view article

Figure 4
Validation plots for the different modelling approaches: axes are not absolute, but have been scaled relative to the minima and maxima of the plotted values, and only the minimum and maximum values are marked on the axes; for all model scores refer to Supporting Information §S1. (a) Plots for Figs. 3[link](a)–3[link](c). The plot confirms the visual inspection of the electron density; the ligand scores are improved across all metrics when refined as an ensemble relative to the ligand modelled alone. The absence of the superposed substrate model has a greater effect on the ligand model than the degradation of the protein model phases. (b) Plots for Figs. 3[link](d)–3[link](f). The ensemble model provides the best model for the ligand. The RSZD is decreased in the degraded-phase model for the reasons explained in the main text and is not related to an improved model. (c) Plots for Figs. 3[link](g)–3[link](i). Once more, the model statistics are improved with the addition of a superposed solvent model, with the caveat that the lower RSZD for degraded phases is not indicative of an improved model. (d) Plots for Figs. 3[link](j)–3[link](l). The inclusion of the solvent model still increases the quality of the model compared with when it is omitted, albeit marginally. The degraded phase model has lower B-factor ratios than either of the other two models owing to a decrease in the B factors of the ligand and a corresponding drop in occupancy.

Journal logoSTRUCTURAL
ISSN: 2059-7983
Follow Acta Cryst. D
Sign up for e-alerts
Follow Acta Cryst. on Twitter
Follow us on facebook
Sign up for RSS feeds