Mean coordinates cannot give all 8 Å majority contacts with one distance limit for the specified 2K39 and 6V5D models
For chain A coordinate models from ubiquitin entries 1D3Z (10 models), 2K39 (116), and 6V5D (176), this analysis compares 8 Å majority-contact labels with contacts from equally weighted mean coordinates after rigid alignment. For 2,278 residue pairs, it tests every graph attainable by a positive distance limit on each mean structure. The minimum error counts for fits on residues 1–70 and 1–76 were 0 and 0 for 1D3Z, 2 and 2 for 2K39, and 2 and 3 for 6V5D. Thus, for these specified coordinates and alignments, no single distance limit recovers every target label for 2K39 or 6V5D; both 1D3Z fits recover all labels. The result is limited to three selected ubiquitin entries, equal model weights, the stated fitting procedure, and calculated float64 distances. Model counts are not independent biological observations or physical state probabilities, and the analysis does not establish experimental contact truth or a coordinate-rounding bound after refitting.
ubiquitin · protein structure · contact maps · coordinate averaging · NMR ensembles · distance thresholds
References
9- usesThree contact rules give overlap above 0.94 between 2K39 and 6V5D with 168 edges per graph. sci:sha256:47adbec543999691f9d48e9817210cda9226ab32ec376c56fae2f175635e3106
The half-model distance supplied an order-statistic comparison. At a fixed distance limit, this value gives the half-or-more target labels. This analysis does not reproduce the equal-edge graph comparison.
- usesAt 8 Å, mean coordinates give three contact labels that do not agree with majority labels in 2K39 and 6V5D sci:sha256:6fc778742736e55967f2f1c3d961adb6884ad5d900f0688d47400d8ab98e3b31
The coordinate files, fitting ranges, equal weights, and reference pair tables supplied the data and comparison for the complete distance-limit sweep. The code also gives the same primary contact counts and mean distances. It does not reproduce the full source analysis.
- usesdoi:10.2210/pdb1d3z%2Fpdb
Uses chain A coordinates from PDB entry 1D3Z, archive version 1.4.
- usesdoi:10.2210/pdb2k39%2Fpdb
Uses chain A coordinates from PDB entry 2K39, archive version 1.3.
- usesdoi:10.2210/pdb6v5d%2Fpdb
Uses chain A coordinates from PDB entry 6V5D, archive version 1.2.
- doi:10.1186/1472-6807-9-12
- doi:10.1371/journal.pcbi.0040043
- doi:10.1126/science.1157092
- doi:10.1007/s10858-019-00288-8
Cited by
4 publicationsPublic catalog entries on Sciscape.
- The method changes the 6V5D minimum error counts from (2, 3) to (3, 2) with 0.00025 Å axis limits.
[Legacy relation: reproduces] New code gives the six primary aligned coordinate-mean minimum error counts, contact votes, and pair distances for the same three source files. It does not reproduce the other representations or the full source sensitivity grid.
- usesThe contact medoid does not give the minimum number of majority contact differences at 8 Å in 2K39 and 6V5D
This publication supplied the frozen PDB files, reference tables, pair scope, and question about representation. No prior code was executed.
- usesA minimum of 50 model removals makes the tested 2K39 and 6V5D contact graphs agree.
The coordinate files and contact-count table gave data and comparison counts for all three PDB entries. All 52 manifest files passed checksum and size checks. This run checked the 6,834 primary 8 Å counts before the model-removal analysis.
- In one 6V5D test, bond rotations within 0.0005 Å axis limits reduce the 1–76 minimum contact errors from three to two
[Legacy relation: reproduces] New code gives the same 6V5D aligned-mean minimum error counts of two and three for alignment on residues 1–70 and 1–76. This relation excludes other entries, other representations, and the full sensitivity grid.