Back

Ankh-score produces better sequence alignments than AlphaFold3

Malec, J.; Rusen, K.; Golding, G. B.; Ilie, L.

2025-09-07 bioinformatics
10.1101/2025.09.03.674029 bioRxiv
Show abstract

Protein sequence alignment is one of the most fundamental procedures in bioinformatics. Due to its many downstream applications, improvements to this procedure are of great importance. We consider two revolutionary concepts that emerged recently as candidates for improving the state-of-the-art alignment methods: AlphaFold and protein embeddings. Alignment improvements can come from structural alignment of AlphaFold-predicted structures or scoring based on similarity of protein embedding vectors, resp. Thorough comparison on many domains from BAliBASE and CDD demonstrates that the Ankh-score method produces much better sequence alignments than structural alignments using US-align of AlphaFold3-predicted structures. Both are better than the traditional method using BLOSUM matrices. This suggests that Ankh embedding vectors may possess certain information that is not available in the AlphaFold3-predicted structures.

Published in Proteins: Structure, Function, and Bioinformatics (predicted rank #6) · training set

Matching journals

The top 5 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.