Table 2 Evaluated metrics
From: Cryo-EM model validation recommendations based on outcomes of the 2019 EMDataResource challenge
Metric class | Package metric definition |
|---|---|
Fit-to-Map | |
Correlation Coefficient, all voxels | Phenix CCbox full grid map versus model-map density correlation coefficient19 TEMPy CCC full grid map versus model-map density correlation coefficient17 |
Correlation Coefficient, selected voxels | Phenix CCmask map versus model-map density, only modeled regions19 Phenix CCpeaks map versus model-map density, only high-density map and model regions19 TEMPy CCC_OV map versus model-map density, overlapping map and model regions18 TEMPy SMOC Segment Manders’ Overlap, map versus model-map density, only modeled regions18 |
Correlation Coefficient, other density function | TEMPy LAP map versus model-map Laplacian filtered density (partial second derivative)16 TEMPy Mutual Information (MI) map versus model-map Mutual Information entropy-based function16 TEMPy MI_OV map versus model-map Mutual Information, only modeled regions18 |
Correlation Coefficient, atom positions | Chimera/MAPQ Q-score map density at each modeled atom versus reference Gaussian density function8 |
FSC | Phenix FSC05 Resolution (distance) of Map-Model FSC curve read at point FSC = 0.5 (ref. 19) CCPEM/Refmac FSCavg FSC curve area integrated to map resolution limit13,42 |
Atom Inclusion | EMDB/VisualAnalysis AI all Atom Inclusion, percentage of atoms inside depositor-provided density threshold14 TEMPy ENV Atom Inclusion in envelope corresponding to sample molecular weight; penalizes unmodeled regions16 |
Sidechain Density | Phenix EMRinger evaluates backbone by sampling map density around Cɣ-atom ring paths for nonbranched residues15 |
Coordinates-only | |
Configuration | Phenix Bond r.m.s.d. of bond lengths21 Phenix Angle r.m.s.d. of bond angles21 Phenix Chiral r.m.s.d. of chiral centers21 Phenix Planar r.m.s.d. of planar group planarity21 Phenix Dihedral r.m.s.d. of dihedral angles21 |
Clashes | MolProbity Clashscore Number of steric overlaps ≥0.4 Å per 1,000 atoms20 |
Conformation | MolProbity Rotamer sidechain conformation outliers20 MolProbity Rama Ramachandran ɸ,ψ main chain conformation outliers20 MolProbity CaBLAM outliers CO and Cɑ-based virtual dihedrals22 MolProbity Calpha outliers Cɑ-based virtual dihedrals and Cɑ virtual bond angle22 |
Versus Reference Model | |
Atom Superposition | Local Global Alignment (LGA) GDT-TS Global Distance Test Total Score, average percentage of model Cɑ that superimpose with reference Cɑ, multiple distance cutoffs23 LGA GDC Global Distance Calculation, average percentage of all model atoms that superimpose with reference, multiple distance cutoffs23 LGA GDC-SC Global Distance Calculation for sidechain atoms only23 OpenStructure/QS CaRMSD r.m.s.d. of Cɑ atoms25 |
Interatomic Distances | LDDT LDDT Local Difference Distance Test, superposition-free comparison of all-atom distance maps between model and reference24 |
Contact Area | CAD CAD Contact Area Difference, superposition-free measure of differences in interatom contacts26 HBPLUS50 HBPR > 6, hydrogen bond precision, nonlocal. fraction of correctly placed hydrogen bonds in residue pairs with >6 separation in sequence |
Comparison among models | |
Atom Superposition, Multiple | DAVIS-QA average of pairwise LGA GDT-TS scores among submitted models27 |
