← Protein structure & design

dEVA

dEVA is a protein structure & design tool whose code is published by gelnesr on GitHub, first public in 2026, described in Science (2026). Design by EVolutionary Algorithm: multi-objective genetic-algorithm protein design that scores sequence proposals on several objectives (e.g. LigandMPNN sequence likelihood plus Metal3D-Cat catalytic-metal probability) and returns the Pareto front; used for zero-shot design of a de novo dizinc metalloenzyme without a natural template.

Visit website ↗

CategoryProtein structure & design
Developer (GitHub)gelnesr
TaskMulti-objective sequence design on a fixed or flexible backbone; any sequence/structure score can be an objective
InterfacePython CLI (python run.py -c config.yml --models seq_model …)
GitHubgelnesr/dEVA
Released2026-11-19 (source)
Paperhttps://www.science.org/doi/10.1126/science.aej3009 · Science (2026)
Code licenseMIT
License notecode MIT (GitHub gelnesr/dEVA and ProteinDesignLab/dEVA); Metal3D-Cat checkpoint licensing not stated in fetched sources
What this allowsPermissive (commercial OK)
Use, modify and ship it inside closed-source commercial products; keep the license text and copyright notices. What do these mean?
License sourceapi.github.com/repos/gelnesr/dEVA · checked 2026-11-20
PricingOpen source
Tags#protein-design#enzyme-design#metalloprotein#genetic-algorithm#ligandmpnn
Last checked · links verified

Can I use dEVA commercially?

Yes. The recorded license is permissive, so closed-source commercial use is allowed if you keep the notices.

What the license types mean · not legal advice.

Specs

TaskMulti-objective sequence design on a fixed or flexible backbone; any sequence/structure score can be an objective
Modalities—
MSA required—
Open weights—
GPU need—
GitHub stars28
InterfacePython CLI (python run.py -c config.yml --models seq_model …)

Compare with…

Pick 1–3 other Protein structure & design tools to compare side by side with dEVA.

Select 1–3 tools

Related

AlphaFold 2

DeepMind's open-source protein structure prediction model (monomer and multimer).

Protein structure & designOpen sourcePermissive (commercial OK)#structure

Structure prediction of proteins, nucleic acids, ligands and their complexes; hosted web server plus open code, weights on request for non-commercial use.

Protein structure & designNo cost (research use)Mixed (code vs weights differ)#structure

Boltz

Open biomolecular interaction models (Boltz-1, Boltz-2) for complex structure and binding-affinity prediction.

Protein structure & designOpen sourcePermissive (commercial OK)#structure