Short- and long-read metagenomics expand individualized structural variations in gut microbiomes

Abstract

In-depth profiling of genetic variations in the gut microbiome is highly desired for understanding its functionality and impacts on host health and disease. Here, by harnessing the long read advantage provided by Oxford Nanopore Technology (ONT), we characterize fine-scale genetic variations of structural variations (SVs) in hundreds of gut microbiomes from healthy humans. ONT long reads dramatically improve the quality of metagenomic assemblies, enable reliable detection of a large, expanded set of structural variation types (notably including large insertions and inversions). We find SVs are highly distinct between individuals and stable within an individual, representing gut microbiome fingerprints that shape strain-level differentiations in function within species, complicating the associations to metabolites and host phenotypes such as blood glucose. In summary, our study strongly emphasizes that incorporating ONT reads into metagenomic analyses expands the detection scope of genetic variations, enables profiling strain-level variations in gut microbiome, and their intricate correlations with metabolome.

Improved Assembly of Metagenome-Assembled Genomes and Viruses in Tibetan Saline Lake Sediment by HiFi Metagenomic Sequencing

“…Pacific Biosciences (PacBio) and Oxford Nanopore Technologies (ONT) have been the most popular long-read sequencing platforms recently. The development of TGS and corresponding advanced bioinformatics tools, along with affordable sequencing costs, has revolutionized de novo genome assembly, which may now retrieve the complete circular genomes of prokaryotes ( 22 ). These technologies have also increased the contiguity of assemblies of complex eukaryotic genomes by orders of magnitude, with read lengths exceeding 10 kbp ( 23 ).…”

“…These technologies have also increased the contiguity of assemblies of complex eukaryotic genomes by orders of magnitude, with read lengths exceeding 10 kbp ( 23 ). Introducing TGS into metagenomic studies would provide an advantage in contig assembly based on much longer reads ( 22 , 24 – 26 ). For example, 20 circularized complete MAGs (cMAGs) from 13 human stool samples were assembled using ONT long-reads, which provided opportunities to investigate the potential functions of repeat elements, although these cMAGs showed lower nucleotide accuracy ( 24 ).…”

“…Moreover, hybrid approaches combining NGS and TGS technologies to perform metagenomic assembly present high potential, including the generation of longer contigs and more complete MAGs for various habitats, such as the human gut, sludge, and soil ( 25 – 27 ). Additionally, genetic variations among microbial genomes in the gut have been identified by hybrid approaches and compared with data from healthy humans ( 22 ).…”

Pitfalls of genotyping microbial communities with rapidly growing genome collections

“…Benchmarking these methods in the context of metagenotyping (Andreu-Sanchez et al, 2021) would reveal if alignment uniqueness increases and/or cross-mapping decreases, as well as the computational resources needed to achieve performance advantages. Another way to increase SNV accuracy, while also disentangling strains of the same species present within the a sample, may be to metagenotype multiple co-occurring SNVs together using some combination of long reads (Chen et al, 2022; Xie et al, 2020; Yahara et al, 2021), haplotype assembly (Ghazi et al, 2022; Li et al, 2019; Pulido-Tamayo et al, 2015), and singlecell metagenomic sequencing (Cole et al, 2020). The idea is to leverage information from genetically linked sites to increase confidence in metagenotypes.…”

Noise reduction strategies in metagenomic chromosome confirmation capture to link antibiotic resistance genes to microbial hosts

“…Using sequencing-based approaches, it remains a challenge to link genes in the gut microbiota to their microbial hosts as metagenomic assemblies are often highly fragmented and metagenome assembled genomes (MAGs) are frequently incomplete or suffer from contaminating sequences [2,3]. While contiguity of assemblies can be improved by the incorporation of long-read sequencing data [4,5], the hosts of plasmids can only be predicted, but not conclusively identified, by a variety of bioinformatic approaches [6]. The linkage of genes to their microbial hosts is particularly important for genes that confer antibiotic-resistant phenotypes to their hosts.…”