What if you could predict therapy response using patient-representative avatars before you test in the wet lab?

1. Train avatars ?

Our models learn parameters of cell signaling, shared from cell lines to patients, enabling generalization to new experiments

INPUT

  • Genetic or pharmacological perturbation response data
  • Drug target profiles
  • Genomic and transcriptomic profiles of additional samples

RESULT

  • Trained avatars for project
Approach

2. Simulate experiments

A large library of in silico avatars (cell lines, PDOs, PDXs, and patient samples) can be perturbed genetically and pharmacologically to generate novel predictions at computational scale

SETUP

  • In silico experiment design based on project scope

RESULT

  • Phenotypic and transcriptomic readouts processed for further analyses
Approach

3. Translate results

Simulated experiments are quick to design and easy to interpret. Our proprietary translational suite helps you sift through millions of simulations and generate mechanistically explained hypotheses

INPUT

  • Processed data from in silico experiments

RESULT

  • Hypotheses and mechanistic insights reinforced by biological domain experts
  • Experimental workplan
Approach

4. Validate hypotheses ?

Receive a decision-ready, data package, that integrates into you existing workflows. You only need to perform the experiments with the highest PoS, designed with the help of simulations

SETUP

  • In silico-guided hypotheses with biological and mechanistic insights, reinforced through scientific literature by domain experts
  • Experimental workplan

RESULT

  • Wet-lab experimental data package
  • Guided validation from in vitro to in vivo
Approach

Generate training data ?

Simulation guided proprietary data generation gains the same model performance from 5X fewer data points than unguided, high-throughput data generation

OUTPUT

• functional genomics
• drug modifier screens
• direct assays (e.g.: apoptosis)
• post-treatment transcriptomics

Approach

Turbine’s design principles enable simulating biology & getting to the right experiment


Our model learns fundamental rules of biology to predict events it hasn't seen
We train parameters that govern signaling and are shared from cell lines to patients. This way our model generalizes to new experiments.

Predicting cytotoxic
drug response
(pIC50 correlation)

Predicting post-KO
gene expression
(Top 20 DEG* correlation)

Turbine

Turbine outperforms best in class models in predicting cell behavior on the phenotypic and molecular levels in experimental conditions the model has never seen

*DEG: Differentially Expressed Gene
**NeST: Zhao et al. Cancer Discov 2024;14:508–23
***scGPT: Cui et al. Nat Meth 2024;21:1470–1480
Simulated experiments are quick to design and easy to interpret
Our proprietary translational suite helps you sift through millions of simulations. Resulting insights are actionable, validated at Turbine.

2x

faster from asking
questions to validation

Turbine

From 9 million simulations to actionable recommendations in 4 weeks. Integrating Turbine, partner delivered key data package 2X faster

We harmonize every bit of data and generate it where it really matters
Diverse data sources are boosted by an integrated lab-in-the loop system. Continuous iterations boost predictivity.

80%

less $ spent on data

Turbine

Simulation guided proprietary data generation gains the same model performance from 5x fewer data points than high-throughput data generation

To shift the paradigm in R&D, we need AI to understand how a novel therapy or target will function where it matters the most – in patients.

It must be nimble, able to be implemented at any point in the R&D process and capable of quickly identifying the right real-life experiments to progress with.

Human biology is governed by the concert of our cells. We believe that at a deeper level, all cellular behavior is driven by proteins and their interactions.

Our platform learns this universal language from scalable experiments and applies its knowledge to predict the response of models it has never seen before, ranging from engineered cell lines to patient samples.

Our breakthrough technology generates actionable results in half the time required to train other AI models.



It also helps explain the biological mechanisms at play to reveal hypotheses with the greatest potential to benefit patients and to identify the exact experiments needed for validation.

Turbine offers a new path forward with Simulated Cell™. A platform to simulate biological experiments before you run them – in a cell, animal or even in patients.

    Close