Transfer learning for T-cell response prediction
Josua Stadelmaier, Brandon Malone, Ralf Eggeling · 2026
Read Transfer learning for T-cell response prediction by Josua Stadelmaier, Brandon Malone, Ralf Eggeling (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/PMC13251289.txt.
Description
Josua Stadelmaier, Brandon Malone, and Ralf Eggeling study how to predict T-cell responses in this 2026 paper. Forecasting whether a small protein fragment will trigger T cells could guide personal cancer vaccines, but training data are scarce and mixed from many sources.
The team shows a trap: models can look strong by learning which source a fragment came from instead of what makes T cells react. They propose a fairer test scheme that respects those sources. A per-source fine-tuning approach then worked well across many sources, and their final model matched the best existing tools for human fragments. Careful testing design, they argue, is half the battle.
Catalog record
| CREATOR | Josua Stadelmaier, Brandon Malone, Ralf Eggeling |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 46414 |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |