All posts
Announcement

Aura, the first AI scientist for biology

What if we could make 100x more discoveries in human health, solve unresolved diseases, and ultimately reach biological freedom?

JCJulio Castellanos
3 minutes read
Aura workflow

Aura is a solution of AI agents that seamlessly combines protein folding, structure generation, property prediction, molecular dynamics and more into a single, AI-accelerated platform. Researchers can move from raw sequences to validated structures and functional insights—all within one intuitive ecosystem.


What Can Aura Do?

Aura empowers scientists across disciplines with flexible, composable tools. For instance, a typical workflow might include:

  • Generating novel structures with the De Novo Generation module
  • Predicting molecular properties like stability and solubility
  • Refining and validating with folding predictions and scoring metrics like PAE and pLDDT

This kind of modularity allows you to iterate faster and explore more hypotheses without being bottlenecked by traditional computational hurdles.

“To become proficient with Aura, it’s best to learn each feature both manually and through automation. This maximizes adoption and ensures smoother, more efficient outcomes.”

Aura workflow

Aura is built for continuous iteration between prediction, generation, and validation—whether you're studying proteins, peptides, or full biological systems.


Real Research, Real Impact

We recently spoke with Luisa Cruz, a plant pathogen researcher from North Carolina State University, about workflows in agricultural science that could benefit from AI acceleration.

“One of the most difficult things is obtaining target molecules and starting to predict interactions with novel mechanisms,” she shared.

Luisa studies Xylella fastidiosa, a destructive plant pathogen that affects vital crops like olives, grapes, and citrus. This bacterium blocks water flow in plants, causing leaf scorch and death—posing a serious threat to global agriculture.

Why is this important?

Understanding and disrupting peptide-protein interactions between host plants and the pathogen is critical, but the process of identifying, modeling, and validating these interactions is both complex and resource-intensive.

With Aura, researchers like Luisa can streamline this pipeline:

  • Instant literature review via integrated PubMed search
  • Retrieve structures from PDB and UniProt with a single prompt
  • Use the TX-Gemma mode (fine-tuned from DeepMind’s open-sourced model) to reason about molecular interactions and potential inhibitors
  • Run simulations using Boltz-1 models and molecular dynamics engines for in silico validation

This AI-accelerated loop—from molecule discovery to computational validation—helps scientists rapidly evaluate new hypotheses and unlock breakthrough insights that would otherwise take months.


Choose Your Aura: Agent Modes

Aura adapts to your research goals with specialized agents:

Aura-X-assistant

Aura workflow

For rapid structure analysis and function prediction:

  • Protein folding prediction
  • PAE and pLDDT plots
  • De novo design of protein sequences
  • Function prediction from structure
Aura workflow

Perfect for iterative design, validation, and hypothesis testing.

Models available:

Folding

  • Alphafold 2
  • Colabfold
  • ESM
  • Bolz-1

Generation and property

  • ProTrek
  • ESM2
  • Denovo-Pinal
  • Evolla + Llama3 8b
  • SaProt
  • BiomedNLP

Deep Properties

  • SignalP 6.0
  • AMP-Net
  • DeepLoc

Aura-Z-thinking

Aura workflow

For advanced modeling and deeper exploration:

  • Molecular dynamics simulations
  • PDB search for small molecules and proteins
  • Structural and interaction analysis workflows
Aura workflow

Ideal for high-complexity environments where fine-grained control and simulation accuracy matter.

Tools available:

Molecular Dynamics

  • OpenMM
  • PDB search
  • UniProt

Aura-TX-gemma

Aura workflow

Tailored for therapeutic discovery and clinical applications:

  • Drug interaction prediction
  • Clinical toxicity assessments
  • Literature review via PubMed integration
  • Translational science with biomedical reasoning
Aura workflow

Designed for drug discovery teams, biotech startups, and clinical researchers.

Tools available:

  • Wikipedia, Web and Pubmed Search

  • BLASTp

  • Antibody affinity

  • Molecule Search

  • SMILES to Therapy

  • SMILES Description

  • Drug-target interaction

  • Antibody affinity

  • Protein-protein interaction

  • Toxicity

  • Pharmacokinetics

  • Developability

  • CRISPR Repair


Why Aura?

Aura isn't just a toolkit—it's an AI lab partner. It brings speed, scalability, and scientific reasoning into your workflows, helping you explore the unknown, validate new ideas, and move from concept to insight faster than ever before.

Whether you're designing next-gen therapeutics, studying agricultural pathogens, or unraveling the secrets of protein folding—Aura is here to accelerate your science.

Ready to get started? Join our early access program and bring your research to the next level.