Evolutionary Dynamics-Aware De Novo Design of a Thermostable NS1-Derived Diagnostic Protein for Yellow Fever Virus

Abstract Serological diagnosis of neglected arboviruses remains hindered by extensive intra-genus cross-reactivity, compromising both assay specificity and epidemiological accuracy. In this study, we focused on yellow fever virus (YFV) as a model system to design and validate a generalizable de novo protein design framework integrating evolutionary genomics with structural modeling to rationally engineer diagnostic antigens with reduced cross-reactivity and improved cost-effectiveness. Through comprehensive evolutionary mapping over time of the YFV polyprotein, we identified conserved yet discriminatory peptide regions within the NS1 protein and selected one peptide to guide scaffold-based protein design while preserving critical structural constraints. As a complementary strategy to develop antibody alternatives for antigen detection assays, we also implemented an in silico RNA aptamer design pipeline to circumvent the time and cost limitations of conventional in vitro Systematic Evolution of Ligands by EXponential enrichment (SELEX) technique. Nevertheless, NS1-targeting aptamers demonstrated limited sensitivity and specificity upon experimental validation. In contrast, structure-guided design of the NS1-derived protein (YFV-Scaffold) yielded a thermally stable, highly expressible recombinant protein, which demonstrated sensitivity and specificity comparable to those of commercially available full-length NS1 in in-house ELISA assays. Circular dichroism analysis revealed minimal denaturation up to 94 °C and complete recovery of secondary structure upon cooling, underscoring its remarkable structural resilience. Collectively, our results establish an evolutionary dynamics-aware strategy for target-antigen design, providing a scalable platform for the rational design of robust, low-cost diagnostic proteins applicable to a wide range of viral pathogens.