A Comprehensive Investigation of Metagenome Assembly by Linked-Read Sequencing

Abstract

Background: The human microbiota are complex systems with important roles in our physiological activities and diseases. Sequencing the microbial genomes in the microbiota can help in our interpretation of their activities. The vast majority of the microbes in the microbiota cannot be isolated for individual sequencing. Current metagenomics practices use short-read sequencing to simultaneously sequence a mixture of microbial genomes. However, these results are in ambiguity during genome assembly, leading to unsatisfactory microbial genome completeness and contig continuity. Linked-Read sequencing is able to remove some of these ambiguities by attaching the same barcode to the reads from a long DNA fragment (10kb-100kb), thus improving metagenome assembly. However, it is not clear how the choices for several parameters in the use of Linked-Read sequencing affect the assembly quality.Results: We first examined the effects of read depth (C) on metagenome assembly from linked-reads in simulated data and a mock community. The results showed that C positively correlated with the length of assembled sequences but had little effect on their qualities. The latter observation was corroborated by tests using real data from the human gut microbiome, where C demonstrated minor impact on the sequence quality as well as on the proportion of bins annotated as draft genomes. On the other hand, metagenome assembly quality was susceptible to read depth per fragment (CR), and DNA fragment physical depth (CF). For the same C, deeper CR resulted in more draft genomes while deeper CF improved the quality of the draft genomes. We also found that average fragment length (μFL), had marginal effect on assemblies, while fragments per partition (NF/P) impacted the off-target reads involved in local assembly, namely, lower NF/P values would lead to better assemblies by reducing the ambiguities of the off-target reads. In general, the use of linked-reads improved the assembly for contig N50 when compared to Illumina short-reads, but not when compared to PacBio CCS (circular consensus sequencing) long-reads.Conclusions: We investigated the influence of Linked-Read sequencing parameters on metagenome assembly comprehensively. While the quality of genome assembly from linked-reads cannot rival that from PacBio CCS long-reads, the case for using Linked-Read sequencing remains persuasive due to its low cost and high base-quality. Our study revealed that the probable best practice in using linked-reads for metagenome assembly was to merge the linked-reads from multiple libraries, where each had sufficient CR but a smaller amount of input DNA.

Generation of lineage-resolved complete metagenome-assembled genomes by precision phasing

“…2 The number of MAGs that had at least one detectable alternative SNP haplotype allele. 3 Haplotype variant length represents the count of polymorphic SNP loci within an identified haplotype. 4 Genomic length was defined as the distance in bases on the assembled contig from the first polymorphic SNP site to the final site.…”

“…Generally, the MAGs assembled from short reads are represented by hundreds or even thousands of contigs, many of which have fragmented open reading frames (ORFs) at their ends. A major source of discontinuity in metagenome assembly appears to be from the prevalence of high sequence identity orthologous genes and operons 3 . These genomic features tend to be repetitive in the community and can preclude complete assembly unless the data include sequencing reads that span the entire shared region.…”

“…The recent development of highly accurate HiFi reads from circular consensus sequencing (CCS) on the Pacific Biosciences platform resulted in long accurate reads with error rates below 1% across the length of the read 19 providing opportunity to improve assembly quality 20 and even resolve both haplotypes of diploid genomes 21,22 . Recent attempts to sequence and assemble metagenomes with long error-prone reads [23][24][25] or linked reads 3 have resulted in few successes. While it has been demonstrated that long error-prone reads result in longer contigs than short reads 23,24 , the assembly of nearly all members of the community in singular circular contigs has still been elusive.…”