journal menu
![[Figure 5]](rr5136fig5mag.jpg)
|
|
Figure 5
PEG fragments ranked by LLDF. The PEG fragments are labelled as numbered in PDB entry 5l6d and are ranked by the LLDF (best to worst), which is the first item in the data columns/row, followed by the real-space R value (RSR) and the real-space correlation coefficient (RSCC) as listed in the validation report. Density levels are 1σ for the 2mFo − DFc map (blue grid, highlighted around the PEG fragments) except for PG4 H305, where the contours of the blue map have been lowered to 0.70σ. Difference density is shown at ±3σ levels. Red LLDF numbers correspond to yellow highlights in the validation report. The reason for the scores is not always transparent, which emphasizes the need for electron-density inspection for ligands and caution against simple acceptance (or condemnation) based on the LLDF score. For example, in PEG H307 it is obvious that the model is insufficient to explain the additional density (perhaps an alternate conformation), while the very plausible hydrogen-bond network of PG4 H301 does not prevent a highlighted (presumably bad) score. Note that in contrast to H301 the central positive density peak in PE8 H302 could not be modelled with a water molecule because it was impossible (or beyond our limit of tedium) to orient the PE8 in one unique conformation where only O atoms would form a reasonable hydrogen-bond network to the neighbouring molecules. Incorrect register of the O versus C atoms frequently causes improper close contacts to neighbouring atoms. 2mFo–DFc density is displayed at 1.2σ. Note also that the assumption that PEG fragment models must always end with a terminal oxygen (–OH) at each end is an unrealistic expectation. This figure was displayed and rendered with Coot (Emsley et al., 2010  ). |
![[Figure 5]](rr5136fig5.jpg)
 | STRUCTURAL BIOLOGY |
ISSN: 2059-7983
Open 
access
access
