Computational Protein Engineering

Better enzymes, designed with models and data

We combine structural modeling, molecular dynamics and machine learning to help R&D teams find stronger enzyme variants faster, then hand over every model, pipeline and method.

01 02 03
Fig. 01Modeled landscape, three ranked candidates

What we do

Services

From the first feasibility review through to full handover.

01

Feasibility and Research Roadmaps

We review public and proprietary databases, assess what is technically achievable, and deliver a written roadmap before major investment begins.

Details on Feasibility and Research Roadmaps
02

Structural Modeling and Docking

We build structural models of wild-type and variant enzymes and analyze active sites, substrate binding and docking to guide design choices.

Details on Structural Modeling and Docking
04

Data and ML Pipelines

We build the data-management and machine-learning pipeline that turns experimental results into predictive models you can benchmark and trust.

Details on Data and ML Pipelines
05

Active-Learning Design Loops

Each round of lab results retrains the models, which then propose the next set of variants. Every cycle makes the search sharper.

Details on Active-Learning Design Loops
06

Assay Data Analysis

We analyze and interpret assay data from your laboratory, linking measured performance back to structure and model predictions.

Details on Assay Data Analysis

How we deliver

Milestone by milestone, with no surprises.

Each engagement is broken into defined deliverables. You review and accept each one before the next phase begins.

  1. 01 Month 0

    Kickoff and scope alignment

    We agree on goals, data access, success criteria and the review schedule for the engagement.

  2. 02 Month 2

    Feasibility report and roadmap

    Target identification, database review, written feasibility assessment and a phased research roadmap.

  3. 03 Month 4

    Structural models delivered

    Wild-type and variant structures with active-site, substrate-binding and docking analysis.

  4. 04 Month 6

    First ranked variant library

    Molecular dynamics results and a ranked library of variants ready for laboratory testing.

  5. 05 Month 7

    ML pipeline and benchmarks

    Data-management and machine-learning pipeline built, predictive models benchmarked, active-learning loop designed.

  6. 06 Month 8

    First-round assay analysis

    We analyze and interpret the first round of assay data returned from your lab.

  7. 07 Month 9

    Models retrained on results

    Predictive models are retrained on real assay data to improve the next round of designs.

  8. 08 Month 10

    Second-round optimized variants

    A new set of optimized variants, informed by the retrained models, is delivered for testing.

  9. 09 Month 11

    Final report and documentation

    A final technical report plus documentation of all code, models and methods.

  10. 10 Month 12

    Handover and team training

    Code and models are transferred to your team, with hands-on training so they can continue independently.

Principles

How we work with you

Milestones you can verify

Every phase ends with a concrete deliverable, such as a report, a model or a variant library, so progress is easy to verify.

Models that learn from the lab

Computation and experiment work as one loop. Your assay results feed back into the models, and each round gets more precise.

You own everything

Code, models, data pipelines and methods are documented and handed over. No black boxes and no lock-in.

Security

Secure by default

Your data and IP are protected at every step of an engagement.

Security practices

Confidential by default

Your sequences, structures and assay data stay confidential and are used only for your project.

Controlled data access

Access to project data is limited to the people working on it and reviewed throughout the engagement.

Your infrastructure, your choice

We can work in your cloud or on-premise environment, so sensitive data never has to leave your control.

Traceable methods

Every model, dataset version and analysis step is documented, so results can be audited and reproduced.

FAQ

Common questions

What is computational enzyme engineering?

It uses structural modeling, molecular dynamics and machine learning to predict which enzyme variants are most likely to perform better. Your lab then tests a short, ranked list instead of screening thousands of random candidates.

Do you run the laboratory experiments?

No. We handle the computational side: modeling, simulation, variant ranking and data analysis. Your team or contract lab runs the assays, and we analyze the results and design the next round.

How long does a typical engagement take?

A full engagement usually runs about 12 months, from feasibility through two design rounds to final handover. Shorter projects, such as a feasibility study or a single modeling phase, are also available.

What is an active-learning loop?

It is a cycle in which lab results retrain our predictive models, and the updated models choose the next variants to test. Each round focuses your experiments on the most promising designs.

Who owns the code and models at the end?

You do. We deliver full documentation of all code, models and methods, transfer them to your team, and provide training so your scientists can keep using and extending them.

What do you need from us to get started?

Your target enzyme or goal, any existing sequence, structure or assay data, and a point of contact in your lab. We use these to complete a feasibility review before committing to a full roadmap.

Let’s talk

Have a project in mind?

Call or email us and tell us what you need. We reply within one business day.