DNA augmentations for semi-supervised learning in genomic sequence classification
Hyunwoo Yoo, Gail L Rosen · 2026
Read DNA augmentations for semi-supervised learning in genomic sequence classification by Hyunwoo Yoo, Gail L Rosen (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/pmc-13678202.txt.
Description
Hyunwoo Yoo and Gail Rosen test ways to stretch scarce labels for DNA classification in a 2026 application note. Gene data is plentiful, but labels cost expert time, so semi-supervised learning trains on a few labeled cases plus many unlabeled ones. The catch is how to perturb DNA copies during training without breaking the biology that the labels rest on.
The pair tried twenty-five mixes of weak and strong changes, from flipping a strand to swapping codons, masking spots, and inserting or deleting letters. Gentle, shape-keeping changes worked best, while heavy masking or mutation hurt. Their winning recipe pairs short masked patches with a mild insertion-deletion rate. They share code and default settings for others to reuse.
Catalog record
| CREATOR | Hyunwoo Yoo, Gail L Rosen |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 23 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |