EpiTEAmDNA: Sequence feature representation via transfer learning and ensemble learning for identifying multiple DNA epigenetic modification types across species

Fei Li, Shuai Liu, Kewei Li, Yaqi Zhang, Meiyu Duan, Zhaomin Yao, Gancheng Zhu, Yutong Guo, Ying Wang, Lan Huang, Fengfeng Zhou

Abstract: Highlights•The generic predictor EpiTEAmDNA covers 3 DNA methylation types and 15 species.•EpiTEAmDNA uses the same architecture and hyperparameters for all the 29 datasets.•EpiTEAmDNA outperforms the SOTA generic predictors and type-specific predictors.•Transfer learning and ensemble learning are integrated for better performance.•EpiTEAmDNA performs well on both small and large datasets.