Capturing short-range and long-range dependencies of nucleotides for identifying RNA N6-methyladenosine modification sites

Guodong Li, Bowei Zhao, Xiao-Rui Su, Yue Yang, Zhi Zeng, Pengwei Hu, Lun Hu

Abstract: Highlights•We incorporate alias sampling strategies to comprehensively capture both short-range and long-range dependency information among nucleotides. This enhancement aims to bolster the accurate identification of RNA m6A modification sites.•We formulate a self-correlation map for each RNA sequence to retain both long-range and short-range dependencies among nucleotides. Leveraging these maps, m6ASLD assesses the significance of Local representations and refines the global representations of sequences.•We design the joint training of local and global representations of sequences, achieving the best performance on all benchmark datasets.

External IDs:dblp:journals/cbm/LiZSYZHH25