Folding the unfoldable 2: using AlphaFold and ESMFold to explore spurious proteins
Ailsa K Orr, Alex Bateman, Dmitrij Frishman · 2026
Read Folding the unfoldable 2: using AlphaFold and ESMFold to explore spurious proteins by Ailsa K Orr, Alex Bateman, Dmitrij Frishman (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/PMC13294457.txt.
Description
Ailsa K Orr, Alex Bateman, and Dmitrij Frishman test whether folding software can spot fake proteins in a 2026 study. Spurious entries arise from gene prediction errors and should not fold, yet earlier AlphaFold work gave short fakes high confidence scores. Three predictors, ESMFold, AlphaFold2, and AlphaFold3, were run on fake and real sets split by length.
All three still scored short fakes suspiciously high, but sorting improved past 100 amino acids. Score mismatches flagged two likely fake shadow entries in a trusted database and one wrongly flagged real entry. A Gaussian process model scanning the big AlphaFold database found more suspects, best used with other checks.
Catalog record
| CREATOR | Ailsa K Orr, Alex Bateman, Dmitrij Frishman |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 43248 |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |