Machine Learning-Assisted Discrimination of Isobaric Histone Acetylations by Nanopore

Abstract

Histone acetylation, a post-translational modification, is associated with cancer progression, invasion, and metastasis, serving as a promising biomarker. The accurate discrimination of acetylation sites has strong potential for cancer diagnostics. Herein, we engineered an R220Q aerolysin nanopore to detect isobaric histone acetylations. By exposing the paired negative charge (D222) at the pore entrance, peptides with a higher positive charge density encounter a lower U* for pore entry and subsequent translocation. With the assistance of machine learning, we achieved direct discrimination of isobaric acetylations in histone peptides. Monoacetylated peptides H3K9ac and H3K14ac, as well as diacetylated peptides H3K4acK9ac and H3K4acK14ac were unambiguously identified in mixtures with identification accuracies of 99.4% and 98.8%, respectively. This work offers a label-free approach for analyzing acetylated biomarkers with potential applications in cancer detection and drug target development.