Dataset for: Comparative study of nanopore phenylalanine clamp variants reveals unique peptide biosensing and classification properties

Abstract Rationally engineering biological nanopores is critical for advancing single-molecule biosensing. Here, we investigate the phenylalanine clamp active site (ϕ clamp) of the anthrax toxin protective antigen (PA) nanopore, a key site for molecular interaction, to test if engineering this site can improve peptide classification. We performed a comparative analysis of wild-type PA and two ϕ-clamp mutants (F427A, F427Y). We report the paradoxical finding that the F427A mutant—known to be a defective large protein translocase—is a superior peptide biosensor. Using a machine learning framework with engineered biophysical features, the F427A pore classifies a diverse peptide set with 93% accuracy. Our analysis suggests this enhanced performance arises because the F427A mutation, while weakening specific interactions, produces more consistent, lower-variance kinetic ‘fingerprints’ that are more easily distinguished by computational models. These findings establish a principle for biosensor design and enable a strategy where engineered pores with complementary specificities are deployed in multiplexed arrays for robust diagnostics.