// Enzyme — ensemble-vs-static demonstration · Protein pipeline, Phase 6 evidence layer
TEM-1 β-lactamase
286 aa (P62593) · predicted structure regenerated live for this page (AlphaFold DB + ANM ensemble + fpocket + cross-frame ranker, same pipeline as every worked example on this site).
// Independently verified against a real PDB structure — not pipeline output
A different kind of finding: the rank-1 pocket only appears in later ANM ensemble frames (member frames 2-3 of 5), absent from the static AlphaFold structure. Cross-walked to the deposited PDB 1M40’s Ambler numbering, it lands on residues 99, 102-107, 110 — squarely inside the independently-published H3-loop-H4 allosteric region (Ambler 99-114), not the specific Bowman-lab cryptic pocket this target was originally chosen to test against. Structural check: whole-chain CA RMSD vs. 1M40 is 0.37 Å, crosswalk-matched over 262 of 286 residues — near-experimental accuracy.
// Sequence
// Findings
Family, precedent, and provenance
Family classification
returned dataBeta-lactamase2 (PF13354) — Beta-lactamase enzyme family
E=1.20e-43 · bit score 149.9 · passes GA threshold: yes
Known ligand precedent
returned data153 total structures in family · 9 distinct ligand scaffolds curated
- 8RMA — A1H1U
- 6WJM — ALA
- 8R2Y — 6M9
- 8R2Z — XR9
- 7Q11 — 6V8
Conservation
returned data56 seed sequences · mean pairwise identity 21.8%
Similar known proteins
returned data- 6NIQ — 43.2% identity · (none)
- 6WJM — 36.4% identity · ALA
- 5NE3 — 36.2% identity · NXL
- 7D5J — 36.1% identity · (none)
- 3W4P — 35.5% identity · (none)
Structure-based (Foldseek)
- 7QOR — TM 1.000 · 100% id (new vs. sequence list)
- 1JWV — TM 1.000 · 100% id (new vs. sequence list)
- 1XXM — TM 1.000 · 99% id (new vs. sequence list)
- 1FQG — TM 1.000 · 100% id (new vs. sequence list)
- 6AYK — TM 1.000 · 99% id (new vs. sequence list)
Interaction fingerprints (Evidence Integration Layer)
no_ligand_bound_structure8 ligand-bound structure(s) exist for this target's Pfam family (PF13354), but none checked aligned to the query at >=50% sequence coverage — likely other members of the same broad family, not this specific protein. Interaction fingerprints require a structure of the query protein itself.
Structural analysis — ranked pocket clusters
returned dataResidue numbers below are pipeline-sequential, with the literature (author-deposited PDB 1M40) number shown in parentheses — 262/286 residues cross-walked.
| Rank | Persistence | Residues |
|---|---|---|
| #1 | 1 | 39(41), 40(42), 41(43), 61(63), 62(64), 63(65), 65(67), 170(172), 171(173), 172(174), 173(175), 174(176), 175(177), 178(180), 180(?), 238(241), 262(266), 263(267), 264(268), 265(269) |
| #2 | 1 | 20(?), 21(?), 22(?), 23(?), 24(26), 27(29), 46(48), 53(55), 54(56), 55(57), 56(58), 255(259), 282(286), 284(288), 285(289), 286(290) |
| #3 | 1 | 61(63), 154(156), 155(157), 156(158), 157(159), 180(?), 182(184), 183(185), 186(188), 187(189), 190(192) |
Pocket functional context (UniProt + ClinVar)
Pocket 1
no UniProt functional features or ClinVar variants overlap with pocket residues
Pocket 2
no UniProt functional features or ClinVar variants overlap with pocket residues
Pocket 3
no UniProt functional features or ClinVar variants overlap with pocket residues
// Limitations & data provenance
Auto-populated from each section’s own status — not hand-maintained.