Misalignments in Generative Protein Design: Thermodynamic Filtering Fails to Predict Specificity

Generative protein design models frequently produce high-confidence binders $\textit{in silico}$, yet wet-lab evaluation demonstrates that these binders often lack the expected biochemical specificity. To probe the boundary conditions of current design and evaluation pipelines, we present a pilot failure-mode analysis by designing binders for asparagine deamidation. This single-residue modification consists of an electrostatic charge shift that can act as a rigorous test of highly specific binder design. When evaluating our designs against this target using a custom $\textit{in vivo}$ split-spGFP biosensor, we identified a confounding pipeline misalignment. While binding was achieved after target optimization, binding specificity was inverted compared to Rosetta's thermodynamic predictions. Curiously, retrospective cross-architecture analysis, using multiple folding models, revealed AlphaFold 3 correctly predicted the $\textit{in vivo}$ specificity natively. However, this prediction would have been overridden by the commonly used $\Delta\Delta$G thermodynamic metric used for evaluation. Ultimately, this case study demonstrates that appending standard thermodynamic filters to generative outputs can potentially degrade pipeline reliability. This exposes flaws both in generative models and in how post-hoc scoring weighs theoretical electrostatic rewards against energetic binding penalties.