2026 Sang - AF3-TurboAb antibody structural prediction
AF3-TurboAb is a practical repertoire-scale wrapper around AF3 for antibody-antigen complexes. It removes the CPU preprocessing bottleneck by using antibody-specific MSA/template libraries and reusing antigen features, then validates ipTM as an operational confidence metric for antibody/nanobody interfaces.
For STRC, this is a method precedent for high-throughput interface triage, decoy specificity checks, and atlas-style epitope mapping. It is not direct evidence that AF3 ipTM predicts affinity for arbitrary STRC protein complexes, peptides, or small molecules.
Citation
Sang Z, Xiang Y, Huang W, Sargunas PR, Kim YJJ, Feng Z, Taylor DJ, Shi Y. “Repertoire-scale antibody structural prediction informs therapeutic design.” Science Advances 12(17), eaef7163, 2026. DOI: 10.1126/sciadv.aef7163.
Numbers that matter
| Claim | Value | Source in paper | STRC use |
|---|---|---|---|
| Standard AF3 complex runtime | ~17 min per complex | Fig. 1A text | Shows preprocessing, not inference, is the scale blocker |
| CPU MSA/template share | 97.2% of runtime | Fig. 1A caption | Reuse/precompute features before any large STRC AF3 sweep |
| Nb MSA runtime | ~670 s to 1.6 s | Results / Fig. 1C | 418x speedup from antibody-specific search space |
| End-to-end TurboAb runtime | ~35 s per seed | Results / fig. S1A | Practical million-scale screening with GPU parallelism |
| 1M Nb screen, single GPU | 8055 h vs 294,722 h standard AF3 | Results | Speedup changes the experimental design space |
| Parallelized 1M Nb screen | 8.4 days on 40 GPUs | Results | Academic-cluster feasible |
| Posttraining Nb benchmark | 243 complexes, ⇐4 A, post-2022 | Fig. 1C/E; Methods | Leakage-reduced validation set |
| TurboAb vs standard AF3 ipTM | Pearson r = 0.95 | Fig. 1E | Optimization preserves AF3 confidence behavior |
| High-confidence threshold | mean ipTM >= 0.8 | Results | Use only as class-calibrated antibody interface threshold |
| High-confidence structural quality | DockQ > 0.49, median epitope overlap 90% | Results | ipTM tracks near-experimental antibody interfaces |
| Cognate vs decoy hit rates | ~23% vs 0.2% and 0.14% | Fig. 1J | Decoy panels are mandatory for specificity claims |
| In-house affinity panel | high 43.8%, medium 16.7%, low 3.2% high-confidence | fig. S2/table S1 | ipTM enriches affinity class, but is not a Kd predictor |
| Repertoire atlas size | 275,371 complexes, 3 seeds each | Results | Demonstrates atlas-scale structural decoding |
| High-confidence atlas models | 39,973 total; 28,013 distinct CDR sequences | Results | Expands structural coverage beyond PDB |
| Training-overlap check | 94.4% have <60% CDR3 identity to PDB CDR3s | Results | Useful leakage-control pattern |
| Cryo-EM validation | 12 HSA-Nb complexes at 3.1 to 3.8 A | Fig. 2 | Multi-epitope experimental validation |
| Cryo-EM agreement | all epitopes correct; 10/12 close by RMSD | Results | Confirms epitope identity more robustly than full antigen conformation |
| HSA induced fit | ~19 deg / 12.7 A and ~15 deg / 8.5 A shifts | Results | AF3 can miss antigen conformational rearrangement |
| Predicted epitope surface coverage | ~60 to 100% of solvent-accessible surface | Fig. 3 | Atlas exposes unmapped antigen surfaces |
| Hotspot architecture | 95% of interfaces contain 1 to 5 hotspots | Fig. 3K/L | Compact residue clusters dominate contacts |
| Hotspot contact share | 31% average; up to 86% with 5 hotspots | Fig. 3L | Design should target compact contact pillars |
| Hotspot chemistry | 83.6% top seven residues, aromatic or charged/polar enriched | Fig. 3N | Matches salt-bridge/H-bond/aromatic interface intuition |
| SARS-CoV-2 RBD triparatopic designs | cover 41.6 to 44.4% of RBD surface | Fig. 5H | Multiparatopic design can combine breadth and potency |
| JN.1 neutralization gain | up to 2716x EC50 improvement over matched monomers | Fig. 5L | Avidity plus epitope complementarity can be very large |
| PD-L1 screen | 700 Nbs modeled; 14 high-confidence hits | Methods | Glycan-aware AF3 inputs can select domain-specific binders |
| PD-L1 P10 binding | 6 nM human, 19 nM mouse cell EC50 | Results | Cross-species binders can emerge from epitope selection |
| PD-L1 glycoform effect | deglycosylation weakens ELISA 4.9 to 27 nM | Results | Glycoform modeling can matter structurally |
| MERS virtual screen | 20,000 Nbs to 25 seeds, expanded to 6290, then 566 hits | Results/Methods | Two-stage seed plus neighbor expansion workflow |
| MERS prospective hit rate | 5/19 expressed binders, 26.3% | Results | ipTM-only triage can enrich de novo binders |
Method takeaway
- The decisive engineering move is not a new neural model. It is replacing repeated full-database MSA/template searches with antibody-focused databases plus precomputed antigen features.
- Mean ipTM >= 0.8 is useful in the validated antibody-antigen regime. Do not import this cutoff as an affinity threshold for STRC/TMEM145, WH2-actin, homodimers, or small-molecule complexes without a local benchmark and decoys.
- The paper’s specificity logic is the part STRC should copy: cognate sets, biologically implausible decoys, repeated seeds, and explicit leakage checks.
- The HSA cryo-EM cases are a good caution: epitope identity can be right while antigen subdomain conformation shifts after binding. For STRC, AF3 interface success still needs geometry/MD/experiment when induced fit is plausible.
- The hotspot analysis is design-relevant beyond antibodies: compact aromatic/charged/polar patches can dominate contact counts, but translating this to STRC requires target-specific validation.
Relevance to STRC
h03 mini-STRC: supports continued use of AF3 as a structural gate only when paired with leakage control, decoys, and experimental follow-up. It does not change the h03 rank because h03 evidence already rests on STRC/TMEM145-specific AF3 and vector audit results.
h09 hydrogel: reinforces the need for decoy and threshold discipline in any AF3 peptide-interface gate. It does not rescue weak confidence-supported WH2-actin contacts, because the validated regime here is antibody-antigen, not short peptide-actin.
h26 engineered homodimer: does not reopen the paused branch. h26 failed its own AF3 G1 structural confidence gate; antibody repertoire screening does not provide a new homodimer scaffold or interface.
Contrast with Harvey 2026
2026-harvey-afm-nanobody-gpcr showed AF-M nanobody screening can enrich GPCR binders but transfers poorly to soluble and non-GPCR targets. Sang 2026 is stronger for antibody-antigen atlas work because it has posttraining benchmarks, decoys, HSA cryo-EM, and prospective MERS validation. The shared STRC lesson is still conservative: structural-confidence enrichment is target-class-specific.
STRC ranking impact
No ranking change. This paper improves method discipline and supports AF3-scale screening workflows, but it does not directly validate or falsify any STRC therapeutic hypothesis.
Follow-up
- If a future STRC binder-library idea emerges, copy the paper’s decoy and leakage-control design before trusting ipTM.
- If using AF3 at scale for STRC proteins, first identify the reusable feature layer analogous to antigen precomputation.
- If citing this paper in a phase proof, cite the specific validated regime: antibody/nanobody-antigen complex prediction with ipTM >= 0.8, not general protein-protein affinity.
- Future AF3/interface screens should start from STRC AF3 Interface Triage Protocol.
Connections
[see-also]AlphaFold 3 Server[see-also]STRC AF3 Static Pocket Blindness to Loop Dynamics[supports]STRC AF3 Interface Triage Protocol[see-also]2026-harvey-afm-nanobody-gpcr[see-also]MRGPRX2 AF-M screen[applies]H03 Mini-STRC Single-Vector[applies]H09 Synthetic Peptide Hydrogel[applies]H26 Engineered Homodimer