MOTIVATION: Identifying chromatin accessibility is one of the key steps in studying the regulation of eukaryotic genomes. The combination of exogenous methyltransferase and nanopore sequencing provides an strategy to identify open chromatin over long genomic ranges at the single-molecule scale. However, endogenous methylation, non-open-chromatin-specific exogenous methylation and base-calling errors limit the accuracy and hinders its application to complex genomes. RESULTS: We systematically evaluated the impact of these three influence factors, and developed a model-based computational method, methyltransferase accessible genome region finder (MAGNIFIER), to address the issues. By incorporating control data, MAGNIFIER attenuates the three influence factors with data-adaptive comparison strategy. We demonstrate that MAGNIFIER is not only sensitive to identify the open chromatin with much improved accuracy, but also able to detect the chromatin accessibility of repetitive regions that are missed by NGS-based methods. By incorporating long-read RNA-seq data, we revealed the association between the accessible Alu elements and non-classic gene isoforms. AVAILABILITY AND IMPLEMENTATION: Freely available on web at https://github.com/Goatofmountain/MAGNIFIER.