Levitate Bio offers two fully managed protein modeling and design products: Levitate Bench, which is an easy to use browser based GUI tool, and Levitate Engine which is a powerful and flexible API system for building sophisticated protein modeling and design pipelines.

Levitate Bench

Cloud-run Rosetta & Artificial Intelligence protocols accessible via an easy-to-use web-based software

  • No Computational Biology experience needed
  • No Installation required
  • Runs in a web browser
  • Full-time professional support
  • Unlimited users
  • Flexible subscription tiers based on compute usage

Subscribe Now

A screenshot of the Levitate Bench application

Levitate Bench Tools

CategoryProtocolFunction
Protein Structure Prediction Single Chain Homology Modeling (HM)Monomer
HM for Variable Region (RosettaAb)Antibody Variable Region
HM for Variable Region (High Res)Antibody Variable Region
HM for Full Length Antibody (High Res)Antibody Full Length
AlphaFold2AI Folding, Monomer
AlphaFold2 MultimerAI Folding, Multimer
OpenFoldAI Folding, Monomer
ESMFoldAI Folding, Monomer, (fast)
Protein Structure Optimization RelaxFind More Favorable Conformations
PrepareConstrained Relax
Loop RemodelFind More Favorable Loops
MinimizeFind Closest Favorable Conformations
RepackOptimize Side Chains
Protein Design/Mutation Sampling Relax Design (region redesign with Relax)Suggest Sets of Mutations
Flex Design (region redesign with local flex)Suggest Sets of Mutations
Fixed Backbone Design (region redesign)Suggest Conservative Mutations
Single Point Mutation Scan (Cartesian DDG)Scan Point Mutations
Single Point Mutation Scan (ThermoMPNN)Scan Point Mutations
Sequence Analysis Immunogenicity ScanPredict MHCII Display

Levitate Engine

Cloud-Run Rosetta & Artificial Intelligence protocols accessible via APIs. 4 Different API packages are provided based on use case. The packages are focused on Rosetta modeling methods, AI/ML modeling methods, Analysis and liability prediction, and De Novo design.

Rosetta Package

The Rosetta Package has physical-based tools from the Rosetta Modeling Suite, including the most used tools with optional flags for flexible modeling of broad protein types and for diverse types of modeling goals.

This package requires a Rosetta license from the University of Washington

CategoryProtocolFunction
Protein Structure Prediction Single Chain Homology Modeling (HM)Monomer
HM for Variable Region (High Res)Antibody Variable Region
HM for Full Length Antibody (High Res)Antibody Full Length
HM with Disulfide ConstraintMonomer w/ Disulfide
HM Accessory Tools (Fragments, BLAST, PatentBLAST)Extra Tools
Protein Structure Optimization RelaxFind More Favorable Conformations
Loop RemodelFind More Favorable Loops
Protein Design/Mutation Sampling Relax Design (region redesign with Relax)Suggest Sets of Mutations
Single Point Mutation Scan (Cartesian DDG)Scan Point Mutations
Disulfidizer (Creates each favorable disulfide)Suggest Disulfide Positions
Epitope Redesign (Reduce MHC II display)Select Less Immunogenic Mutations
Structure Analysis RosettaHoles (Check core for gaps)Show Holes in Core
Interface Analysis (Energy, SASA, and more)Analyze Protein-Protein Interface

Artificial Intelligence Package

The Artificial Intelligence Packages provide the newest protocols for structure prediction and Rosetta community’s recent cutting-edge sequence design and structure generation tools.

FoldingSequenceStructureProtocolFunction
✔️AlphaFold2Monomer AI
✔️AlphaFold2 MultimerMonomer AI
✔️Openfold for Single ChainMonomer AI
✔️ESMFold for Single ChainMonomer AI Fast
✔️MolprobityID Structure Problems
✔️Single Point Mutation Scan (ThermoMPNN)Scan Point Mutations
✔️ProteinMPNN (sequence generation)Generate sequences from structures
✔️LigandMPNN (sequence generation)Generate sequences from structures including ligands
✔️RosettaFold Diffusion (RFDiffusion, structures)Generate de novo structures
✔️RosettaFold All Atom (RFAA, structures)Generate de novo structures including ligands
✔️Radius of Gyration (for RFDiffusion filtering)Protein Compactness

Analysis Package

The Analysis Package has machine learning-based predictions for developability

CategoryProtocolFunction
Structure/Sequence Analysis N-linked Glycosylation PredictionPredict Glycosylation
PTM prediction (deamidase, isomerization, oxidation, etc)Predict Degradation
Immunogenicity Scan (MHC II affinity)Predict Immunogenicity
Solubility/Aggregation PredictionPredict Solubility

De Novo Package

The De Novo Package has a full set of tools that facilitate end to end de novo protein design, featuring the Rosetta community’s AI protocols. The Lite version excludes tools that require a Rosetta license from the University of Washington.

CategoryProtocolFunctionLite Version
Protein Structure PredictionAI FoldingMonomer/Multimer Structure Prediction✔️
Structure Optimization RelaxFind More Favorable Conformations
Loop RemodelFind more favorable loops
De Novo Design ProteinMPNN (sequence generation)Generate Sequence from structure✔️
RosettaFold Diffusion (RFDiffusion)Generate de novo structures✔️
De Novo Design Molprobity (for AIFolding structure analysisIdentify Structure Problems✔️
Radius of Gyration (for RFDiffusion filtering)Generate de novo structures✔️
Interface Analysis (Energy, SASA, and more)Analyze Protein-Protein interfaces
Solubility/Aggregation PredictionPredict solubility✔️

De Novo Design toolkit demo video