Skip to page content
SAVRN
Search Plan a project

Open-weight model

Proteus-Structures

by Inserloft Research InserloftResearch/Proteus-Structures

Proteus-Structures is an open-weight model from Inserloft Research, released under Creative Commons Attribution 4.0. Its published files total 463.1 KB.

Computationally generated and predicted molecular structures from Project Proteus, the biological discovery research program of Inserloft Research.

Parameters—
Context—
Weights463.1 KB
Licensecc-by-4.0
AccessOpen weights
Monthly Downloads—

Model Card

By Inserloft Research, published under cc-by-4.0, revision 89974befe763.

Computationally generated and predicted molecular structures from Project Proteus, the biological discovery research program of Inserloft Research. This repository contains structural representations associated with computational protein-design experiments conducted within Project Proteus. Proteus Structures is a structural repository for molecular candidates generated or investigated through the Proteus research program. The structures in this repository are intended to support: Unless explicitly stated otherwise, structures contained here are computational models and should not be interpreted as experimentally determined structures. Additional structures may be added as Proteus research…

Read Inserloft Research's full model card

Computationally generated and predicted molecular structures from Project Proteus, the biological discovery research program of Inserloft Research.

AI for the next generation of biological discovery.

This repository contains structural representations associated with computational protein-design experiments conducted within Project Proteus.


About

Proteus Structures is a structural repository for molecular candidates generated or investigated through the Proteus research program.

The structures in this repository are intended to support:

  • Structural inspection
  • Computational analysis
  • Protein-design research
  • Reproducibility
  • Visualization
  • Further experimental investigation

Unless explicitly stated otherwise, structures contained here are computational models and should not be interpreted as experimentally determined structures.


Repository Structure

Proteus-Structures/
│
├── CNB-1/
│   ├── Macromolecular_PET_Hydrolase_Alpha.pdb
│   ├── Alanine-Glycine-Aspartate.pdb
│   └── ...
│
└── README.md

Additional structures may be added as Proteus research progresses.


CNB-1

The first experimental model developed for Project Proteus is Caelis Neural Base-1 (CNB-1).

CNB-1 has been used to investigate computational protein design and generate defined molecular candidates for further structural and biological analysis.

Documented candidates

Macromolecular PET Hydrolase Alpha

A computationally designed protein candidate of approximately 350 amino acids.

The candidate was generated around a structural design seed associated with natural PET-degrading enzymes and was investigated as a potential hydrolytic enzyme candidate targeting polyethylene terephthalate (PET).

The computational hypothesis involves PET hydrolysis toward:

  • Terephthalic acid
  • Ethylene glycol

Proposed catalytic features remain hypotheses requiring further structural, biochemical, and experimental validation.

AGD-F1

AGD-F1 — Alanine–Glycine–Aspartate Structural Candidate is a 377-amino-acid computationally designed protein candidate generated by CNB-1.

Its structure was computationally predicted using ESMFold.

Current characterization:

Property Status
Length 377 aa
Generation model Caelis Neural Base-1
Structure prediction ESMFold
Experimental structure Not determined
Biological function Uncharacterized
Experimental stability Not determined
Experimental validation Not performed

The AGD-F1 designation reflects prominent amino-acid representation in the candidate sequence and does not imply a demonstrated biochemical function.


Structure Prediction

Some structures in this repository are computational predictions rather than experimentally determined structures.

For predicted structures, the corresponding metadata should identify:

  • The generation model
  • The structure-prediction method
  • The associated candidate
  • The computational nature of the structure
  • Relevant attribution

Computational structure predictions should therefore be treated as structural hypotheses.

A predicted structure does not establish:

  • Biological function
  • Experimental stability
  • Enzymatic activity
  • Binding affinity
  • Folding behavior in vivo
  • Experimental structure
  • Biological efficacy

PDB Files

The repository uses the Protein Data Bank (PDB) format for molecular structure representation.

PDB files can be inspected with molecular visualization software such as:

  • PyMOL
  • UCSF ChimeraX
  • VMD
  • Mol*
  • 3Dmol.js

Web-based Proteus structure viewers may also be provided by Inserloft Research.


Research Philosophy

Proteus is built around a simple research principle:

Generate. Structure. Analyze. Validate. Repeat.

Computational generation is only one stage of the process.

The long-term objective is to move from computational molecular hypotheses toward rigorous structural and experimental investigation.

Proteus explores how AI can assist researchers in investigating biological structures and designing candidate molecules relevant to difficult scientific problems.

For a better world. For a better life.


Scientific Status

The structures in this repository represent ongoing research.

Computational results should not be interpreted as established biological discoveries unless explicitly supported by experimental evidence.

Where experimental validation has not been performed, the relevant structure or candidate should be considered a computational research candidate.


Attribution

Inserloft Research Project Proteus

Research and computational modeling by Inserloft Research.

Where third-party models, predictors, datasets, or software are used, their respective authors, publications, and licensing terms remain applicable.


License

Unless a specific file or directory states otherwise, the repository's licensing terms apply to the structural files and accompanying documentation.

Third-party components, predicted structures, models, datasets, and software may be subject to their own licenses and attribution requirements.

See the individual files and repository metadata for applicable terms.


Disclaimer

These structures are provided for research and computational analysis purposes.

Computationally generated or predicted molecular structures are not equivalent to experimentally determined structures.

Nothing in this repository constitutes a claim that a candidate protein has a validated biological function, therapeutic effect, environmental effect, safety profile, or experimental activity unless such evidence is explicitly documented.


Project

Project Proteus Inserloft Research

AI for the next generation of biological discovery.

Identity and Version

Repository
InserloftResearch/Proteus-Structures
Publisher
Inserloft Research
Task
Not stated by the source
Modality
Other
Library
Not stated by the source
Parameters
Not stated by the source
Languages
en
Revision
89974befe763023980e3d676aa9cd5b49d571178
First published
2026-09-18
Last updated
2026-09-22

Files and Weights

4 files, 463.1 KB in total.

Documentation1 file · 6.1 KB
Other2 files · 455.4 KB
Repository1 file · 1.5 KB
Every file
FileTypeSizeSHA-256
README.mdDocumentation6.1 KB —
CNB-1/Alanine-Glycine-Aspartate.pdbOther231.1 KB —
CNB-1/Macromolecular_PET_Hydrolase_Alpha.pdbOther224.3 KB —
.gitattributesRepository1.5 KB —

License and Download

License
cc-by-4.0
Access
Open weights, no gate
Download from Inserloft Research

Released by Inserloft Research through its official repository on Hugging Face. Read the license.

Questions About Proteus-Structures

Can I use Proteus-Structures commercially?

Yes. Proteus-Structures is released under Creative Commons Attribution 4.0. CC BY 4.0 permits sharing and adapting the work, including commercially, provided the creator is credited and changes are indicated.