Hugging Face's logo

Datasets:
tonynzh2
/
afdb6

Tags:
biology
protein-structure
alphafold
motif-search
Dataset card Files
xet
 Community
tonynzh2 commited on
Commit
20bfdf4
·
verified ·
1 Parent(s): 249b720

Publish reference-coverage audit and mark historical InterPro v2 limitations

Browse files
Files changed (10) hide show
  1. annotations/biolip_interpro_catalytic_comparison_20261002/coverage_audit/REPORT.md +97 -0
  2. annotations/biolip_interpro_catalytic_comparison_20261002/coverage_audit/per_reference.csv +0 -0
  3. annotations/biolip_interpro_catalytic_comparison_20261002/coverage_audit/per_reference.jsonl +0 -0
  4. annotations/biolip_interpro_catalytic_comparison_20261002/coverage_audit/summary.json +387 -0
  5. annotations/complete_ligand_sets_20261002/README.md +6 -0
  6. annotations/complete_ligand_sets_20261002/REPORT.md +6 -0
  7. annotations/complete_ligand_sets_20261002/coverage_audit/REPORT.md +97 -0
  8. annotations/complete_ligand_sets_20261002/coverage_audit/per_reference.csv +0 -0
  9. annotations/complete_ligand_sets_20261002/coverage_audit/per_reference.jsonl +0 -0
  10. annotations/complete_ligand_sets_20261002/coverage_audit/summary.json +387 -0
annotations/biolip_interpro_catalytic_comparison_20261002/coverage_audit/REPORT.md ADDED
@@ -0,0 +1,97 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Why only 245 InterPro reference sets were initially evaluable
2
+
3
+ The original 245/876 is **coverage of one annotation-and-role-matching
4
+ implementation**, not a count of enzymes that use ligands. No reference was
5
+ established to be ligand-free by this analysis. The audit identified real
6
+ normalization omissions, so the original v2 selection is historical and is
7
+ superseded by the separately versioned comparison.
8
+
9
+ ## What was audited
10
+
11
+ The population is all 876 frozen motif queries, representing 863 M-CSA entries.
12
+ Their available reference chains produce 880 distinct native strings. We read
13
+ the completed exact-string InterPro Matches API/fresh-scan cache, not new
14
+ predictions, generated sequences, or annotations retrieved only by accession.
15
+
16
+ Selected traditional evidence is Pfam 38.2, CDD 3.21, PIRSR 2025_05,
17
+ PROSITE patterns/profiles 2026_01 and SFLD 4. InterPro-N and other neural sources
18
+ are excluded before library-version checks. Reference catalytic roles came from
19
+ current M-CSA cofactor pages retrieved on 2 October, their historical curated
20
+ mechanistic-role table, the frozen 23 September reaction snapshot, and explicit
21
+ catalytic-site descriptors. Presence in the cofactor list alone was insufficient.
22
+
23
+ We traced every query through reference mapping, annotation retrieval, literal
24
+ ligand extraction, and catalytic-role joining. Categories below are mutually
25
+ exclusive and sum to 876. `per_reference.jsonl` contains the underlying records.
26
+
27
+ ## The original v2 coverage funnel
28
+
29
+ | Last successful stage / reason for unresolved coverage | Queries | Queries with a cofactor listed in current M-CSA |
30
+ |---|---:|---:|
31
+ | Evaluable nonempty catalytic reference set | 245 | 227 |
32
+ | No binding descriptor extracted | 250 | 93 |
33
+ | Only generic or explicitly excluded descriptors | 112 | 51 |
34
+ | Specific wording remained unmapped or ambiguous | 53 | 27 |
35
+ | Named ligand found, but no accepted catalytic-role join | 105 | 66 |
36
+ | Relevant specific identity unresolved, blocking completeness | 24 | 24 |
37
+ | Reference mapping or existing length/sequence scope unresolved | 87 | 58 |
38
+
39
+ The middle four rows account for the earlier combined **520** unknown cases.
40
+ Of these, **237 actually have a cofactor listed in M-CSA**. The other 283 have no
41
+ listed cofactor in the retrieved field; that does not establish absence of a
42
+ substrate, cofactor, or other ligand.
43
+
44
+ There were **zero missing cached API records** among the 880 available strings,
45
+ and no incompatible selected-library records. Of those strings, 879 had a
46
+ selected traditional-library match. Thus the main problem was not an API outage.
47
+ InterPro can identify a family or domain without providing a specific ligand ID.
48
+
49
+ The 87 reference-mapping/scope cases partition into 45 overlength references,
50
+ 25 multiple-distinct-string mappings, 9 uncertain mappings, 6 missing references,
51
+ and 2 nonstandard strings. These are input-policy or mapping issues, not evidence
52
+ about ligand dependence. In particular, the existing 1,024-residue cap on
53
+ reference annotation is a pipeline choice; it is not logically required just
54
+ because training endpoints have that cap. It was held fixed for this comparison.
55
+
56
+ ## Concrete examples
57
+
58
+ | Reference | What the records say | Why v2 did not establish a usable set |
59
+ |---|---|---|
60
+ | Phospholipase C, M-CSA 27 | M-CSA records catalytic zinc; CDD/Pfam/PROSITE explicitly say zinc-dependent | Our title extractor required binding/ligand wording and ignored the dependency title. This was an extraction omission. |
61
+ | Trypsin, M-CSA 173 | Compatible InterPro matches include a generic substrate-binding descriptor | This does not identify a specific ligand. It is not a missing API result. |
62
+ | RNase HI, M-CSA 163 | InterPro describes RNA/DNA-hybrid binding; M-CSA lists magnesium | The retrieved descriptor does not establish a specific magnesium-binding label. |
63
+ | Betaine-aldehyde dehydrogenase, M-CSA 100 | InterPro supplies NAD labels; the reaction lists NAD(1−) | A charge-qualified name failed to join the NAD role family. This was a normalization bug. |
64
+ | Cytokinin dehydrogenase, M-CSA 115 | FAD binding and catalytic FAD roles are present | Coordinated wording such as “FAD and cytokinin binding” was parsed too narrowly. |
65
+ | Dialkylglycine decarboxylase, M-CSA 482 | PLP is annotated and listed as a cofactor | The historical role-table field used by the selector was empty. The current M-CSA narrative does describe a catalytic PLP role. This is a role-source coverage limitation, not proof that PLP is neutral. |
66
+ | Sulfite oxidase, M-CSA 121 | Heme and molybdenum-cofactor descriptors are present | Specific cofactor identity remained unresolved under the conservative InterPro normalizer. |
67
+ | Glutamate racemase, M-CSA 1 | No cofactor field was extracted; the curated reaction still names glutamate | An empty cofactor field cannot be interpreted as “no ligand.” |
68
+
69
+ The [M-CSA PLP example](https://www.ebi.ac.uk/thornton-srv/m-csa/entry/482/)
70
+ illustrates why absence from one structured role field is weaker than absence
71
+ of documented evidence in the resource as a whole. No new narrative-derived
72
+ roles were silently added to this comparison.
73
+
74
+ ## Confirmed corrections and their consequences
75
+
76
+ The NAD(1−)/NADP(3−) role-family error affected 20 previously unevaluable queries.
77
+ It also omitted NAD from the nonempty requirements of M-CSA 312, 500 and 501.
78
+ Those three groups had 6,107 supported memberships under v2, so the earlier
79
+ selector was not simply conservative: it could also require too few labels.
80
+
81
+ ChEBI identifies [NAD(1−)](https://www.ebi.ac.uk/chebi/CHEBI:57540) as a
82
+ protonation form of NAD+. This correction concerns reference role evidence;
83
+ it does not let arbitrary redox or ligand identities substitute for one another
84
+ when matching reference and hit annotations.
85
+
86
+ Version 3 separately fixes the role join, retains explicit ligand-dependency
87
+ titles, and handles the observed coordinated FAD-binding wording. It reparses
88
+ the full saved cache identically for references and hits. It gives **274
89
+ evaluable references, 6,568,858 supported unique motif–sequence memberships,
90
+ 5,018,150 distinct supported strings and 206 supported groups**. The remaining
91
+ reference statuses are 491 without a usable named catalytic set, 24 with
92
+ unresolved relevant identity, and the same 87 mapping/scope cases.
93
+
94
+ Original v2 artifacts and counts remain intact. Version 3 and the BioLiP2
95
+ comparison are in `../../biolip_interpro_catalytic_comparison_20261002/`.
96
+ These are annotation-supported subset indices, not activity measurements or
97
+ aligned training-pair releases.
annotations/biolip_interpro_catalytic_comparison_20261002/coverage_audit/per_reference.csv ADDED
The diff for this file is too large to render. See raw diff
 
annotations/biolip_interpro_catalytic_comparison_20261002/coverage_audit/per_reference.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
annotations/biolip_interpro_catalytic_comparison_20261002/coverage_audit/summary.json ADDED
@@ -0,0 +1,387 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "frozen_policy": "interpro_complete_catalytic_ligands_20261002_v2",
3
+ "query_groups": 876,
4
+ "reference_strings": 880,
5
+ "cache_not_found": 0,
6
+ "references_with_annotation_issues": 0,
7
+ "reference_strings_with_selected_matches": 879,
8
+ "breakdown": {
9
+ "reference_mapping_or_scope": {
10
+ "groups": 87,
11
+ "cofactor_listed": 58,
12
+ "no_selected_matches": 6,
13
+ "charged_name_join_flags": 0,
14
+ "role_table_gap": 8,
15
+ "named_dependency_title": 1
16
+ },
17
+ "no_extracted_binding_descriptors": {
18
+ "groups": 250,
19
+ "cofactor_listed": 93,
20
+ "no_selected_matches": 1,
21
+ "charged_name_join_flags": 0,
22
+ "role_table_gap": 0,
23
+ "named_dependency_title": 7
24
+ },
25
+ "evaluable": {
26
+ "groups": 245,
27
+ "cofactor_listed": 227,
28
+ "no_selected_matches": 0,
29
+ "charged_name_join_flags": 3,
30
+ "role_table_gap": 17,
31
+ "named_dependency_title": 14
32
+ },
33
+ "named_labels_without_role_join": {
34
+ "groups": 105,
35
+ "cofactor_listed": 66,
36
+ "no_selected_matches": 0,
37
+ "charged_name_join_flags": 20,
38
+ "role_table_gap": 42,
39
+ "named_dependency_title": 14
40
+ },
41
+ "only_generic_or_excluded_descriptors": {
42
+ "groups": 112,
43
+ "cofactor_listed": 51,
44
+ "no_selected_matches": 0,
45
+ "charged_name_join_flags": 0,
46
+ "role_table_gap": 0,
47
+ "named_dependency_title": 3
48
+ },
49
+ "specific_identity_blocks_reference": {
50
+ "groups": 24,
51
+ "cofactor_listed": 24,
52
+ "no_selected_matches": 0,
53
+ "charged_name_join_flags": 0,
54
+ "role_table_gap": 1,
55
+ "named_dependency_title": 2
56
+ },
57
+ "unmapped_or_unresolved_descriptors": {
58
+ "groups": 53,
59
+ "cofactor_listed": 27,
60
+ "no_selected_matches": 0,
61
+ "charged_name_join_flags": 0,
62
+ "role_table_gap": 0,
63
+ "named_dependency_title": 0
64
+ }
65
+ },
66
+ "mapping_priority_breakdown": {
67
+ "overlength": 45,
68
+ "multiple_distinct_reference_strings": 25,
69
+ "mapping_uncertain": 9,
70
+ "missing": 6,
71
+ "nonstandard": 2
72
+ },
73
+ "charged_role_join_affected": {
74
+ "unknown_no_named_catalytic_reference_labels": [
75
+ {
76
+ "query_id": "mcsa_100_1a4s_assembly1",
77
+ "compounds": [
78
+ {
79
+ "count": 1,
80
+ "type": "reactant",
81
+ "chebi_id": "57540",
82
+ "name": "NAD(1-)",
83
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
84
+ }
85
+ ],
86
+ "existing_required_labels": []
87
+ },
88
+ {
89
+ "query_id": "mcsa_240_1szd_assembly1",
90
+ "compounds": [
91
+ {
92
+ "count": 1,
93
+ "type": "reactant",
94
+ "chebi_id": "57540",
95
+ "name": "NAD(1-)",
96
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
97
+ }
98
+ ],
99
+ "existing_required_labels": []
100
+ },
101
+ {
102
+ "query_id": "mcsa_336_2hdh_assembly1",
103
+ "compounds": [
104
+ {
105
+ "count": 1,
106
+ "type": "reactant",
107
+ "chebi_id": "57540",
108
+ "name": "NAD(1-)",
109
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
110
+ }
111
+ ],
112
+ "existing_required_labels": []
113
+ },
114
+ {
115
+ "query_id": "mcsa_429_1brm_assembly1",
116
+ "compounds": [
117
+ {
118
+ "count": 1,
119
+ "type": "reactant",
120
+ "chebi_id": "58349",
121
+ "name": "NADP(3-)",
122
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/58349_FtUFgHJ.mol"
123
+ }
124
+ ],
125
+ "existing_required_labels": []
126
+ },
127
+ {
128
+ "query_id": "mcsa_450_1dgs_assembly1",
129
+ "compounds": [
130
+ {
131
+ "count": 1,
132
+ "type": "reactant",
133
+ "chebi_id": "57540",
134
+ "name": "NAD(1-)",
135
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
136
+ }
137
+ ],
138
+ "existing_required_labels": []
139
+ },
140
+ {
141
+ "query_id": "mcsa_458_1a4i_assembly1",
142
+ "compounds": [
143
+ {
144
+ "count": 1,
145
+ "type": "reactant",
146
+ "chebi_id": "58349",
147
+ "name": "NADP(3-)",
148
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/58349_FtUFgHJ.mol"
149
+ }
150
+ ],
151
+ "existing_required_labels": []
152
+ },
153
+ {
154
+ "query_id": "mcsa_576_1j49_assembly1",
155
+ "compounds": [
156
+ {
157
+ "count": 1,
158
+ "type": "reactant",
159
+ "chebi_id": "57540",
160
+ "name": "NAD(1-)",
161
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
162
+ }
163
+ ],
164
+ "existing_required_labels": []
165
+ },
166
+ {
167
+ "query_id": "mcsa_584_1ldm_assembly1",
168
+ "compounds": [
169
+ {
170
+ "count": 1,
171
+ "type": "reactant",
172
+ "chebi_id": "57540",
173
+ "name": "NAD(1-)",
174
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
175
+ }
176
+ ],
177
+ "existing_required_labels": []
178
+ },
179
+ {
180
+ "query_id": "mcsa_607_1r16_assembly1",
181
+ "compounds": [
182
+ {
183
+ "count": 1,
184
+ "type": "reactant",
185
+ "chebi_id": "57540",
186
+ "name": "NAD(1-)",
187
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
188
+ }
189
+ ],
190
+ "existing_required_labels": []
191
+ },
192
+ {
193
+ "query_id": "mcsa_67_1pjb_assembly1",
194
+ "compounds": [
195
+ {
196
+ "count": 1,
197
+ "type": "reactant",
198
+ "chebi_id": "57540",
199
+ "name": "NAD(1-)",
200
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
201
+ }
202
+ ],
203
+ "existing_required_labels": []
204
+ },
205
+ {
206
+ "query_id": "mcsa_711_2esd_assembly1",
207
+ "compounds": [
208
+ {
209
+ "count": 1,
210
+ "type": "reactant",
211
+ "chebi_id": "58349",
212
+ "name": "NADP(3-)",
213
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/58349_FtUFgHJ.mol"
214
+ }
215
+ ],
216
+ "existing_required_labels": []
217
+ },
218
+ {
219
+ "query_id": "mcsa_741_1kae_assembly1",
220
+ "compounds": [
221
+ {
222
+ "count": 2,
223
+ "type": "reactant",
224
+ "chebi_id": "57540",
225
+ "name": "NAD(1-)",
226
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
227
+ }
228
+ ],
229
+ "existing_required_labels": []
230
+ },
231
+ {
232
+ "query_id": "mcsa_803_1o04_assembly1",
233
+ "compounds": [
234
+ {
235
+ "count": 1,
236
+ "type": "reactant",
237
+ "chebi_id": "57540",
238
+ "name": "NAD(1-)",
239
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
240
+ }
241
+ ],
242
+ "existing_required_labels": []
243
+ },
244
+ {
245
+ "query_id": "mcsa_889_2pgd_assembly1",
246
+ "compounds": [
247
+ {
248
+ "count": 1,
249
+ "type": "reactant",
250
+ "chebi_id": "58349",
251
+ "name": "NADP(3-)",
252
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/58349_FtUFgHJ.mol"
253
+ }
254
+ ],
255
+ "existing_required_labels": []
256
+ },
257
+ {
258
+ "query_id": "mcsa_890_1yve_assembly2",
259
+ "compounds": [
260
+ {
261
+ "count": 1,
262
+ "type": "reactant",
263
+ "chebi_id": "58349",
264
+ "name": "NADP(3-)",
265
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/58349_FtUFgHJ.mol"
266
+ }
267
+ ],
268
+ "existing_required_labels": []
269
+ },
270
+ {
271
+ "query_id": "mcsa_891_1ybv_assembly1",
272
+ "compounds": [
273
+ {
274
+ "count": 1,
275
+ "type": "reactant",
276
+ "chebi_id": "58349",
277
+ "name": "NADP(3-)",
278
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/58349_FtUFgHJ.mol"
279
+ }
280
+ ],
281
+ "existing_required_labels": []
282
+ },
283
+ {
284
+ "query_id": "mcsa_902_2nac_assembly1",
285
+ "compounds": [
286
+ {
287
+ "count": 1,
288
+ "type": "reactant",
289
+ "chebi_id": "57540",
290
+ "name": "NAD(1-)",
291
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
292
+ }
293
+ ],
294
+ "existing_required_labels": []
295
+ },
296
+ {
297
+ "query_id": "mcsa_92_1dli_assembly1",
298
+ "compounds": [
299
+ {
300
+ "count": 2,
301
+ "type": "reactant",
302
+ "chebi_id": "57540",
303
+ "name": "NAD(1-)",
304
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
305
+ }
306
+ ],
307
+ "existing_required_labels": []
308
+ },
309
+ {
310
+ "query_id": "mcsa_766_1s3i_assembly1_reviewed_v2",
311
+ "compounds": [
312
+ {
313
+ "count": 1,
314
+ "type": "reactant",
315
+ "chebi_id": "58349",
316
+ "name": "NADP(3-)",
317
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/58349_FtUFgHJ.mol"
318
+ }
319
+ ],
320
+ "existing_required_labels": []
321
+ },
322
+ {
323
+ "query_id": "mcsa_984_4e5k_assembly1_reviewed_v2",
324
+ "compounds": [
325
+ {
326
+ "count": 1,
327
+ "type": "reactant",
328
+ "chebi_id": "57540",
329
+ "name": "NAD(1-)",
330
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
331
+ }
332
+ ],
333
+ "existing_required_labels": []
334
+ }
335
+ ],
336
+ "evaluable": [
337
+ {
338
+ "query_id": "mcsa_312_1kq3_assembly1",
339
+ "compounds": [
340
+ {
341
+ "count": 1,
342
+ "type": "reactant",
343
+ "chebi_id": "57540",
344
+ "name": "NAD(1-)",
345
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
346
+ }
347
+ ],
348
+ "existing_required_labels": [
349
+ "Zn2+"
350
+ ]
351
+ },
352
+ {
353
+ "query_id": "mcsa_500_1e3i_assembly1",
354
+ "compounds": [
355
+ {
356
+ "count": 1,
357
+ "type": "reactant",
358
+ "chebi_id": "57540",
359
+ "name": "NAD(1-)",
360
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
361
+ }
362
+ ],
363
+ "existing_required_labels": [
364
+ "Zn2+"
365
+ ]
366
+ },
367
+ {
368
+ "query_id": "mcsa_501_reviewed_v2",
369
+ "compounds": [
370
+ {
371
+ "count": 1,
372
+ "type": "reactant",
373
+ "chebi_id": "57540",
374
+ "name": "NAD(1-)",
375
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
376
+ }
377
+ ],
378
+ "existing_required_labels": [
379
+ "Zn2+"
380
+ ]
381
+ }
382
+ ]
383
+ },
384
+ "frozen_counts_are_provisional": true,
385
+ "selection_outputs_changed": false,
386
+ "reason": "Coverage mixes database specificity, extraction policy, role-table gaps and confirmed name-joining omissions; three evaluable groups omit a required NAD role label."
387
+ }
annotations/complete_ligand_sets_20261002/README.md CHANGED
@@ -1,5 +1,11 @@
1
  # Complete catalytic reference-ligand matching
2
 
 
 
 
 
 
 
3
  This experiment asks whether a baseline-validated motif hit has every **named,
4
  documented catalysis-associated binding label** found in its native reference.
5
  It applies the user's 2 October 2026 restriction: incidental, regulatory,
 
1
  # Complete catalytic reference-ligand matching
2
 
3
+ **Version note:** this folder preserves the original v2 implementation and
4
+ outputs. The [coverage audit](coverage_audit/REPORT.md) identifies confirmed
5
+ normalization omissions. Corrected v3 and the full matched BioLiP2 comparison
6
+ live in `../biolip_interpro_catalytic_comparison_20261002/`; use that release for
7
+ subsequent dataset selection.
8
+
9
  This experiment asks whether a baseline-validated motif hit has every **named,
10
  documented catalysis-associated binding label** found in its native reference.
11
  It applies the user's 2 October 2026 restriction: incidental, regulatory,
annotations/complete_ligand_sets_20261002/REPORT.md CHANGED
@@ -1,5 +1,11 @@
1
  # Full reference-ligand-set matching: catalytic labels only
2
 
 
 
 
 
 
 
3
  Completed 2026-10-02T07:23:45.714912+00:00. Policy: `interpro_complete_catalytic_ligands_20261002_v2`.
4
 
5
  ## Question and method
 
1
  # Full reference-ligand-set matching: catalytic labels only
2
 
3
+ **Historical v2 result.** The [reference coverage audit](coverage_audit/REPORT.md)
4
+ found confirmed normalization omissions, including three groups with an omitted
5
+ NAD requirement. Use the corrected v3 and matched BioLiP2 comparison in
6
+ `../biolip_interpro_catalytic_comparison_20261002/` for subsequent selection.
7
+ The original counts below are preserved rather than silently rewritten.
8
+
9
  Completed 2026-10-02T07:23:45.714912+00:00. Policy: `interpro_complete_catalytic_ligands_20261002_v2`.
10
 
11
  ## Question and method
annotations/complete_ligand_sets_20261002/coverage_audit/REPORT.md ADDED
@@ -0,0 +1,97 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Why only 245 InterPro reference sets were initially evaluable
2
+
3
+ The original 245/876 is **coverage of one annotation-and-role-matching
4
+ implementation**, not a count of enzymes that use ligands. No reference was
5
+ established to be ligand-free by this analysis. The audit identified real
6
+ normalization omissions, so the original v2 selection is historical and is
7
+ superseded by the separately versioned comparison.
8
+
9
+ ## What was audited
10
+
11
+ The population is all 876 frozen motif queries, representing 863 M-CSA entries.
12
+ Their available reference chains produce 880 distinct native strings. We read
13
+ the completed exact-string InterPro Matches API/fresh-scan cache, not new
14
+ predictions, generated sequences, or annotations retrieved only by accession.
15
+
16
+ Selected traditional evidence is Pfam 38.2, CDD 3.21, PIRSR 2025_05,
17
+ PROSITE patterns/profiles 2026_01 and SFLD 4. InterPro-N and other neural sources
18
+ are excluded before library-version checks. Reference catalytic roles came from
19
+ current M-CSA cofactor pages retrieved on 2 October, their historical curated
20
+ mechanistic-role table, the frozen 23 September reaction snapshot, and explicit
21
+ catalytic-site descriptors. Presence in the cofactor list alone was insufficient.
22
+
23
+ We traced every query through reference mapping, annotation retrieval, literal
24
+ ligand extraction, and catalytic-role joining. Categories below are mutually
25
+ exclusive and sum to 876. `per_reference.jsonl` contains the underlying records.
26
+
27
+ ## The original v2 coverage funnel
28
+
29
+ | Last successful stage / reason for unresolved coverage | Queries | Queries with a cofactor listed in current M-CSA |
30
+ |---|---:|---:|
31
+ | Evaluable nonempty catalytic reference set | 245 | 227 |
32
+ | No binding descriptor extracted | 250 | 93 |
33
+ | Only generic or explicitly excluded descriptors | 112 | 51 |
34
+ | Specific wording remained unmapped or ambiguous | 53 | 27 |
35
+ | Named ligand found, but no accepted catalytic-role join | 105 | 66 |
36
+ | Relevant specific identity unresolved, blocking completeness | 24 | 24 |
37
+ | Reference mapping or existing length/sequence scope unresolved | 87 | 58 |
38
+
39
+ The middle four rows account for the earlier combined **520** unknown cases.
40
+ Of these, **237 actually have a cofactor listed in M-CSA**. The other 283 have no
41
+ listed cofactor in the retrieved field; that does not establish absence of a
42
+ substrate, cofactor, or other ligand.
43
+
44
+ There were **zero missing cached API records** among the 880 available strings,
45
+ and no incompatible selected-library records. Of those strings, 879 had a
46
+ selected traditional-library match. Thus the main problem was not an API outage.
47
+ InterPro can identify a family or domain without providing a specific ligand ID.
48
+
49
+ The 87 reference-mapping/scope cases partition into 45 overlength references,
50
+ 25 multiple-distinct-string mappings, 9 uncertain mappings, 6 missing references,
51
+ and 2 nonstandard strings. These are input-policy or mapping issues, not evidence
52
+ about ligand dependence. In particular, the existing 1,024-residue cap on
53
+ reference annotation is a pipeline choice; it is not logically required just
54
+ because training endpoints have that cap. It was held fixed for this comparison.
55
+
56
+ ## Concrete examples
57
+
58
+ | Reference | What the records say | Why v2 did not establish a usable set |
59
+ |---|---|---|
60
+ | Phospholipase C, M-CSA 27 | M-CSA records catalytic zinc; CDD/Pfam/PROSITE explicitly say zinc-dependent | Our title extractor required binding/ligand wording and ignored the dependency title. This was an extraction omission. |
61
+ | Trypsin, M-CSA 173 | Compatible InterPro matches include a generic substrate-binding descriptor | This does not identify a specific ligand. It is not a missing API result. |
62
+ | RNase HI, M-CSA 163 | InterPro describes RNA/DNA-hybrid binding; M-CSA lists magnesium | The retrieved descriptor does not establish a specific magnesium-binding label. |
63
+ | Betaine-aldehyde dehydrogenase, M-CSA 100 | InterPro supplies NAD labels; the reaction lists NAD(1−) | A charge-qualified name failed to join the NAD role family. This was a normalization bug. |
64
+ | Cytokinin dehydrogenase, M-CSA 115 | FAD binding and catalytic FAD roles are present | Coordinated wording such as “FAD and cytokinin binding” was parsed too narrowly. |
65
+ | Dialkylglycine decarboxylase, M-CSA 482 | PLP is annotated and listed as a cofactor | The historical role-table field used by the selector was empty. The current M-CSA narrative does describe a catalytic PLP role. This is a role-source coverage limitation, not proof that PLP is neutral. |
66
+ | Sulfite oxidase, M-CSA 121 | Heme and molybdenum-cofactor descriptors are present | Specific cofactor identity remained unresolved under the conservative InterPro normalizer. |
67
+ | Glutamate racemase, M-CSA 1 | No cofactor field was extracted; the curated reaction still names glutamate | An empty cofactor field cannot be interpreted as “no ligand.” |
68
+
69
+ The [M-CSA PLP example](https://www.ebi.ac.uk/thornton-srv/m-csa/entry/482/)
70
+ illustrates why absence from one structured role field is weaker than absence
71
+ of documented evidence in the resource as a whole. No new narrative-derived
72
+ roles were silently added to this comparison.
73
+
74
+ ## Confirmed corrections and their consequences
75
+
76
+ The NAD(1−)/NADP(3−) role-family error affected 20 previously unevaluable queries.
77
+ It also omitted NAD from the nonempty requirements of M-CSA 312, 500 and 501.
78
+ Those three groups had 6,107 supported memberships under v2, so the earlier
79
+ selector was not simply conservative: it could also require too few labels.
80
+
81
+ ChEBI identifies [NAD(1−)](https://www.ebi.ac.uk/chebi/CHEBI:57540) as a
82
+ protonation form of NAD+. This correction concerns reference role evidence;
83
+ it does not let arbitrary redox or ligand identities substitute for one another
84
+ when matching reference and hit annotations.
85
+
86
+ Version 3 separately fixes the role join, retains explicit ligand-dependency
87
+ titles, and handles the observed coordinated FAD-binding wording. It reparses
88
+ the full saved cache identically for references and hits. It gives **274
89
+ evaluable references, 6,568,858 supported unique motif–sequence memberships,
90
+ 5,018,150 distinct supported strings and 206 supported groups**. The remaining
91
+ reference statuses are 491 without a usable named catalytic set, 24 with
92
+ unresolved relevant identity, and the same 87 mapping/scope cases.
93
+
94
+ Original v2 artifacts and counts remain intact. Version 3 and the BioLiP2
95
+ comparison are in `../../biolip_interpro_catalytic_comparison_20261002/`.
96
+ These are annotation-supported subset indices, not activity measurements or
97
+ aligned training-pair releases.
annotations/complete_ligand_sets_20261002/coverage_audit/per_reference.csv ADDED
The diff for this file is too large to render. See raw diff
 
annotations/complete_ligand_sets_20261002/coverage_audit/per_reference.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
annotations/complete_ligand_sets_20261002/coverage_audit/summary.json ADDED
@@ -0,0 +1,387 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "frozen_policy": "interpro_complete_catalytic_ligands_20261002_v2",
3
+ "query_groups": 876,
4
+ "reference_strings": 880,
5
+ "cache_not_found": 0,
6
+ "references_with_annotation_issues": 0,
7
+ "reference_strings_with_selected_matches": 879,
8
+ "breakdown": {
9
+ "reference_mapping_or_scope": {
10
+ "groups": 87,
11
+ "cofactor_listed": 58,
12
+ "no_selected_matches": 6,
13
+ "charged_name_join_flags": 0,
14
+ "role_table_gap": 8,
15
+ "named_dependency_title": 1
16
+ },
17
+ "no_extracted_binding_descriptors": {
18
+ "groups": 250,
19
+ "cofactor_listed": 93,
20
+ "no_selected_matches": 1,
21
+ "charged_name_join_flags": 0,
22
+ "role_table_gap": 0,
23
+ "named_dependency_title": 7
24
+ },
25
+ "evaluable": {
26
+ "groups": 245,
27
+ "cofactor_listed": 227,
28
+ "no_selected_matches": 0,
29
+ "charged_name_join_flags": 3,
30
+ "role_table_gap": 17,
31
+ "named_dependency_title": 14
32
+ },
33
+ "named_labels_without_role_join": {
34
+ "groups": 105,
35
+ "cofactor_listed": 66,
36
+ "no_selected_matches": 0,
37
+ "charged_name_join_flags": 20,
38
+ "role_table_gap": 42,
39
+ "named_dependency_title": 14
40
+ },
41
+ "only_generic_or_excluded_descriptors": {
42
+ "groups": 112,
43
+ "cofactor_listed": 51,
44
+ "no_selected_matches": 0,
45
+ "charged_name_join_flags": 0,
46
+ "role_table_gap": 0,
47
+ "named_dependency_title": 3
48
+ },
49
+ "specific_identity_blocks_reference": {
50
+ "groups": 24,
51
+ "cofactor_listed": 24,
52
+ "no_selected_matches": 0,
53
+ "charged_name_join_flags": 0,
54
+ "role_table_gap": 1,
55
+ "named_dependency_title": 2
56
+ },
57
+ "unmapped_or_unresolved_descriptors": {
58
+ "groups": 53,
59
+ "cofactor_listed": 27,
60
+ "no_selected_matches": 0,
61
+ "charged_name_join_flags": 0,
62
+ "role_table_gap": 0,
63
+ "named_dependency_title": 0
64
+ }
65
+ },
66
+ "mapping_priority_breakdown": {
67
+ "overlength": 45,
68
+ "multiple_distinct_reference_strings": 25,
69
+ "mapping_uncertain": 9,
70
+ "missing": 6,
71
+ "nonstandard": 2
72
+ },
73
+ "charged_role_join_affected": {
74
+ "unknown_no_named_catalytic_reference_labels": [
75
+ {
76
+ "query_id": "mcsa_100_1a4s_assembly1",
77
+ "compounds": [
78
+ {
79
+ "count": 1,
80
+ "type": "reactant",
81
+ "chebi_id": "57540",
82
+ "name": "NAD(1-)",
83
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
84
+ }
85
+ ],
86
+ "existing_required_labels": []
87
+ },
88
+ {
89
+ "query_id": "mcsa_240_1szd_assembly1",
90
+ "compounds": [
91
+ {
92
+ "count": 1,
93
+ "type": "reactant",
94
+ "chebi_id": "57540",
95
+ "name": "NAD(1-)",
96
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
97
+ }
98
+ ],
99
+ "existing_required_labels": []
100
+ },
101
+ {
102
+ "query_id": "mcsa_336_2hdh_assembly1",
103
+ "compounds": [
104
+ {
105
+ "count": 1,
106
+ "type": "reactant",
107
+ "chebi_id": "57540",
108
+ "name": "NAD(1-)",
109
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
110
+ }
111
+ ],
112
+ "existing_required_labels": []
113
+ },
114
+ {
115
+ "query_id": "mcsa_429_1brm_assembly1",
116
+ "compounds": [
117
+ {
118
+ "count": 1,
119
+ "type": "reactant",
120
+ "chebi_id": "58349",
121
+ "name": "NADP(3-)",
122
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/58349_FtUFgHJ.mol"
123
+ }
124
+ ],
125
+ "existing_required_labels": []
126
+ },
127
+ {
128
+ "query_id": "mcsa_450_1dgs_assembly1",
129
+ "compounds": [
130
+ {
131
+ "count": 1,
132
+ "type": "reactant",
133
+ "chebi_id": "57540",
134
+ "name": "NAD(1-)",
135
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
136
+ }
137
+ ],
138
+ "existing_required_labels": []
139
+ },
140
+ {
141
+ "query_id": "mcsa_458_1a4i_assembly1",
142
+ "compounds": [
143
+ {
144
+ "count": 1,
145
+ "type": "reactant",
146
+ "chebi_id": "58349",
147
+ "name": "NADP(3-)",
148
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/58349_FtUFgHJ.mol"
149
+ }
150
+ ],
151
+ "existing_required_labels": []
152
+ },
153
+ {
154
+ "query_id": "mcsa_576_1j49_assembly1",
155
+ "compounds": [
156
+ {
157
+ "count": 1,
158
+ "type": "reactant",
159
+ "chebi_id": "57540",
160
+ "name": "NAD(1-)",
161
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
162
+ }
163
+ ],
164
+ "existing_required_labels": []
165
+ },
166
+ {
167
+ "query_id": "mcsa_584_1ldm_assembly1",
168
+ "compounds": [
169
+ {
170
+ "count": 1,
171
+ "type": "reactant",
172
+ "chebi_id": "57540",
173
+ "name": "NAD(1-)",
174
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
175
+ }
176
+ ],
177
+ "existing_required_labels": []
178
+ },
179
+ {
180
+ "query_id": "mcsa_607_1r16_assembly1",
181
+ "compounds": [
182
+ {
183
+ "count": 1,
184
+ "type": "reactant",
185
+ "chebi_id": "57540",
186
+ "name": "NAD(1-)",
187
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
188
+ }
189
+ ],
190
+ "existing_required_labels": []
191
+ },
192
+ {
193
+ "query_id": "mcsa_67_1pjb_assembly1",
194
+ "compounds": [
195
+ {
196
+ "count": 1,
197
+ "type": "reactant",
198
+ "chebi_id": "57540",
199
+ "name": "NAD(1-)",
200
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
201
+ }
202
+ ],
203
+ "existing_required_labels": []
204
+ },
205
+ {
206
+ "query_id": "mcsa_711_2esd_assembly1",
207
+ "compounds": [
208
+ {
209
+ "count": 1,
210
+ "type": "reactant",
211
+ "chebi_id": "58349",
212
+ "name": "NADP(3-)",
213
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/58349_FtUFgHJ.mol"
214
+ }
215
+ ],
216
+ "existing_required_labels": []
217
+ },
218
+ {
219
+ "query_id": "mcsa_741_1kae_assembly1",
220
+ "compounds": [
221
+ {
222
+ "count": 2,
223
+ "type": "reactant",
224
+ "chebi_id": "57540",
225
+ "name": "NAD(1-)",
226
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
227
+ }
228
+ ],
229
+ "existing_required_labels": []
230
+ },
231
+ {
232
+ "query_id": "mcsa_803_1o04_assembly1",
233
+ "compounds": [
234
+ {
235
+ "count": 1,
236
+ "type": "reactant",
237
+ "chebi_id": "57540",
238
+ "name": "NAD(1-)",
239
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
240
+ }
241
+ ],
242
+ "existing_required_labels": []
243
+ },
244
+ {
245
+ "query_id": "mcsa_889_2pgd_assembly1",
246
+ "compounds": [
247
+ {
248
+ "count": 1,
249
+ "type": "reactant",
250
+ "chebi_id": "58349",
251
+ "name": "NADP(3-)",
252
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/58349_FtUFgHJ.mol"
253
+ }
254
+ ],
255
+ "existing_required_labels": []
256
+ },
257
+ {
258
+ "query_id": "mcsa_890_1yve_assembly2",
259
+ "compounds": [
260
+ {
261
+ "count": 1,
262
+ "type": "reactant",
263
+ "chebi_id": "58349",
264
+ "name": "NADP(3-)",
265
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/58349_FtUFgHJ.mol"
266
+ }
267
+ ],
268
+ "existing_required_labels": []
269
+ },
270
+ {
271
+ "query_id": "mcsa_891_1ybv_assembly1",
272
+ "compounds": [
273
+ {
274
+ "count": 1,
275
+ "type": "reactant",
276
+ "chebi_id": "58349",
277
+ "name": "NADP(3-)",
278
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/58349_FtUFgHJ.mol"
279
+ }
280
+ ],
281
+ "existing_required_labels": []
282
+ },
283
+ {
284
+ "query_id": "mcsa_902_2nac_assembly1",
285
+ "compounds": [
286
+ {
287
+ "count": 1,
288
+ "type": "reactant",
289
+ "chebi_id": "57540",
290
+ "name": "NAD(1-)",
291
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
292
+ }
293
+ ],
294
+ "existing_required_labels": []
295
+ },
296
+ {
297
+ "query_id": "mcsa_92_1dli_assembly1",
298
+ "compounds": [
299
+ {
300
+ "count": 2,
301
+ "type": "reactant",
302
+ "chebi_id": "57540",
303
+ "name": "NAD(1-)",
304
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
305
+ }
306
+ ],
307
+ "existing_required_labels": []
308
+ },
309
+ {
310
+ "query_id": "mcsa_766_1s3i_assembly1_reviewed_v2",
311
+ "compounds": [
312
+ {
313
+ "count": 1,
314
+ "type": "reactant",
315
+ "chebi_id": "58349",
316
+ "name": "NADP(3-)",
317
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/58349_FtUFgHJ.mol"
318
+ }
319
+ ],
320
+ "existing_required_labels": []
321
+ },
322
+ {
323
+ "query_id": "mcsa_984_4e5k_assembly1_reviewed_v2",
324
+ "compounds": [
325
+ {
326
+ "count": 1,
327
+ "type": "reactant",
328
+ "chebi_id": "57540",
329
+ "name": "NAD(1-)",
330
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
331
+ }
332
+ ],
333
+ "existing_required_labels": []
334
+ }
335
+ ],
336
+ "evaluable": [
337
+ {
338
+ "query_id": "mcsa_312_1kq3_assembly1",
339
+ "compounds": [
340
+ {
341
+ "count": 1,
342
+ "type": "reactant",
343
+ "chebi_id": "57540",
344
+ "name": "NAD(1-)",
345
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
346
+ }
347
+ ],
348
+ "existing_required_labels": [
349
+ "Zn2+"
350
+ ]
351
+ },
352
+ {
353
+ "query_id": "mcsa_500_1e3i_assembly1",
354
+ "compounds": [
355
+ {
356
+ "count": 1,
357
+ "type": "reactant",
358
+ "chebi_id": "57540",
359
+ "name": "NAD(1-)",
360
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
361
+ }
362
+ ],
363
+ "existing_required_labels": [
364
+ "Zn2+"
365
+ ]
366
+ },
367
+ {
368
+ "query_id": "mcsa_501_reviewed_v2",
369
+ "compounds": [
370
+ {
371
+ "count": 1,
372
+ "type": "reactant",
373
+ "chebi_id": "57540",
374
+ "name": "NAD(1-)",
375
+ "mol_file": "www.ebi.ac.uk/thornton-srv/m-csa/media/compound_mols/57540_mTGTBRN.mol"
376
+ }
377
+ ],
378
+ "existing_required_labels": [
379
+ "Zn2+"
380
+ ]
381
+ }
382
+ ]
383
+ },
384
+ "frozen_counts_are_provisional": true,
385
+ "selection_outputs_changed": false,
386
+ "reason": "Coverage mixes database specificity, extraction policy, role-table gaps and confirmed name-joining omissions; three evaluable groups omit a required NAD role label."
387
+ }