On the relationship between cumulative correlation coefficients and the quality of crystallographic data sets

Abstract

In 2012, Karplus and Diederichs demonstrated that the Pearson correlation coefficient CC is a far better indicator of the quality and resolution of crystallographic data sets than more traditional measures like merging R-factor or signal-to-noise ratio. More specifically, they proposed that CC be computed for data sets in thin shells of increasing resolution so that the resolution dependence of that quantity can be examined. Recently, however, the CC values of entire data sets, i.e., cumulative correlation coefficients, have been used as a measure of data quality. Here, we show that the difference in cumulative CC value between a data set that has been accurately measured and a data set that has not is likely to be small. Furthermore, structures obtained by molecular replacement from poorly measured data sets are likely to suffer from extreme model bias.

Heterogeneous Composition of Oxygen-Evolving Complexes in Crystal Structures of Dark-Adapted Photosystem II

“…The structural biology community has often misused B -factors as fudge factors for model refinement under invalid assumptions. Whenever invalid assumptions were made or invalid models were built for model refinement, the resulting ED maps could have significant assumption-based or model-based biases, which could sometimes lead to overinterpretations of structural data. ,,− ,− For some time, we were very concerned about the validity of some earlier data processing procedures of XFEL data sets, the quality of processed data, structures being derived, and overinterpretation of some XFEL data for PSII intermediates. − ,,, It was often assumed that the occupancy and B -factor could not be simultaneously refined in crystal structures determined at ∼2.0 Å resolution. Because the B -factor affects the slope of the ED distribution in the log rr plot and occupancy affects the entirety of the ED distribution curves linearly, the two parameters can be accurately decoupled in omit maps using the methods described in this study.…”

“…13,14,38−40,44−57 For some time, we were very concerned about the validity of some earlier data processing procedures of XFEL data sets, the quality of processed data, structures being derived, and overinterpretation of some XFEL data for PSII intermediates. [38][39][40]54,56,57 It was often assumed that the occupancy and B-factor could not be simultaneously refined in crystal structures determined at ∼2.0 Å resolution. Because the B-factor affects the slope of the ED distribution in the log rr plot and occupancy affects the entirety of the ED distribution curves linearly, the two parameters can be accurately decoupled in omit maps using the methods described in this study.…”

Mix-and-inject XFEL crystallography reveals gated conformational dynamics during enzyme catalysis

“…Special care was taken to avoid model bias in the XFEL structure, which has been reported to be a potential concern for sparse data generated by some serial femtosecond crystallography experiments (56). In addition to refining a model with the single conformation for helix H (Figure 1B), a dual conformation model for this helix was also refined against the XFEL data.…”

Simulating diffraction photographs based on molecular dynamics trajectories of a protein crystal: a new option to examine structure-solving strategies in protein crystallography

“…One statistic that deserves a separate discussion is R merge , which is widely used as a measure of precision for experimental crystallographic data (Holton et al, 2014;Wang et al, 2017). In our recovered structures, this parameter is only borderline acceptable: 0.105 and 0.136 for the structures at 250 and 298 K, respectively (Supplementary Table S1).…”