Nanoq: ultra-fast quality control for nanopore reads

Nanopore sequencing is now routinely used in a variety of genomics applications, including whole genome assembly One of the first steps in many workflows is to assess the quality of reads, obtain summary statistics, and filter fragmented or low quality reads. With increasing throughput on scalable nanopore platforms like GridION or PromethION, fast quality control of sequence reads and the ability to generate summary statistics on-the-fly are required. Benchmarks indicate that nanoq is as fast as seqtk for small datasets (100,000 reads) and ~1.5x as fast for large datasets (3.5 million reads). Without quality scores, computing summary statistics is around ~2-3x faster than rust-bio-tools and seq-kit stats, 44x faster than seqtk, and up to ~450x faster than NanoStats (> 1.2 million reads per second). In read filtering applications, nanoq is considerably faster than other commonly used tools (NanoFilt, Filtlong). Memory consumption is consistent and tends to be lower than other applications (~5-10x). Nanoq offers nanopore-specific quality scores, read filtering options, and output compression. It can be applied to data from the public domain, as part of automated pipelines, in streaming applications, or to rapidly check progress of active sequencing runs.