Canu

Canu is a free, open-source genome assembler designed for long-read sequencing data, such as those produced by Pacific Biosciences SMRT sequencing and Oxford Nanopore Technologies. It is a continuation of legacy Celera Assembler work and is widely used for de novo genome assembly, including large and complex genomes. Canu focuses on handling the high error rates characteristic of long reads and aims to produce accurate assemblies from single data sets or mixed data.

The Canu workflow consists of three main stages: read error correction, read trimming, and assembly. The software employs an overlap-layout-consensus approach and emphasizes adaptive k-mer weighting to distinguish repeats and improve assembly continuity. This design helps the assembler scale to large genomes and to manage heterozygosity and repeats that complicate assembly. Polishing steps can be used to improve consensus accuracy, often by integrating short-read data or other polishing tools after the initial assembly.

Canu runs on Unix-like operating systems and is typically deployed on high-performance computing infrastructure due to substantial memory and CPU requirements, especially for large eukaryotic genomes. Outputs commonly include assembled contigs or unitigs, along with summary statistics and optional assembly graphs. The pipeline accepts various long-read chemistries and can be tuned to genome size, read length, and expected error profiles.

Impact and usage: since its introduction, Canu has become a standard tool in de novo genome assembly, applied to bacteria, fungi, plants, and animals. It is frequently used in combination with polishing steps and supplementary data to produce high-quality genome assemblies and has contributed to numerous reference-grade genomes in genomics research.