# Researchers unveil TangleDiff to design entangled protein hydrogels

Digest AI · Research · published 2026-09-21T00:00:00Z

Canonical: https://digestai.news/story/researchers-unveil-tanglediff-to-design-entangled-protein-hydrogels

## Summary

A team of authors introduced TangleDiff, a deep‑learning framework that creates homodimeric entangled proteins with programmable features.  In silico, TangleDiff generates diverse foldable sequences with a success rate exceeding 70%, far above the roughly 1% reported for current models.  When conditioned on inter‑chain binding energy, about 70% of the successful designs fall within the targeted energy ranges.

The researchers experimentally tested nine designed homodimers covering various binding energies.  Seven of these homodimers formed hydrogels, and the observed stress‑relaxation dynamics aligned with the specified binding‑energy values.  The work demonstrates a general strategy for entangled protein design, potentially expanding biomaterial innovation for 3D stem‑cell and organoid culture.

## Key points

- TangleDiff achieves in‑silico success exceeding 70%, versus ~1% for existing models.
- Conditioning on binding energy yields approximately 70% of designs matching target ranges.
- Experimental validation of nine homodimers produced seven hydrogels with expected stress‑relaxation behavior.

## Why it matters

A high‑success, programmable protein‑design tool could accelerate creation of custom biomaterials for tissue engineering and regenerative medicine.

## Sources

1. [Generating protein hydrogels with customizable stress relaxation behavior via deep learning-driven entanglement design](https://nature.com/articles/s41467-026-77607-9) (Nature Machine Learning, 2026-09-21, primary source)

## Cite

Digest AI, "Researchers unveil TangleDiff to design entangled protein hydrogels", 21 September 2026, https://digestai.news/story/researchers-unveil-tanglediff-to-design-entangled-protein-hydrogels

---

Written by Digest AI's editorial model from the linked sources; the sources are the record. Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse
JSON: https://digestai.news/story/researchers-unveil-tanglediff-to-design-entangled-protein-hydrogels.json