Characterizing Highly Conserved Fragments in 3'UTRs via Computational and Transfer Learning Approaches
Eric S. Ho, Ash Baeck-Hubloux, Nathan Dinh, Ava Severino, Ciara Troy
The paper identifies highly conserved fragments in 3'UTRs across mammals using computational methods and a transformer model, finding they are A/T-rich and enriched in neuronal tissues with roles in neurodevelopment and RNA processing.
bioRxiv · Jan 20, 2026 · PreprintGenomics