BPS2026 – Peptide sequencing with single-acid resolution using a sub-nanometer-diameter pore

Protein sequencing with high sensitivity is fundamental for understanding biological function and identifying disease biomarkers but remains a major challenge. Existing approaches, including mass spectrometry, typically required millions of molecules, limiting sensitivity for rare targets. Here, a label-free sub-nanopore based approach is demonstrated that can directly “read” peptide sequences at single-amino acid (AA) resolution and build consensus from only a few hundred molecules. A 0.3–0.4 nm diameter pore with biconical geometry was fabricated in a few-nanometer-thick amorphous silicon (a-Si) membrane using a tightly focused 300 keV electron beam sputtering in an aberration-corrected scanning transmission electron microscope (STEM). An applied electric field drives denatured peptides through the pore, producing partial ionic current depending on the properties of individual AA. Signals were then amplified at 1.8 MHz bandwidth and sampled at 3.6 MS/s. Signal to noise ratio was improved by building consensus from clusters of similar blockades generated using affinity propagation (AP) algorithm. To interpret these fluctuations, molecular dynamics (MD) simulations incorporating residue volume, charge, mobility, solvation, and pore geometry was used to model the pore currents. Sequencing capability was demonstrated using amyloid-beta (Aβ 1-42 ), an Alzheimer’s biomarker, its scrambled variant (SAβ 1-42 ), and peptides with post-translational modifications and point mutations. Comparison of z-scored blockade currents from experiment and MD model yielded Pearson correlation coefficients (PCCs) of 35%–42%. A UniProt-derived database of 1,843 human peptides with 40–44 residues were then compared using PCC, true peptides ranked to top 12. To address mobility variations, dynamic time warping (DTW) was applied, improving PCCs to 92%–96%, a near perfect match, and elevating the true peptides to the top-rank among top 15 candidates. These results demonstrate that label-free, single-amino acid resolution sequencing with high-sensitivity is feasible with sub-nanopores.