Explore model lineages, foundation-model dependencies, benchmark peers, open reimplementations, workflow complements and platform integrations without reducing the ecosystem to a flat list.
51connected nodes43visible relationships
Structure Prediction · 2020
AlphaFold 2 / 3
The model that solved the 50-year protein-folding problem.
Google DeepMind
Recorded relationships
Open ReimplementationOpenFold
OpenFold reimplements the AlphaFold architecture for trainable open research.
Architectural LineageChai-1 / Chai-2
Chai belongs to the post-AlphaFold generation of multimodal complex predictors and design systems.
Architectural LineageBoltz-1 / Boltz-2
Boltz provides an open AlphaFold 3-class complex-prediction lineage.
Open ReimplementationProtenix
Protenix is an open, trainable AlphaFold 3-style reproduction.
Workflow ComplementAlphaProteo
AlphaProteo extends DeepMind's structure work into target-conditioned binder design.
Open ReimplementationHelixFold3
HelixFold3 is an open academic AF3-class biomolecular structure-prediction effort.
Edges are either explicitly curated scientific or product lineages, or same-organization relationships derived from the dataset. They do not imply technical compatibility or endorsement.
β
Ask BioAtlasFREE
Evidence-aware scientific research copilotConnecting…
β
BIOATLAS RESEARCH COPILOT
Ask, research, compare, verify or plan.
Ask BioAtlas now orchestrates the native model registry, evidence claims, Watchtower, frontier research, benchmark governance and scientific planning layers before synthesis.