battery: the refine arm gets three protein sets and a ribosome

193L (tetragonal lysozyme 1.33 A, Na+ and Cl-), 1EXR (Ca2+-calmodulin 1.0 A,
four EF-hand Ca2+), 1K4C (KcsA in high K+, 2.0 A, the selectivity-filter K+
sites) and 4YBB (E. coli 70S ribosome 2.1 A, tag "large", to show that a
ribosome refines at all). All X-ray with deposited structure factors.

4YBB on an RTX 5080 (16 GB): completes on the GPU, 489,290 atoms with riding
hydrogens, 3.06 M reflections; wall 2:09, peak GPU memory 7.4 GB, peak host
RSS 6.9 GB; R-free 0.2876 as given, 0.2782 after placement, 0.2808 final
(deposited 0.2338).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01HVeDQqN9qRD57vnShzTC37

tools/battery/README.md

@@ -540,6 +540,8 @@ scored. Run it on its own (`--arm refine`) or beside the others (`--arm open,ref
 
 A row is an id and why the entry is there: `{"id": "1hr2", "tags": [...], "why": "..."}`. Public
 depositions only; the PDB ID may be named. To add one, add the row - nothing else is needed.
+The set tagged `large` (4ybb, a 70S ribosome of ~490,000 atoms with its riding hydrogens) is there to show that
+a ribosome refines at all: about two minutes and 7 GB of GPU memory on a 16 GB card, 7 GB of host memory.
 
 ### The REFMAC model check (open arm, opt-in)
 

tools/battery/refine.json

@@ -1,4 +1,8 @@
 {"arm": "refine", "sets": [
- {"id": "1hr2", "tags": ["rna", "mg"], "why": "Tetrahymena group I intron P4-P6 domain (J5/5a hinge mutant), 2.25 A, P212121, two molecules, 42 Mg2+ and 265 waters in the deposition (buffer MgCl2, Na cacodylate, NaCl, spermine); its Mg2+ sites are re-examined in the Cat_Wiz paper (Naleem et al. 2026, Nucleic Acids Res. 54, gkag120, Table 3) and its stacked-purine N7/O6 site in Leonarski et al. 2017 (Nucleic Acids Res. 45, 987-1004)"},
- {"id": "2r8s", "tags": ["rna", "protein", "mg"], "why": "the same P4-P6 domain bound to a synthetic Fab, 1.95 A, ammonium acetate and MgCl2 in the drop; Leonarski et al. 2017 (Nucleic Acids Res. 45, 987-1004) note that a water replaces the Mg2+ of 1HR2's stacked-purine N7/O6 site here"}
-]}
+ {"id": "1hr2", "tags": ["rna", "mg"], "why": "Tetrahymena group I intron P4-P6 domain (J5/5a hinge mutant), 2.25 A, P212121, two molecules, 42 Mg2+ and 265 waters in the deposition (buffer MgCl2, Na cacodylate, NaCl, spermine); its Mg2+ sites are re-examined in the Cat_Wiz paper (Naleem et al. 2026, Nucleic Acids Res. 54, gkag120, Table 3) and its stacked-purine N7/O6 site in Leonarski et al. 2017 (Nucleic Acids Res. 45, 987-1004)"}, 
+ {"id": "2r8s", "tags": ["rna", "protein", "mg"], "why": "the same P4-P6 domain bound to a synthetic Fab, 1.95 A, ammonium acetate and MgCl2 in the drop; Leonarski et al. 2017 (Nucleic Acids Res. 45, 987-1004) note that a water replaces the Mg2+ of 1HR2's stacked-purine N7/O6 site here"}, 
+ {"id": "193l", "tags": ["protein", "na", "cl"], "why": "tetragonal hen egg-white lysozyme, 1.33 A, P43212, crystallised from NaCl: the classic Na+ site (loop 60-73) and Cl- sites of lysozyme; a standard test protein"}, 
+ {"id": "1exr", "tags": ["protein", "ca"], "why": "calmodulin with four Ca2+ in its EF hands, 1.0 A (Wilson & Brunger 2000, J. Mol. Biol. 301, 1237-1256): Ca2+ sites of carboxylate-rich, well-determined pentagonal-bipyramidal geometry"}, 
+ {"id": "1k4c", "tags": ["protein", "k"], "why": "the KcsA potassium channel with a Fab, 2.0 A, in high K+ (Zhou et al. 2001, Nature 414, 43-48): the K+ sites of the selectivity filter, eight carbonyl O each"}, 
+ {"id": "4ybb", "tags": ["rna", "protein", "mg", "large"], "why": "the E. coli 70S ribosome, 2.1 A, two ribosomes in the asymmetric unit, ~295,000 atoms with Mg2+ and Zn2+ (Noeske et al. 2015, Nat. Struct. Mol. Biol. 22, 336-341): whether a ribosome refines at all"}
+]}