Developed community standards on setting the diffraction resolution yields improved density maps and a new monosaccharide model in pdb_00004dda

A systematic survey of monosaccharides in the PDB archive threw up a possible improvement in model fit in pdb_00004dda. Reprocessing of the publicly open archive of the raw diffraction images was made using EVAL. The diffraction resolution was assessed by the CC_1/2 = 0.5 criterion and confirmed by paired model refinement implemented in PDB-REDO. The resulting change in diffraction resolution, from 2.4 Å to 2.0 Å, improved the clarity of the electron density at the monosaccharide binding site. The NAG (N-acetyl-β-D-glucosa­mine) was refitted as NDG (N-acetyl-α-D-glucosa­mine). This example illustrates the advantages of raw diffraction data archiving, as aimed at by FAIR data polices, allowing reprocessing as agreement on community metrics.

Introduction

A systematic survey of monosaccharides in the PDB archive threw up a possible improvement in model fit in pdb_00004dda. The structure was published in Tanley et al. (2012) with a diffraction resolution of <I/σI> = 4.0. The ways to establish the diffraction resolution have since then developed into community-agreed improved criteria. Reprocessing of the publicly open archive of the raw diffraction images was made using EVAL (Schreurs et al. 2010). The diffraction resolution was assessed by the CC_1/2 = 0.5 criterion (Karplus & Diederichs 2012) and confirmed by paired-model refinement using Refmacat (Yamashita et al. 2023) implemented in PDB-REDO (Joosten et al. 2014). The resulting change in diffraction resolution, from 2.4 Å to 2.0 Å, improved the clarity of the electron density at the monosaccharide binding site. The NAG (N-acetyl-β-D-glucosa­mine) was refitted as NDG (N-acetyl-α-D-glucosa­mine). The raw diffraction images for pdb_00004dda in the original RAXIS.osc format, along with the reprocessing files, can be found at https://zenodo.org/records/18258544. The raw data were converted to fullCBF with the make-cbf tool as part of imgCIF_Creator (Kluyver et al., 2026) and deposited to Zenodo (https://zenodo.org/records/20326720). This example illustrates the advantages of raw diffraction data archiving, as aimed at by FAIR data polices (Wilkinson et al., 2016), allowing reprocessing as agreement on community metrics improve.

The benefits of raw diffraction data archiving linked to deposited PDB files were described with an example by the IUCr abstract (Helliwell et al., 2023) as a case study within the Global Open Science Cloud (Chen et al. 2023). Their example documented a case of a good choice of diffraction limit by the authors (Sato et al. 2021) and PDB entry pdb_00007ccy. In the case of pdb_00004dda we can now illustrate below the corrective power and an improvement with one of our own studies.

Methods and results

Data were processed in the same way as in Tanley et al. (2012) and detailed in Tanley et al. (2013) by EVAL (Schreurs et al. 2010), which uses a diffraction data integration method based on ab initio predicted profiles, but only extending the resolution and with the summary statistics now including CC_1/2. The new diffraction resolution was chosen to be at a CC_1/2 criterion of 0.5 and confirmed by paired model refinement implemented in PDB-REDO (Joosten et al. 2014). The electron density at the monosaccharide binding site after the reprocessing and re-refinement is shown in Fig. 1. The raw data collection details are provided in Table 1. The data statistics of pdb_00004dda and of the reprocessed and re-refined models are given in Table 2. The metadata for the raw diffraction images have been automatically checked by the IUCrData checkcif-for-raw-data software (https://publbio.iucr.org/imgcif/).

Table 1
Experimental details

Raw Data
DOI	https://doi.org/10.5281/zenodo.18258544
	https://doi.org/10.5281/zenodo.20326720
Data archive	Zenodo
Data Format	OSC and CBF
Data Collection
Diffractometer	Rigaku Micromax-007 generator
Detector type	R-AXIS IV image plate
Radiation type	Cu Kα
Wavelength (Å)	1.5418
Beam centre (mm) (fast, slow)	150.00, −149.89
Detector axis	−Z
Detector distance (mm)	140.00†
Pixel size (mm)	0.100 × 0.100
No. of pixels	3000 × 3000
No. of scans	1
Scan axis	ω, χ
Start angle, increment per frame (°)	0.0, 1.0
Scan range (°)	180.0
No. of frames	180
Exposure time per frame (s)	4.5
The detector distance in the raw image files is given as 140.0 mm, while after indexing and refinement it is approximately 135 mm.

Table 2
4dda X-ray crystallographic data processed via EVAL15 and refinement statistics for HEWL co-crystallized with N-acetyl-D-glucosamine (Tanley et al 2012) and new data processing and refinement statistics undertaken by PDB-REDO with TLS refinements and hydrogens in riding positions

Space group	P43212	P43212
Unit-cell parameters (Å)	a = b = 78.37, c = 36.58	a = b = 78.37, c = 36.58
Molecular mass (Da)	14700	14700
Molecules per asymmetric unit	1	1
Observed reflections	49543	93073 (10160)
Unique reflections	4120	7468 (796)
Resolution (Å)	19.59–2.40 (2.48–2.40)	19.59–2.0 (2.07–2.0)
Completeness (%)	85.3 (100)	91.5 (100.0)
Rmerge	0.147 (0.607)	0.243 (1.960)
<I/σ(I)>	13.6 (4.02)	8.15 (1.14)
CC1/2	†	0.996 (0.508)
Multiplicity	12.1 (12.8)	12.5 (12.8)
Cruickshank DPI (Å)	0.806	0.287
Average Bfactor (Å^2)	29.5	21.9
Refinement Rfactor/Rfree(%)	20.0/28.5	20.4/26.0
R.m.s. Z bonds	1.201	0.204
R.m.s. Z angles	0.967	0.439
Molprobity Clashscore	5.53	2.00
Ramachandran Z-score	−2.12	−0.06
Rotamer normality Z-score	−1.88	−1.11
Note 1. CC1/2 was not a metric in use in 2012.

Figure 1 Electron density 2mFo−DFc (1.58 r.m.s.) with the monosaccharide NDG fitted (this study at 2.0 Å). The real-space correlation coefficient of NDG is 0.93. This figure was produced using Coot (Emsley et al., 2010).

Discussion and conclusions

As community metrics in macromolecular crystallography evolve, the processing of archived raw diffraction images can be successfully re-evaluated as shown in this case with an improved diffraction resolution limit and model. The full range of situations suitable for a Raw Data Letter are described in Kroon-Batenburg et al. (2022).