2019 Model Metrics Challenge Summary Statistics

Challenge Target

T1 Apoferritin 1.8 Å (EMD-20026) 16
T2 Apoferritin 2.3 Å (EMD-20027) 15
T3 Apoferritin 3.1 Å (EMD-20028) 15
T4 Alcohol Dehydrogenase 2.9 Å (EMD-0406) 17

Which target map was used for (final) fitting?

primary map of the EMDB entry (emd_#####.map) 42
unsharpened/unmodified map from reconstruction software (emd_#####_additional_#.map) 16
half-map 1 for the FSC calculation (emd_#####_half_map_1.map) 5

Did you modify the map?

yes 21
no 42

Map preparation procedures used (answered by 21 responding yes to "did you modify the map?")


yes no
low pass filter 9 12
high pass filter 4 17
local normalization 4 17
segmentation 5 16
applied a mask 10 11
sharpened w/ constant B-value 12 9
sharpened w/ variable B-value 4 17
converted to structure factors 11 10

Modelling Category

Created an ab initio model 51
Optimized a known model 12

Effort type

fully automated 22
partially automated, some manual steps 40
manual 1

Refinement space

real 51
reciprocal 12

Final model refinement (choose one)

ARP/wARP/refmac 8
Chimera 10
COOT 1
CDMD 4
MDFF 4
pathwalker 2
phenix 13
refmac 4
rosetta 16
VMD 1

All modelling software used (indicate all used)

ARP/wARP 8
Buccaneer 4
Chimera 35
COOT 19
CDMD 4
direX 3
mainmast 10
MDFF 10
pathwalker 2
phenix 16
pymol 4
refmac 12
rosetta 17
situs 1
TEMPy 4
VMD 4

Target criteria used to identify model improvements (indicate all used)

map vs. model correlation coefficient 28
map vs. model FSC curve 19
energy function 25
cross-validation procedure 17
Other (additional reported criteria listed below) 24
--Reciprocal Space Target (refmac) 8
--EMRinger 6
--geometry (e.g. molprobity) 6
--phenix cryo-EM tools 4
--Segment-based Mander's overlap coefficient 4
--Q-score 1

Atomic Displacement (B)

Type of B factor Treatment All targets T1 T2 T3 T4
single overall B value applied to entire model 5 1 1 1 2
grouped B (segments) 3 1 1 1 0
grouped B (per residue) 6 1 1 2 2
individual B for each atom 23 6 6 6 5
none 30 8 7 6 9

Was the map scale (voxel size) adjusted during the modelling process?

no 63

Validation Checks


yes no
internal model consistency 60 3
fit of model to the target map 60 3
fit of model to map other than target (cross-validation) 25 38