Table 1 Examples of quality indicators pointing to potential problems in PDB entries, calculated using our automatic evaluation pipeline

From: Making the invisible enemy visible

Key indicators in evaluation

Number of depositions (Percentage of total)

X-ray crystallography (999 depositions)

Completeness <80%

14 (1.4%)

Rfree >35%

2 (0.2%)

Potential twinning

52 (5.2%)

Contaminated by ice diffraction

93 (9.3%)

Incorrect mask

86 (8.6%)

Single-particle cryo-EM (360 depositions)

Average model–map FSC <0.4

46 (13%)

MI score <0.4

56 (16%)

SMOC >10%

64 (18%)

Other indicators (1,392 depositions, including 31 NMR and 2 neutron diffraction)

CaBLAM outlier conformations >2.0%

318 (23%)

CaBLAM severe Cα outliers >1.0%

122 (8.8%)

Sequence mismatch

23 (1.7%)

  1. Potential twinning was identified by L-Test7, and ice diffraction and incorrect mask were identified by visual inspection of AUSPEX plots8. The chosen cutoffs for FSC and MI score16 indicate poor overall agreement between the map and model. A SMOC15 score of >10% indicates that more than 10% of the residues of a structure fit poorly with the map and could potentially be improved.