Systematic Benchmarking of DNA Sequence Encoding Strategies for Predicting Regulatory Effects of Non-Coding SNPs
Hui Jin, Yihang Bao, Wenhao Li, Chengyi Yang, Weidi Wang, Wenxiang Cai, Zhe Liu, Guan Ning Lin · 2026
Read Systematic Benchmarking of DNA Sequence Encoding Strategies for Predicting Regulatory Effects of Non-Coding SNPs by Hui Jin, Yihang Bao, Wenhao Li, Chengyi Yang, Weidi Wang, Wenxiang Cai, Zhe Liu, Guan Ning Lin (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/PMC13465526.txt.
Description
Hui Jin and seven co-authors benchmark six ways to encode DNA letters for models that predict what non-coding variants do. Letter changes outside genes steer regulation and disease, yet turning them into numbers shapes every model downstream. Six encoding styles, from plain categories to meaning-rich embeddings, faced three trait-linked tasks.
Nine learning models, from classic methods to deep nets, each ran the full encoding set under repeated, controlled trials. No single winner emerged; instead, encoding, model type, and data preparation bent results together. Readability, richness, and computing cost each pulled the choice a different way. The standardized frame gives genomics teams a fair test track for picking encodings before they commit to a pipeline.
Catalog record
| CREATOR | Hui Jin, Yihang Bao, Wenhao Li, Chengyi Yang, Weidi Wang, Wenxiang Cai, Zhe Liu, Guan Ning Lin |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 67223 |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |