Validation

Testing every prediction in the laboratory, so the platform stays answerable to real measurements.

A high-key laboratory macro of sample vials and a pipette tip catching the light, rendered in soft violet and cyan on a pale background. Illustrative generated image.

Introduction

Computation only counts once it survives the bench.

Wet-laboratory science in Bellinzona keeps every computational prediction answerable to experiment.

How it works

Bellinzona laboratories

Predictions about disordered proteins are only useful if they hold up experimentally. Peptone runs an 8,000 sq ft Biosafety Level 2 laboratory in Bellinzona, Switzerland, so expression, purification, biophysical characterization, and functional assays sit alongside the computational platform rather than downstream of it.

Measured against reality

The laboratory produces the proteins the platform reasons about, measures them with the biophysical methods that report on disorder, and tests whether designed binders engage their intended states. When the data disagree with a model, the model changes. This is what keeps the platform calibrated to reality.

Data that compounds

Because measurement and computation share the same targets and the same loop, discovery decisions are grounded in evidence at every step. Each program also returns proprietary ensemble and assay data that improves the methods applied to the next one.

A translational loop linking expression, characterization, and assays back to the computational platform.

At a glance

Protein expression and purification tuned for disordered and dynamic targets
Biophysical characterization that reports directly on conformational behaviour
Functional assays that test whether designed binders engage their target states
A feedback path where experimental data recalibrates the computational models
Proprietary ensemble and assay data compounding across every program