Cradle

¿Qué es Cradle?

Cradle is an AI protein engineering platform for scientists working on antibodies, enzymes, peptides, and vaccines. It connects candidate design, experimental testing, and optimization in a workflow that learns from a team's assay results. Researchers can start from a target or an existing sequence and refine candidates toward desired efficacy, developability, and manufacturability profiles. Scientists choose which candidates to test and upload the resulting assay data to guide subsequent design rounds.

¿Qué puede hacer Cradle?

Project-specific models

Train models on your assay data to learn the fitness landscape of an individual protein. Results from Cradle-generated candidates and your own designs inform subsequent rounds.

Multi-property optimization

Define a primary objective and constraints in a Target Candidate Profile. Cradle evaluates trade-offs across specified properties and estimates the probability of meeting each objective.

Candidate plate forecasting

Score proposed plates, including designs from other methods, before laboratory testing. Review expected hit rates, score distributions, and remaining room for improvement to decide whether to run, redesign, or stop.

Sequence constraints

Apply fixed regions, liability motifs, length limits, and other constraints during candidate generation so returned sequences fit the project's construction requirements.

Guided rounds and analytics

Follow guided setup from raw data to candidates for testing. Reports visualize model outputs and help scientists inspect variants, compare mutations, and choose subsequent experiments.

Web application and API

Bench scientists can design and optimize candidates through a no-code interface. Computational teams can access the same capabilities programmatically and incorporate them into existing workflows.

Casos de uso

“Therapeutic candidate optimization”

Biopharma teams can refine antibodies or peptides against potency and developability requirements, including projects where earlier screening or optimization has stalled.

“Enzyme engineering”

Industrial biotechnology researchers can seek enzymes with stronger catalytic activity, expression, and thermostability for demanding operating conditions.

“Hit confirmation after screening”

Teams with noisy screening pools, such as NGS enrichment counts from selection campaigns, can down-select sequences into a diverse plate for hit confirmation.

“Academic protein research”

Researchers conducting non-commercial studies can apply to the limited-availability academic program, which includes publication rights and expert help getting started.

Preguntas frecuentes

Do I need assay data to start?

No. Cradle can generate a first round from a single protein sequence without assay data. Experimental results can then be uploaded to retrain the model before the next round.

How is Cradle licensed?

Cradle offers annual software licenses without royalties or milestone payments. Standard licensing is based on active projects; Enterprise licensing supports deployment across programs and divisions. Commercial licensing inquiries go through its team.

Is academic access free?

The license fee is waived for non-commercial research through Cradle's limited-availability academic program. Researchers must apply and can receive API and/or web interface access, publication support, and expert assistance.

How is customer data protected?

Cradle states that its platform is ISO 27001 and SOC 2 Type II certified. It runs on Google Cloud with data isolation, continuous monitoring, and access restricted to people authorized by the customer.