Local Error Estimates Dramatically Improve the Utility of Homology Models for Solving Crystal Structures by Molecular Replacement

Gábor Bunkóczi,Björn Wallner,Randy J Read

doi:10.1016/j.str.2014.11.020

Gábor Bunkóczi, Björn Wallner + Show 1 more

Open Access

https://doi.org/10.1016/j.str.2014.11.020

Copy DOI

Abstract

SummaryPredicted structures submitted for CASP10 have been evaluated as molecular replacement models against the corresponding sets of structure factor amplitudes. It has been found that the log-likelihood gain score computed for each prediction correlates well with common structure quality indicators but is more sensitive when the accuracy of the models is high. In addition, it was observed that using coordinate error estimates submitted by predictors to weight the model can improve its utility in molecular replacement dramatically, and several groups have been identified who reliably provide accurate error estimates that could be used to extend the application of molecular replacement for low-homology cases.

Highlights

About two-thirds of crystal structures deposited in the Protein Data Bank (PDB) (Berman et al, 2003) are solved by the method of molecular replacement (MR), and it has been estimated that about 80% could have been solved by MR using templates available at the time of deposition (Long et al, 2008)
The likelihood gain (LLG) Score The LLG score provides a direct measure for evaluating the quality of a predicted structure for MR
As the difficulty of MR is proportional to the discriminative power of the LLG score for correct solutions versus noise, it is clear that MR is unlikely to work with models having a GDT_TS below 40–50, since the LLG score is broadly flat in that region, at least for molecules of the size typically explored in CASP

Summary

SUMMARY

Predicted structures submitted for CASP10 have been evaluated as molecular replacement models against the corresponding sets of structure factor amplitudes. It has been found that the log-likelihood gain score computed for each prediction correlates well with common structure quality indicators but is more sensitive when the accuracy of the models is high. It was observed that using coordinate error estimates submitted by predictors to weight the model can improve its utility in molecular replacement dramatically, and several groups have been identified who reliably provide accurate error estimates that could be used to extend the application of molecular replacement for low-homology cases

INTRODUCTION

RESULTS AND DISCUSSION

Score Rms B Factor

Conclusions

EXPERIMENTAL PROCEDURES