A detailed guide to assessing genome assembly based on long-read sequencing data using Inspector

Abstract

Long-read sequencing technologies yield extended DNA sequences capable of spanning intricate, repetitive genome regions, thereby facilitating the generation of more precise and comprehensive genome assemblies. However, assembly errors are inevitable owing to inherent genomic complexity and limitations of sequencing technology and assembly algorithms, making assembly evaluation crucial. The genome assembly evaluation tool Inspector presents several advantages over existing long-read de novo assembly evaluation tools, including (1) both reference-free and reference-guided assembly evaluation; (2) the ability to detect both small- and large-scale structural errors; (3) the option of assembly error correction, which can improve the quality value of the original assembly; and (4) the ability to perform haplotype-resolved assembly evaluation. Inspector can provide not only basic contig and alignment statistics, but also the precise locations and types of the different structural errors. These advantages provide a robust framework for long-read assembly evaluation. In this Protocol, we showcase four procedures to demonstrate the different applications of Inspector for long-read assembly evaluation. Inspector software and additional guides can be found at https://github.com/ChongLab/Inspector_protocol . This protocol describes a versatile computational tool for reference-free or reference-guided assessment and correction of errors in de novo genome assembly based on long-read sequencing data, showcased in four different use-case scenarios.