Abstract
FBXO11 binds SKP1 within an SCF E3 ubiquitin ligase complex, where it recognizes substrates for ubiquitination and degradation. Pathogenic FBXO11 variants cause neurodevelopmental disorders, yet several retain normal SKP1 binding, indicating that disrupted SKP1 binding alone cannot account for disease. We integrated multi-conformational AlphaFold3 models with FoldX and Rosetta stability predictions to evaluate 44 missense variants (23 pathogenic, 21 benign). Pathogenic variants showed greater predicted destabilization than benign controls by both methods (FoldX 1.86 vs. 0.54 kcal/mol, Cohen's d = 0.50; Rosetta 6.85 vs. 1.31 kcal/mol, Cohen's d = 0.78; both p < 0.01, FDR-corrected). Discrimination was consistent in both the FBXO11 monomer and the FBXO11-SKP1 complex contexts. Among ten pathogenic variants with experimentally validated normal SKP1 binding, 9/10 by FoldX and 8/10 by Rosetta exceeded a 1.0 kcal/mol threshold (10/10 by either method). Benchmarking against AlphaMissense, REVEL, and CADD indicated that physics-based ΔΔG provides complementary mechanistic information to existing predictors. Exploratory molecular dynamics simulations (300 ns total) suggested elevated backbone RMSD in pathogenic variants. These findings support a hypothesis of binding-independent destabilization in FBXO11-associated pathogenesis, although requiring further computational analysis and experimental validation.