Linking crystallographic model and data quality

By Paul Karplus, K. Diederichs. 2012

1,786 citations recorded by the source corpus.

Finessing Crystal Analysis Protein crystallography has revolutionized our understanding of a whole variety of biological processes (see the Perspective by Evans). In crystallography, the measure of agreement between the data and the calculated model is not on the same scale as the measure of data quality, making it challenging to choose an optimal high resolution limit beyond which the data should be discarded. Now, Karplus and Diederichs (p. 1030) introduce a statistical model that assesses agreement of model and data accuracy on the same scale. Determining the structures of biological macromolecules by x-ray crystallography requires solving the phase problem. The two techniques that dominate phase evaluation (multi- and single-wavelength anomalous diffraction) rely on element-specific scattering from incorporated heavy atoms. Liu et al. (p. 1033) present procedures for routine structure determination of native proteins with no heavy atom incorporation. The technique, which relies on combining data from multiple crystals, was used to determine the structures of four native proteins, including a 1200-residue complex. A statistical method places model and data quality on the same scale and indicates how far one can model. In macromolecular x-ray crystallography, refinement R values measure the agreement between observed and calculated data. Analogously, Rmerge values reporting on the agreement between multiple measurements of a given reflection are used to assess data quality. Here, we show that despite their widespread use, Rmerge values are poorly suited for determining the high-resolution limit and that current standard protocols discard much useful data. We introduce a statistic that estimates the correlation of an observed data set with the underlying (not measurable) true signal; this quantity, CC*, provides a single statistically valid guide for deciding which data are useful. CC* also can be used to assess model and data quality on the same scale, and this reveals when data quality is limiting model improvement.

Subfields: Materials Chemistry, Molecular Biology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Materials Chemistry

In Molecular Biology

Citations in the corpus

Decision making in xia2 (2013, methodology): “Recent systematic studies suggest that methods based on correlation coefficients may give a more robust insight (Evans & Murshudov, 2013; Karplus & Diederichs, 2012); however, these have yet to be implemented in xia2.”
Higher-Resolution Structure of the Human Insulin Receptor Ectodomain: Multi-Modal Inclusion of the Insert Domain (2016, methodology): “3 Å based on a CC1/2 criterion (rather than Rmerge- or <I/σ(I)>-based criteria) to permit maximum exploitation of weak data during maximum-likelihood refinement (Karplus and Diederichs, 2012).”
MrpH, a new class of metal-binding adhesin, requires zinc to mediate biofilm formation (2020, methodology): “Resolution cut-offs were based on CC1⁄2 [90] and completeness of the data.”
Conserved Neutralizing Epitope at Globular Head of Hemagglutinin in H3N2 Influenza Viruses (2014, background): “a CC 1/2, percentage of correlation between intensities from random half-datasets (60).”
Computational design of environmental sensors for the potent opioid fentanyl (2017, methodology): “Resolution cutoffs were determined by monitoring the refinement statistics in the context of the reflection data completeness and the CC 1⁄2 and I/sI values (Karplus and Diederichs, 2012).”
DIALS: implementation and evaluation of a new integration package (2018, methodology): “While the Rmeas value in the outer shell may appear excessive (in excess of 100%), the half sets of data are still significantly correlated, with CC1/2 = 0.669 (Karplus & Diederichs, 2012), and thus contribute usefully to the data set.”
Structure of the Shroom-Rho Kinase Complex Reveals a Binding Interface with Monomeric Shroom That Regulates Cell Morphology and Stimulates Kinase Activity* (2016, methodology): “Based on this analysis as well as values for completeness and CC* (57), we chose to truncate the data at 3.57-Å resolution.”
Crystal structure of a soluble cleaved HIV-1 envelope trimer (2013, methodology): “π Rpim = Σhkl (1/(n-1))1/2 Σi | Ihkl,i - <Ihkl> | / Σhkl Σi Ihkl,I, where Ihkl,i is the scaled intensity of the i measurement of reflection h, k, l, <Ihkl > is the average intensity for that reflection, and n is the redundancy ς CC1/2= correlation coefficient of half-datasets (77) ‡ Rcryst= Σhkl||Fo”
Can I solve my structure by SAD phasing? Planning an experiment, scaling data and evaluating the useful anomalous correlation and anomalous signal (2016, methodology): “…of each reflection, then the expected correlation CC* of the intensities in the merged data set with the true intensities is given approximately by (Karplus & Diederichs, 2012) CC ¼ 2CC1=2 1þ CC1=2 1=2 ; ð6Þ where CC1/2 is the correlation between intensities in the two half data sets, known as the…”