Novel multigene molecular characterization of avian reovirus strains and associated embryonic pathogenicity.

Avian reovirus (ARV) infection causes a variety of health problems in poultry, including tenosynovitis/arthritis, hepatitis, malabsorption, etc., leading to huge economic losses. Therefore, regular monitoring of ARV strains is crucial for detecting new variants to mitigate clinical disease. Traditional classification approaches have focused primarily on a single gene, σC, which encodes the viral spike protein that elicits neutralizing antibodies. However, additional capsid proteins could play a role by inducing a host antiviral immune response. In this study, the λC, μB, σC, and σB proteins encoded by the L3, M2, S1, and S3 genomic segments, respectively, were analyzed to classify clinical ARV strains with respect to embryonic pathogenicity and genotype-phenotype relationships. Embryo inoculation revealed strain-dependent pathogenicity ranging from the absence of tendon lesions (strain 23-272) to multifocal ulceration with fibrinoheterophilic crusting (22-806) and multifocal necrotizing hepatitis (23-087). Pairwise log-rank and Cox regression tests demonstrated highly significant strain-dependent differences in survival, defining a gradient from rapid embryonic lethality to delayed mortality, with one low-virulence strain exhibiting partial survival. While the λB-σB dual genotyping approach showed insignificant concordance with virulence clusters, a secondary structure-based μB-λC-σC triple model emerged as a better predictor of virulence. Notably, σC exhibited substantial variability in antigenic epitopes, whereas σB remained relatively conserved. These results might explain reduced vaccine efficacy against recent variants and support shifting vaccine design from σC to σB antigen or a combination to achieve broader protection. In conclusion, this multigene, structure-informed framework enhances ARV classification, links genotypes with virulence, and informs improved vaccination strategies. IMPORTANCE: Avian reovirus (ARV) represents a major threat to poultry health and production, primarily through its association with tenosynovitis/arthritis and the emergence of vaccine-resistant strains driven by high genetic diversity. Effective control requires an ARV classification system beyond traditional schemes based solely on the σC protein, which have proven insufficient to capture pathogenic diversity. We propose a novel, structure-informed triple genotyping approach incorporating the major capsid (μB), the turret (λC), and the virus attachment (σC) proteins. This method improves the prediction of disease severity in chicken embryos while reducing reliance on extensive animal challenge models. Furthermore, our protein-based analysis highlights that the conserved σB protein or a combination of σC & σB protein might be a promising vaccine target capable of providing broader protection across diverse ARV strains, thereby offering new avenues for both improved strain classification and rational vaccine design.