ONT basics

A high-level overview of nanopore sequencing is provided by many reviews, including our structural variant (SV) review:

• https://pubmed.ncbi.nlm.nih.gov/37931775/

Key features exploited by functions in ont-mdi-tools include:

• (very) long reads support improved structural assembly and SV detection
• duplex sequencing of both DNA strands supports high accuracy basecalling
• direct sequencing of genomic DNA molecules supports modification analysis

ONT support software

Detailed descriptions of Oxford Nanopore Technologies (ONT) methods and data types is beyond our scope here, but well documented by ONT:

• ONT on GitHub: https://github.com/nanoporetech
• POD5 files and utility: https://github.com/nanoporetech/pod5-file-format
• Dorado basecaller: https://github.com/nanoporetech/dorado
• Remora modification modeleing: https://github.com/nanoporetech/remora

ont-mdi-tools provides wrappers that make it easier to install and use these programs.

POD5 files are IO intensive

ONT flowcells generate electrical signal data as “traces”, or “squiggles”, one DNA molecule at a time, in each of up to a few thousand nanopore “channels”. That data are stored in POD5 files for basecalling and modification analysis.

It is critical to understand that:

• POD5 files can easily grow to 1T or more in net size
• programs that use POD5 files are often IO-bound

Thus, disk speed is a main limit on program performance. ont-mdi-tools implements strategies to maximize file IO to keep processing speed high. Large files are moved en bloc to fast drives for iterative access by ONT software.

---