Bioinformatics solutions for addressing post-processing challenges in Nanopore sequencing data

Nanopore sequencing (Oxford Nanopore Technology) has rapidly become more accessible and widely adopted in research laboratories, offering a real-time, cost-effective, and portable sequencing technology. This advancement has significantly accelerated scientific discoveries, particularly in metagenomic studies, by enabling comprehensive analyses of diverse microbial and viral communities. However, the inherent characteristics of Nanopore sequencing introduce several challenges. The high error rate, for instance, complicates the assembly of reads, as well as the detection and removal of Nanopore adapters and barcodes. Additionally, the ligation of adapters to DNA fragments (using the Ligation Sequencing Kit) and/or the overloading of DNA in the flowcell, appear to contribute to a substantial number of chimeric and inverted repeated reads, especially when dealing with short DNA fragments. These issues significantly complicate the processing of sequencing data from multiplexed DNA samples, as many reads are either misallocated to incorrect samples (e.g., chimeric reads) or poorly assembled (e.g., inverted repeated reads). Existing bioinformatics tools, such as Dorado, Guppy, and Porechop, do not fully address these challenges. Consequently, we identified the need for specialized, custom-built tools designed to tackle these specific problems. To address this, we are currently developing three computational tools: . DeSIR: a tool for detecting and splitting inverted repeat reads, . Tripatouille: a tool for identifying and removing adapters and barcodes (custom or Nanopore), particularly suited for demultiplexing and cleaning multiplexed amplicon reads, such as those derived from the virion-associated nucleic acid (VANA)-based metagenomics approach, . ChimeraKiller: a tool for detecting and splitting chimeric reads (currently under development). These tools, still in the prototype phase, require further refinement, particularly in terms of execution speed and adjustments to sensitivity and specificity.