data quality and model parameterisation martyn winn ccp4, daresbury laboratory, u.k. prague, april...
TRANSCRIPT
Data quality and model parameterisation
Martyn WinnCCP4, Daresbury Laboratory, U.K.
Prague, April 2009
Model Parameters
E.g. asymmetric unit contains n copies of a protein of N atoms
Coordinates
3 x N x n xyz co-ordinates
or ... 6 x M x n if each protein modelled as M rigid bodies
or ... ~ 0.5 x N x n torsion angles
Displacement parameters
1 x N x n B factors
or ... 6 x N x n anisotropic U factors
or ... 20 x M x n if each protein has M TLS groups
Model Parameters (2)Occupancies
Usually fixed at 1.0 for protein
... except for alternative conformations (usually sum to 1.0)
Water/ligand occupancies
Scaling parameters etc.
koverall, Boverall, kBabinet, BBabinet, ksolvent, Bsolvent
twin fraction
Ultra-high resolution
Multipolar expansion coefficients
Interatomic scatterers
Reflection DataNumber of independent reflections, dependent on:
– spacegroup– resolution– completeness
For each reflection, one has at least F/sigF.Might also have reliable experimental phases φ or F(+)/F(-)
Data / parameter ratioRefinement means minimise -log(likelihood):
Nonlinear function of model parameters.
Global minimum and many local minima.
Need good data/parameter ratio.
Strong dependence on resolution.
No strong dependence on protein size.
Generally not enough data ....Reduce number of parameters - constraintsAdd data - restraints
RestraintsExpected geometry of the protein treated as additional data
bond lengthsbond anglestorsions / dihedral (but not φ,ψ)chirality (e.g. chiral volume)planaritynon-bonded (VdW, H-bonds, etc.)B factors (between bonded atoms)U factor restraints (similarity, sphericity, rigid bond)NCS (position or conformation)
Data / parameter ratio
Not really true ... assumes all data independentbond lengths and angles and planar restraints in ring systembond length restraint vs. high resolution diffraction data
Estimate as: no. reflections + no. restraints no. parameters
Restraints may be more necessary in poorly determined parts of the structure.
Restraints have associated weights:Overall w.r.t. reflection data
Individual weights e.g. WB
calmodulin at 1.8 Å (1clm) 1132 protein atoms, 4 Ca atoms, 71 waters 4828 x, y, z, B factors
No. of unique reflections 10610 (deposited 1993 no test set!)
data/parameter = 2.2
Bond restraints: 1144Angle restraints: 1536Torsion restraints: 429Chiral restraints: 170Planar restraints: 874Non-bonded restraints: 1391B factor restraints: 2680(no NCS)
total restraints = 8224 data/parameter = 3.9
calmodulin at 1.0 Å (1exr)
1467 protein atoms (inc. alt. conf.), 5 Ca atoms, 178 waters 4950 x, y, z+ 9900 anisotropic U factors+ 316 occupancy parameters total parameter count = 15166
No. of unique reflections 77150No. in test set 7782 (10%)Data for refinement 69368
No. of restraints (PDB header) 22732
data/parameter = 4.6
data/parameter = 6.1
GCPII at 1.75 Å (3d7g)
5724 protein atoms (inc. alt. conf.), 211 ligand atoms, 617 waters 26046 x, y, z, B factors + 162 anisotropic U factors (S, Zn, Ca, Cl only)+ 225 occupancy parameters total parameter count = 26433
No. of unique reflections 105077No. in test set 1550 (1.5%)Data for refinement 103527
No. of restraints (PDB header) 44652
data/parameter = 3.9
data/parameter = 5.6
Thioredoxin reductase at 3.0Å (1h6v)
22514 protein atoms, 552 ligand atoms, 9 waters 92300 x, y, z, residual B factors 6 TLS groups 120 TLS parameters
No. of unique reflections 69328No. in test set 3441 (5%)Data for refinement 65887
No. of restraints 209378(inc. 44484 NCS restraints)
data/parameter = 0.7
data/parameter = 3.0
Getting a good R-factorThe old way:1.Refine parameters so that Fcalc (from model) agrees
with Fobs for all reflections
2.Calculate: R = |Fobs| - s | Fcalc | / |Fobs|
(Note: precise value may depend on scaling used)
3.Add parameters until R is sufficiently low
What’s wrong with that ? ?
Avoiding overfitting: RfreeWhat's wrong?:• Can add any old parameters to improve R-factor, when low
data/parameter ratio• May not be physically correct – "overfitting"
Solution:• Calculate R-factor on a set of reflections not used in
refinement = "Rfree"• If changes to model improve Rfree as well as R, then they are
good.• Note: Rfree is global number - useful for refinement
strategies, not useful for assessing changes to a few atoms
Choosing your free reflections
• Usually a randomly chosen subset.• Typically 5-10% (CCP4 default is 5%)• If you have enough reflections, impose
maximum number (2000 in phenix.refine)• Free set also used in maximum likelihood to
estimate σA parameters
Rfree and NCS• NCS operators map different regions of reciprocal asymmetric
unit onto each other. Reflections in these regions are correlated.
gaps = free set
working reflections
free reflections
Rfree and NCS• Solution: choose free set from thin shells in reciprocal space
Pros:NCS operators link regions of same resolution which should be both in a shell or outside it
Cons:Large number of shells thin shells most free reflections close to edge and correlated to non-free reflectionsSmall number of shells significant gaps in resolution range, poor determination of σA
SFTOOLS: RFREE 0.05 SHELL 0.001
3rd argument = width of shells in Å-1
Also DATAMAN.
Width 0.013 shells
Width 0.001320 shells(default)
1xmp (1.8 Å)
Width 0.0053 shells
Width 0.000520 shells(default)
XXX (3.8 Å)
• Can increase size of free set to mitigate edge effects• Or use NCS-related free set islands
• Reflections also correlated to immediate neighbours in reciprocal space - can exclude these from working and free sets
Fabiola, Korostelev & Chapman, Acta Cryst D62, 227, (2006)• Rapidly run out of working reflections!
Be aware that correlations can artificially reduce your Rfree
Rfree and NCS
Rfree and twinning
Twinning operator might relate e.g. reflection (1,2,3) to (2,1,-3)
These two reflections should both be in the working set or the free set.
1. Select free set in thin shells (as NCS) 2. Select free reflections in higher lattice symmetry
Transferring free R setsUse the same free set for:
additional datasets for same proteindatasets from isomorphous proteins (derivatives, complexes, etc.)(how isomorphous is not clear, but play safe ...)
Otherwise initial R & Rfree will be similar and low for second structure - it has been refined against most of your free reflections
Further refinement may lead to divergence of R & Rfree, masking the bias. Harder to detect over-fitting. Although may eventually reset Rfree.
How:Use "CAD" / "Merge MTZ files (CAD)" in CCP4.
Useful resources
http://ccp4wiki.org/ - CCP4 Wikihttp://strucbio.biologie.uni-konstanz.de/ccp4wiki/ - CCP4
community wikiProceedings of Study Weekend 2004 (Acta Cryst D, Dec 2004)