What researchers are trying
IsoDDE is presented as a unified drug-design engine trained to reason jointly about ligands, proteins and conformational change instead of docking into a rigid precomputed pocket.
Induced-fit co-folding beyond familiar targets
Can structure models represent large ligand-driven protein rearrangements when the target, pocket or conformational transition is far from training examples?
Why it matters
Real binding sites move. Correctly modeling induced fit could reduce false poses, improve selectivity reasoning and expose conformations missed by rigid docking.
Evidence boundary
Out-of-distribution claims depend strongly on benchmark construction, training-set leakage controls and exact success thresholds.
Organizations represented
Isomorphic Labs
Demonstrated evidence
What has actually been shown.
- Developer-reported improvement on difficult Runs N' Poses generalization cases relative to AlphaFold 3.
Unresolved questions
- How are unseen chemotypes and target families isolated from training data?
- Does structural accuracy translate into enrichment or medicinal-chemistry decisions?
- How stable are alternative conformational ensembles?
Signals to watch next
- Temporal-split benchmarks
- Prospective challenge sets
- Independent co-folding evaluation
- Conformational ensemble outputs
Connected evidence graph
Related BioAtlas model passports.
AlphaFold 2 / 3
The model that solved the 50-year protein-folding problem.
4/7 evidence fields documentedBoltz-1 / Boltz-2
Open-source AF3-quality structure — plus binding affinity.
4/7 evidence fields documentedChai-1 / Chai-2
An AlphaFold3-class complex predictor, made freely usable.
4/7 evidence fields documentedNeuralPLexer / Enchant
Physics-aware structure + multi-task ADMET foundation models.
2/7 evidence fields documentedPrimary and evaluation sources