A review of new science, in plain language

Nature Communications  ·  1 September 2026  ·  From issue 30

A machine learning method for calculating highly localized protein stabilities

A computer program calculates how easily each tiny part of a protein opens, even in huge proteins.

Chenlin Lu, Kyle C. Weber et al.

Proteins are not stiff. Small parts of them open and close like tiny hinges. How easily each part opens changes how the protein works. A new computer program called PFNet reads results from a common lab test and quickly figures out how easily each small piece of a large protein opens up. This gives scientists a fast, detailed map of a protein's floppy and stable spots.

What it could change A tool that turns a common lab test into a detailed map of protein flexibility.

Shows the method can work on proteins of any size, not specific results for any protein.

Method paper only, no specific protein or disease results yet.

Integrity screen: passed (3 checks) Checked 2 September 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.

Get it free, every two days

Five papers free every issue, about 15 emails a month. No spam. Leave in one click.

Also in issue 30 15 papers
More from Nature Communications