Sequence features within human uORFs found to block accumulation of the proteins they encode

Ribosome-profiling studies have repeatedly shown that short upstream open reading frames (uORFs) embedded in the 5′ untranslated regions of human messenger RNAs are frequently translated — yet the proteins they would theoretically produce are almost never detected by proteomics. A preprint posted to bioRxiv on 17 August 2026 from Cold Spring Harbor Laboratory reports a systematic attempt to resolve this discrepancy.

Researchers quantified the intracellular accumulation of 111 evolutionarily conserved human uORF-encoded proteins (uORFps) using a unified reporter platform. They report that sequence-intrinsic barriers — properties embedded in the uORF sequence itself rather than in post-translational degradation pathways — define a class of translation-restricted uORFs whose products fail to accumulate to detectable levels even when ribosomal occupancy is confirmed.

The findings have implications for understanding the true coding potential of the human genome and for interpreting ribosome-profiling datasets, which have increasingly been used to infer the existence of novel small proteins. The work also raises questions about whether apparent uORF translation represents biologically meaningful protein production or a byproduct of ribosomal scanning mechanics. The study has not yet been peer-reviewed.