note that this list is based on the distance-definition of the interface, and one can not calculate f(IR) and f(OP) from it
Both f(IR) and f(OP) are split in a receptor and a ligand part
IA is the interface solvent accessible surface area for the complex, where the interface area is defined as
d(L) is the distance between the backbone-only geometrical centers or centers-of-mass of target and prediction ligand molecule(s), after a LSQ RMS fit of either the receptor interface, or the receptor conformational change invariant residues backbone (see settings)
A PDB file with the prediction structure on top of the target structure (MODEL 0) is supplied under L_rmsd, with the geometrical or mass-weighted centers in MODEL 9999, chain Z, residues XYZ 9998 (target) and XYZ 9999 (prediction)
L_rmsd is the RMSD between target and prediction ligand molecule(s) after the same LSQ RMS fit as was done in the calculation of d(L)
The fitted structures are supplied in a PDB file
Please note that links to PDB files will only work once the coordinates of the target have been released
I_rmsdbb is the interface RMSD after a backbone-only LSQ fit of the prediction onto the target interface residues
The fitted structures are supplied in a PDB file
I_rmsdsc is the interface RMSD after a side-chain-only LSQ fit of the prediction onto the target interface residues
theta(L) is the angle mismatch between prediction and target ligand molecule(s), after a superimposed LSQ fit
seq ID is the lowest sequence identity between any matching target and prediction ligand or receptor chain
M_rmsd is the molecular RMSD after a LSQ fit of both the conformational change invariant and the interface residues. Values for both ligand and receptor are given
The fitted structures are supplied in a PDB file
The classification is the final classification of the submitted model applying the usual criteria of f(nat), L_rmsd and I_rmsdbb, filtering for clashes and sequence identity
The source is only listed for scorer submitted files and indicates the uploader model that the revised submission was based upon