AntigenLM: Structure-Aware DNA Language Modeling for Influenza
Yue Pei, Xuebin Chi, Yu Kang · 2026
Read AntigenLM: Structure-Aware DNA Language Modeling for Influenza by Yue Pei, Xuebin Chi, Yu Kang (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/arxiv-2602-09067.txt.
Description
DNA base models often trail niche tools at flu forecasting, for reasons that were unclear. Yue Pei and two co-authors present AntigenLM, a generative DNA language model trained on flu genomes with functional units kept intact and aligned. That structure-aware training lets the model absorb evolutionary constraints and transfer across tasks.
Fine-tuned on time-ordered hemagglutinin and neuraminidase sequences, the model predicts coming antigenic variants across areas and subtypes, including unseen ones, beating phylogenetic and evolution-based rivals. It also nears perfect subtype classification. Ablations show that fragmenting or shuffling the genome wrecks performance, proving unit integrity matters.
Catalog record
| CREATOR | Yue Pei, Xuebin Chi, Yu Kang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 43 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-09067