Back

Can AlphaFold3 predict unseen amyloid polymorphs?

Price, R. M.; Maerivoet, A.; Mesdaghi, S.; LaCourse, B.; Madine, J.; Rigden, D. J.

2025-12-09 bioinformatics
10.64898/2025.12.05.692529 bioRxiv
Show abstract

Amyloid fibrils are highly ordered, aggregation-prone protein assemblies implicated in numerous neurodegenerative diseases. Their formation arises from protein misfolding, producing structurally distinct polymorphs that have historically been difficult to predict computationally. While AlphaFold2 excels at predicting monomeric globular proteins, it struggles with protein complexes. Here, we show that AlphaFold3 provides a new opportunity to model amyloid structure and capture polymorph heterogeneity. Using TM-score based clustering of known -synuclein fibril structures, we establish a polymorph nomenclature and demonstrate that AlphaFold3 can reproduce known fibril architectures and predict alternative conformations. These results highlight AlphaFold3s potential for studying amyloid proteins, enabling systematic investigation of amyloid polymorphism.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.