Enhancing Long-Read-Based Strain-Aware Metagenome Assembly

Abstract

Microbial communities are usually highly diverse and often involve multiple strains from the participating species due to the rapid evolution of microorganisms. In such a complex microecosystem, different strains may show different biological functions. While reconstruction of individual genomes at the strain level is vital for accurately deciphering the composition of microbial communities, the problem has largely remained unresolved so far. Next-generation sequencing has been routinely used in metagenome assembly but there have been struggles to generate strain-specific genome sequences due to the short-read length. This explains why long-read sequencing technologies have recently provided unprecedented opportunities to carry out haplotype- or strain-resolved genome assembly. Here, we propose MetaBooster and MetaBooster-HiFi, as two pipelines for strain-aware metagenome assembly from PacBio CLR and Oxford Nanopore long-read sequencing data. Benchmarking experiments on both simulated and real sequencing data demonstrate that either the MetaBooster or the MetaBooster-HiFi pipeline drastically outperforms the state-of-the-art de novo metagenome assemblers, in terms of all relevant metagenome assembly criteria, involving genome fraction, contig length, and error rates.

Keywords: genome assembly; haplotype; long reads; metagenome; strain.

FIGURE 1

Workflow overview of MetaBooster and MetaBooster-HiFi. MetaBooster takes raw reads as input and outputs contigs using Canu (blue), whereas MetaBooster-HiFi takes raw reads as input and outputs contigs using HiCanu (red). The dotted arrow lines denote running Strainberry is optional. Canu (correction) means running Canu’s correction and trimming modules only. Canu (assembly) means running Canu’s assembly module only.