Nature Communications · 5 October 2026 · From issue 47
ProFormer: generalizable classification of single-cell and plasma proteomes using deep learning
A new computer program sorts cells and blood samples by disease status from raw protein data.
Karl K. Krull, Arlene Kühn et al.
A computer program can now sort cells and blood samples by reading raw data from protein measuring machines. It does this without scientists manually decoding the data first. The program, called ProFormer, beat 14 other computer methods at this task. It sorted single cells by type and stage, and told patients apart by disease status using blood samples.
Shows the program worked across several test datasets of cells and blood samples, not in a clinical trial.
Lab method only, no patient counts or accuracy numbers given yet.
Integrity screen: passed (3 checks) Checked 6 October 2026. Retraction record: none. DOI resolves at doi.org. Metadata record found (Nature Communications). Read the source
This screen checks process signals. It does not tell you the findings are true.