Functional landscape of non-canonical open reading frames in coordinating cell fate

The human genome harbors thousands of unannotated short open reading frames (sORFs) with the potential to encode microproteins, yet their physiological roles remain largely unexplored. Here, we developed sORF-seq, a functional screen that identified hundreds of sORF-encoded microproteins regulating cellular differentiation. Among these, we discovered lncPRESS1 as a critical regulator of cell fate, and remarkably, it acts as a bifunctional RNA. In the nucleus, lncPRESS1 functions as a long non-coding RNA (lncRNA), guiding the genomic distribution of SWI/SNF and orchestrating developmental gene expression programs. In the cytoplasm, lncPRESS1 acts as an mRNA, translated into a microprotein that directs lineage commitment through Sonic hedgehog (SHH) signaling pathways and interactions with the primary cilium. This dual functionality allows lncPRESS1 to coordinate nuclear and cytoplasmic regulatory networks, shaping early embryogenesis and human brain development. Our findings unveil an unexpected paradigm of non-canonical ORFs in choreographing complex pathways, expanding our understanding of the functional genome beyond traditional coding genes.