SHIMS 3.0: Highly efficient single-haplotype iterative mapping and sequencing using ultra-long nanopore reads

Abstract

The reference sequence of structurally complex regions can only be obtained through a highly accurate clone-based approach that we call Single-Haplotype Iterative Mapping and Sequencing (SHIMS). In recent years, improvements to SHIMS have reduced the cost and time required by two orders of magnitude, but internally repetitive clones still require extensive manual effort to transform draft assemblies into reference-quality finished sequences. Here we describe SHIMS 3.0, using ultra-long nanopore reads to augment the Illumina data from SHIMS 2.0 assemblies and resolve internally repetitive structures. This greatly minimizes the need for manual finishing of Illumina-based draft assemblies, allowing a small team with no prior finishing experience to sequence challenging targets with high accuracy. This protocol proceeds from clone-picking to finished assemblies in 2 weeks for about $80 (USD) per clone. We recently used this protocol to produce reference sequence of structurally complex palindromes on chimpanzee and rhesus macaque X chromosomes. Our protocol provides access to structurally complex regions that would otherwise be inaccessible from whole-genome shotgun data or require an impractical amount of manual effort to generate an accurate assembly.

Fig 1. Overview of the SHIMS3.0 protocol.

A timeline of a single iteration of the SHIMS 3.0 protocol, showing the major protocol steps, with key quality controls on the right. During a two-week iteration, 24 clones are processed in parallel to rapidly generate finished sequence from structurally complex clones. A single technician can proceed from a list of clones to full-length nanopore libraries in 8 d. After a brief MinION run overnight, a bioinformatics specialist can demultiplex fastq sequences, identify full-length reads, then polish and edit the consensus of these reads to generate finished clone sequence.

Fig 2. Editing clone assemblies in Gap5.

Screenshots from Gap5 with reads sorted by technology (Illumina on top; nanopore on bottom), showing two instances where errors in the consensus can be resolved by correcting to the consensus of the Illumina reads: a) frequent insertion and deletion errors at homopolymer runs, and b) more rare substitution errors.