Studying lncRNA Sequence–Function Relationships Using Minimally Supervised Model
Long non-coding RNAs (lncRNAs) constitute a major regulatory component of the human transcriptome, yet their mechanisms of action and functional organization remain poorly understood. Enabled by advances in RNA sequencing, genomic resources, and molecular interaction assays, this study integrates computational and experimental approaches to characterize lncRNA regulatory functions and subcellular localization. We applied a Minimally Supervised Model bioinformatic framework to identify functional domains within lncRNA sequences and combined it with experimental validation, revealing stimulus-dependent localization dynamics. In parallel, optimization of a biotinylated RNA pulldown assay uncovered novel lncRNA–protein interactions, including a previously uncharacterized regulatory mechanism involving SNHG14 and the identification of LINC01001 as a chromatin-associated lncRNA. Together, these findings highlight the power of combining language-inspired computational models with molecular biology techniques to uncover hidden regulatory features of the non-coding genome
