Metapneumovirus genomes harbor a non-random layer of reverse ORFs

Abstract

Metapneumoviruses (MPVs) are negative-sense RNA viruses associated with recurrent respiratory disease outbreaks, yet their full coding potential remains incompletely characterized. We performed a genome-scale survey of reverse open reading frames (rORFs) located on the negative-sense genomic RNA across human metapneumovirus (HMPV) and related MPV genomes, benchmarking against codon-aware and composition-preserving null models. At the recurrent P-overlapping locus, we identified an unusually long rORF (244 aa) overlapping most of the phosphoprotein (P) gene; the HMPV reference genome contains a shorter variant at the same locus (186 aa), which we term P-AS1. Across the dataset, the upper tail of rORF lengths and their clustering into recurrent hotspots exceeded expectations under both controls, consistent with non-random organization. In sequence- and structure-based analyses of selected HMPV reference rORF candidates, several showed features consistent with small membrane-associated proteins, including predicted transmembrane helices. AlphaFold-based modeling of rORF90 suggested a pore-like oligomeric assembly consistent with a tentative viroporin-like hypothesis, although the model remains hypothesis-generating and does not demonstrate expression or function. We assessed translational permissiveness using initiation-site prediction and a Kozak-like start-context screen, supporting prioritization of selected candidates without implying in vivo translation. Together, these analyses support a non-random layer of rORFs in MPV genomes and provide candidates and resources for targeted experimental validation.