A University of Warsaw team has reported an experimentally determined protein structure containing an 8_3 knot with eight crossings, matching an earlier AlphaFold prediction. The significance is not that AI settled the question. It generated an implausible lead that researchers then tested in nature and in the laboratory.
According to Nauka w Polsce, the structure had appeared incompatible with previous assumptions about how proteins fold. The team, led by Professor Joanna Sulkowska at the University of Warsaw’s Centre of New Technologies, identified candidate proteins in anaerobic bacteria associated with the human gut. Researchers grew the relevant protein and determined its structure experimentally. After several years of work, the knot matched the model’s prediction. They named the family EIGHT fold.
That sequence matters. A structure prediction can direct attention to a possibility that established theory had discounted; experimental structure determination is what turns that possibility into a biological finding. The result is therefore a case for using AI as a disciplined source of testable hypotheses, not for treating a model output as a conclusion.
The report describes 8_3 as the most complex protein knot observed to date. Unlike simpler knots thought possible through one threading event, this configuration requires the protein end to pass through twice. It challenges a prior boundary in protein-folding theory, but it does not yet supply a replacement mechanism.
The central scientific question is still how a cell makes the knot. The supplied report does not specify the new family’s biological function, detailed experimental methods, or a medical or industrial use. The team says reproducing the conditions for such folding would matter for efforts to synthesize arbitrary proteins. For now, the clearer achievement is narrower: an AI prediction that seemed improbable has become a precise experimental problem for protein science.
Sources
- Nauka w Polsce · Fundacja PAP — Źródło: Nauka w Polsce (Fundacja PAP) (https://naukawpolsce.pl/) · Source date: 2026-11-12
Independent editorial content. WUA is not affiliated with any political party. 2026-11-12