High-quality genome sequences of uncultured microbes by assembly of read clouds

Abstract

Although shotgun metagenomic sequencing of microbiome samples enables partial reconstruction of the strain-level community structure, it remains difficult to obtain high-quality microbial genome drafts without isolation and culture. Here we present a novel application of read clouds, short read sequences tagged with long-range information, to microbiome samples. We present Athena, a de novo assembler that uses read clouds to improve metagenomic assemblies. We apply this approach to sequence stool samples from two healthy individuals, and compare it to existing short-read and synthetic long-read metagenomic sequencing techniques. Read cloud metagenomic sequencing and Athena assembly produce the most complete individual genome drafts with high contiguity (>200 kbp N50, <10 contigs), even for bacteria that have relatively low (20x) raw short-read sequence coverage. We also sequence a complex marine sediment sample and generate 24 intermediate-quality genome drafts (>70% complete, <10% contaminated), nine of which are complete (>90% complete, <5% contaminated). Thus, our approach allows culture-free generation of high-quality microbial genome drafts using a single shotgun experiment.

Tracking microbial evolution in the human gut using Hi-C

“…The approach is well poised to make significant inroads towards an understanding of complex microbial community structures and dynamics, such as those found in soil, which routinely defy standard binning and other approaches. While promising alternative approaches based on long-reads exist 39,40 , Hi-C is notable for its ability to provide proximity information across millions of base-pairs of contiguous sequence, including inter-molecular contacts, as demonstrated by the association of plasmids with their respective host chromosomes.…”

Analysis procedures for assessing recovery of high quality, complete, closed genomes from Nanopore long read metagenome sequencing

“…However substantial limitations of these approaches have become evident, including problems related to the use of multi–sample co–assemblies [20, 74, 75], the challenges of resolving genomes to strain level [76], difficulties related to extracting MAGs from communities of high ecological complexity [77, 78], and the limitations of automated binning procedures, requiring careful evaluation of recovered genomes [79]. In response to these challenges, recent efforts have combined short read with emerging complementary techniques such as HiC metagenomics [80, 81, 82], synthetic long reads [83, 84], or long read sequencing, and collectively these results suggest substantial improvement can be made in the quality and completeness of metagenome assembled genomes using multiple types of sequence data. In the present study, we make use of DNA extractions that are co–assayed with both long and short sequencing, or DNA extractions from sampling events close enough together in time, that we can discount the influence of the ecogenomic differences as major influence on any observed differences between the types of sequence data.…”

Generating closed bacterial genomes from long-read nanopore sequencing of microbiomes

“…Short read and read cloud data and assemblies for samples P1 and P2-A were used without modification as previously described 9 . The standard approach used to extract DNA for these libraries produced fragmented DNA which incurred a severe loss with size selection, necessitating approximately 300 mg input stool to assure the 1 nanogram (ng) final HMW DNA mass required for read cloud library preparation.…”

“…Contigs from each approach were binned to form draft genomes, which were evaluated and assigned ‘High Quality’, ‘Complete’ and ‘Incomplete’ labels as described 9 . Briefly, drafts at least 90% complete and with at most 5% contamination are termed ‘Complete’, and drafts also containing at least one each of the 5S, 16S and 23S rRNA loci, as well as at least 18 tRNA loci, are labeled ‘High Quality’.…”