BRIDGE-KO: auditable transfer learning with bridged knockoffs improves cross-cohort microbiome prediction of immunotherapy response
Jie Sun, Zhenjiang Fan, Rebecca Caldino Bohn, Soyeon Kim, Hyun Jung Park · 2026
Read BRIDGE-KO: auditable transfer learning with bridged knockoffs improves cross-cohort microbiome prediction of immunotherapy response by Jie Sun, Zhenjiang Fan, Rebecca Caldino Bohn, Soyeon Kim, Hyun Jung Park (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-13660471.txt.
Description
Jie Sun and co-authors present a transfer learning method for gut microbe data in a 2026 paper. Models trained on one study often fail on another because data quirks do not transfer. Their method builds fake shadow features that mirror a source study's structure. Those shadows flag which signals are real and which are study specific noise.
The test task was forecasting response to immune therapy in skin cancer from gut profiles. A huge microbe data set served as the source for the small cancer cohorts. Plain transfer scored near chance, while careful matching lifted scores clearly. The method also audits how microbe patterns shift in responders, the team reports.
Catalog record
| CREATOR | Jie Sun, Zhenjiang Fan, Rebecca Caldino Bohn, Soyeon Kim, Hyun Jung Park |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 76 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |