Exceptional Conservation of SARS-CoV-2 ORF10 Across 9.3 Million Genomes Suggests Functional Constraint

ORF10 is a small open reading frame at the 3' end of the SARS-CoV-2 genome, often dismissed as non-functional due to its lack of confirmed protein expression and frequent truncations in some lineages. However, large-scale genomic evidence for its evolutionary constraint remains limited and until now, no study has systematically mapped co-mutation patterns across millions of sequences. We analyzed 9,356,279 high-quality SARS-CoV-2 genomes to assess the integrity of the ORF10 coding region (positions 29558–29674). Using streaming bioinformatics, we identified co-occurring mutations without full decompression of the data. To evaluate selective pressure, we calculated the pN/pS ratio the proportion of variable non-synonymous sites relative to synonymous sites a more robust metric than dN/dS for low-diversity regions like ORF10. Of the 9.36 million sequences analyzed, 9,105,582 (97.3%) preserved a full-length ORF10 region, indicating exceptional conservation. We observed structured co-mutation patterns, including a dual insertion (29611:->T, 29612:->G) present in over 244,000 sequences and detected in 9 countries. The pN/pS ratio was 0.32, well below 1, suggesting strong purifying selection. Traditional dN/dS could not be reliably estimated due to low synonymous variation and ambiguous base calls (e.g., 'N'), but the near-universal preservation of the ORF10 reading frame provides compelling evidence of evolutionary constraint. The scale and consistency of ORF10 conservation across global sequences challenge the assumption that it is entirely dispensable. While we cannot confirm protein expression from genomic data alone, the strong signal of purifying selection reflected in both pN/pS and the rarity of disruptive mutations suggests that ORF10, or its overlapping genomic elements, may play a functional role. Rather than dismissing it as junk, we should keep investigating.