NEW YORK (GenomeWeb) – In conjunction with a related product launch, Phase Genomics has released early data demonstrating the potential of its Hi-C proximity chromatin capture-based "ProxiMeta" approach for deconvoluting microbial genomes from complex metagenomic samples.
"This method gives a cost-effective way for microbiome researchers to gain vastly better information out of samples," according to Phase Genomics Founder and CEO Ivan Liachko.
The Seattle, Washington-based firm started offering a ProxiMeta service in May. This week, Phase Genomics announced that it will also offer ProxiMeta as a kit, making it possible for investigators to perform the ProxiMeta experimental protocols in their lab at a reduced price relative to the existing service. The resulting data will be analyzed at Phase Genomics using an analytical pipeline consistent with that offered alongside the service.
"We've been developing the platform for the last few years and a few months ago we released it as a service," Liachko said. "We still offer that. But, because in the microbiome field people usually work with so many samples, they need something that's more scalable."
"We're releasing this kit-based package where … you get a kit in the mail, you do it yourself, you upload the data, and we will do the analysis. The analysis comes to you," he explained.
In a BioRxiv preprint online yesterday, Liachko and colleagues from Phase Genomics and the Los Alamos National Laboratory's Bioscience Division shared initial findings from ProxiMeta-based analyses of microbes in a single fecal sample from a healthy adult human.
By pairing Illumina Hi-C data with sequences generated by traditional shotgun metagenomic sequencing, they reportedly picked up more than 250 microbial genome clusters, including 50 almost-complete microbial genomes and 64 genomes that were more than 70 percent complete. Within that collection were genome sequences for 14 previously undocumented bacteria that were each at least 80 percent complete.
The investigators also touted the method as a means of tying plasmid sequences to specific host microbes. And, in their hands ProxiMeta compared favorably to an analytical approach called MaxBin that draws individual genomes out of a metagenomic mix based on read coverage- and tetranucleotide frequency-based binning.
"There's a concept going around of the 'dark matter' of the microbiome. What Hi-C metagenomics does is it allows you to basically delve into that dark matter," Liachko said. "You get high-quality, reference-grade genomes for things that no one's ever sequenced before."
Liachko said his team was keen to publish findings from the fecal sample application of ProxiMeta first, since the microbes found in this gut microbe community represent a good mix of new and known microbial genomes. He is scheduled to present related research at the American Society for Microbiology conference on applied next-generation sequencing and bioinformatics in Washington, DC next week.
"It's important when you look at a [preprint] article like this to take it with a few grains of salt," cautioned Christopher Beitel, who was not involved in the study.
Beitel was first author on a study published in PeerJ in 2014 that applied Hi-C to a simpler, synthetic metagenomic community. He was a doctoral researcher in Jonathan Eisen's University of California, Davis lab when that study was performed. He recently completed post-doctoral work at the Lawrence Berkeley National Lab and is now pursuing opportunities in the tech industry.
He would like to see more detail on the ProxiMeta experimental protocol and analytical software in the peer-reviewed version of the paper, but said the potential for investigating microbiome structure with this general type of method is high.
Beitel noted that it is a valuable accomplishment if they have indeed dialed in the Hi-C-based protocol and analysis such that it can be used reliably across a variety of types of communities. "The key next step is to look at the results that various users external to them get when using those kits and to look at their raw data and results," he said.
Targeted 16S ribosomal RNA sequencing is widely used to take a relatively quick and affordable look at the bacterial representatives present in a given microbial community. But that barcode-type approach does not provide information on full genome features or genetic variation that may exist within the species or genera being identified.
At the other end of the microbiome sequencing spectrum, metagenomic sequencing offers a look at sequences from across the genome in a mixed microbe collection. Teasing those sequences apart to pick individual genomes out of this genomic stew often relies on indirect sequence binning based on features such as read depth, contig sequence abundance, or sequence composition cues, Liachko noted.
"Sequencing has become cheaper, so more and more people are moving away from 16S-type approaches and moving into whole-genome sequencing — whole-genome metagenomics," he said. "The problem when you do that, of course, is that you're stuck with huge numbers of short contigs, short sequences, but you don't even know which organism they came from. So, you have to try to reconstruct that."
"It's sort of like taking hundreds of jigsaw puzzles, dumping all the pieces together and then trying to reconstruct all the puzzles," Liachko said.
While imputation-based methods have been useful, he said, they spit out individual genomes that are still approximations. The range of options available for successfully binning and assembling sequences from the same organism has started to expand as investigators interested in metagenomics start applying some of the proximity ligation methods used to gauge DNA interactions in the genome assembly setting.
In the case of Hi-C-assisted methods such as ProxiMeta, for example, researchers cross-link interacting bits of DNA in intact microbial cells before blasting them apart and sequencing the DNA-DNA junctions to retrace sequences originating in the same cell.
"You can use that data to separate all the jigsaw puzzles back to where they came from," Liachko said, returning to his pile-of-puzzles analogy. "It allows you to go into a metagenome sample, and with no a priori information, get many really high-quality genomes for things that you don't have to culture."
Along with the PeerJ study Beitel, Eisen, and colleagues published in 2014, Liachko was part of a team headed by the University of Washington's genome sciences researcher Jay Shendure that used Hi-C contacts to put together individual microbial genomes in a mixed microbial sample, as they reported in the journal G3 in 2014.
Dovetail Genomics, a Santa Cruz, California-based firm that offers its own Hi-C assembly services, is currently developing in vitro proximity ligation genome assembly methods and related informatics tools for metagenomics — an effort that will be aided by a two-year grant from the National Institutes of Health worth $938,000.
Both Dovetail and Phase Genomics representatives have said that crosslinking achieved by modified Chicago or ProxiMeta methods, respectively, can link plasmid sequences to broader genome sequences from its host strain.
Without such strategies, "there's nothing that will tell you where that plasmid came from, because it's not physically connected to the chromosome and they can jump around from species to species," Phase Genomics' Liachko noted.
Liachko declined to comment on the current price of the ProxiMeta kit, but said "the point of the kit is that it is cheaper" than the service. He suggested that investigators who are interested in applying the firm's approach should contact Phase Genomics for details on the kit and its optimization for a specific research context.
The firm is collaborating with several research teams and sees applications for its approach for studies focused on everything from to infectious disease to human microbiome profiling, agricultural, or environmental sample assessments.