About

Both halves of the problem.

This sits between two fields that rarely share a person. The measurement is hard, the modelling is hard, and the failures come from the seam between them.

01 — Stage

Early, and honest about it.

There is no product to sell, no benchmark published, and no claim here that the model works. What exists is a thesis, the published work below, and a plan for the measurements that come first.

We would rather this page understate where things stand than have to walk something back later. When there is a result, it will be on the site, with the baseline it was scored against.

02 — Publications

The measurements this approach is built on are already in print.

Three perturbation classes — genetic, chemical dose, enzymatic — each read across glycosylation and phosphorylation on the same material. Raw data is deposited and public under the accessions listed.

03 — Get in touch

Worth an email if any of these describe you.

/ 01

You build cell models

Virtual-cell, perturbation-response, or foundation models for biology, and you have wondered what the protein layer would add.

/ 02

You run the instruments

Proteomics at scale, particularly modification-enriched workflows and designed perturbation series.

/ 03

You think this is wrong

Genuinely welcome. The fastest way to improve the thesis is to hear the strongest argument against it.

contact@cellucent.com