CAPRI scoreset v2022
Introduction Browse Download Help

Column Name Description
1 target_id Unique identification of the target/interface combination
2 identification Identification of the decoy; note that the same identification may occur over multiple interfaces of the same target
3 round CAPRI Round where the target was offered
4 target Target ordinal
5 interface Interface ordinal
6 p_id Predictor ID used for this Round and target
7 p_type Predictor type; can be P for Predictors, U for Uploaders, and S for Scorers
8 p_rank Predictor ordinal
9 name Name of the PI of the Predictor Group
10 uid Unique predictor ID; this ID is invariant over the entire scoreset database
11 model Rank of the submitted model
12 e Boolean indicating whether the model was included in the clash calculation
13 nclash Number of clashes in the model
14 threshold Clash threshold above which a model is disqualified
15 fnat Fraction of native contacts
16 fnonnat Fraction of non-native contacts
17 f1 F1, calculated from fnat and fnonnat
18 dockq DockQ score, calculated from fnat, lrms and irms
19 lrms Ligand RMS
20 irms Interface RMS
21 srms Side-chain RMS
22 ia Interface area (Ų)
23 dL Ligand displacement (Å)
24 thetaL Ligand misorientation angle (degrees)
25 nchain Not used
26 (*_coverage) Residue coverage vs the target; can be above 100 due to unresolved residues in the target structure
27 fIR_L Ligand interface recall based on loss of accessibility
28 fIR_R Receptor interface recall based on loss of accessibility
29 fOP_L Ligand interface overprediction based on loss of accessibility
30 fOP_R Receptor interface overprediction based on loss of accessibility
31 mrms_L Ligand molecular RMS
32 mrms_R Receptor molecular RMS
33 classification CAPRI classification of the model
34 main_classification CAPRI classification, regardless of disqualification
35 stars CAPRI classification in terms of number of stars attributed (0=incorrect; 3=high)
36 clashes Boolean indicating disqualification due to too many clashes
37 low_id Boolean indicating disqualification due to too low sequence identity with the target
38 md5 Unique MD5 code indicating the origin of the Scorer-submitted model
Note that not all these may be usefull in a context outside of the CAPRI experiment

The CAPRI scoreset v2022 is developed and maintained by:

  • Marc F. Lensink, CNRS & University of Lille, France
  • Theo Mauri, CNRS & University of Lille, France
  • Guillaume Brysbaert, CNRS & University of Lille, France
  • Shoshana J. Wodak, VUB-VIB, Belgium