OpenBase: a universal framework for high-accuracy single-molecule detection of diverse non-canonical DNA bases using nanopore sequencing

Abstract

Nanopore sequencing holds great potential for the direct detection of non-canonical DNA bases from electrical signals, yet current approaches remain limited to a few classical epigenetic marks. Here we present OpenBase, an open and universal framework that standardizes training-data generation and deep learning based-modeling for single-molecule identification of diverse non-canonical DNA bases. OpenBase transforms nanopore sequencing into a broadly accessible platform for exploring the chemical diversity of DNA.