Status
successful_prediction
Median lDDT
0.9708
Public Path
Workers -> science.eigenatlas.com -> M3 Max
Run Boundary
This is still a fixture-controlled engineering smoke, not a publication-grade biological benchmark. Its value is architectural: the public route can now perform real inference and scoring.
Per-Target Results
| Target | pLDDT mean | Wall clock | lDDT | TM-score | RMSD |
|---|---|---|---|---|---|
1L2Y | 97.2065 | 112.86 s | 0.9708 | 0.6739 | 0.39 |
1CRN | 90.5352 | 111.36 s | 0.9867 | 0.9738 | 0.35 |
1UBQ | 87.9837 | 110.86 s | 0.9650 | 0.9659 | 0.81 |
Verdict
The deterministic fixture verdict reports median lDDT 0.9707634092773411 versus fixture value 0.9694790633021089, a relative error of 0.1324779486065049%.
Artifacts
card.json
Machine-readable result card with public-tunnel predictions, scores, validation, and audit count.
predictions.json
Per-target prediction records and predicted PDB text.
scores.json
Per-target lDDT, TM-score, and RMSD.
verdict.json
Deterministic fixture validation verdict.
Remaining Gate
The major remaining Phase 0 gate is no-fixture Reader extraction. Until that works, the project should not claim paper reproduction.