By Monica Heger
When paired with next-gen sequencing technology, OpGen's Argus optical mapping technology can correct errors in assembled genomes and help close gaps, a company official said last week at a presentation during a one-day conference of BGI users in Rockville, Md.
Trevor Wagner, a senior scientist at OpGen, presented data on how the company has used the Argus platform to find errors in microbial assemblies from the Human Microbiome Project, as well as in finished genomes, and to close introduced gaps in sequenced human genomes.
While the platform has mostly been used for smaller genomes like bacteria and microbes, Wagner said that the company is now also moving into mammalian and plant genomes, using a "hybrid approach" that combines next-gen sequencing with single-molecule restriction maps.
In order to create its optical maps, OpGen first stretches DNA on a glass surface, then uses a restriction enzyme to cut the DNA. Because the cut is made on the glass surface as opposed to in solution, the order of the fragments is maintained. Then, the company takes thousands of images across the glass surface, annotates and measures the fragments, and converts the optical data into digital data to construct a single-molecule map.
The technology generates overlapping patterns, which can be lined up to perform a de novo assembly without a reference genome. "At each location, we have a certain amount of depth that votes 'yay' or 'nay' on restriction location, which then converts to a consensus map," said Wagner.
One way Wagner said the company has been using the technology is to do comparative genomics. This can be particularly useful for an unknown bacterial strain to quickly see how similar it is to a known strain. For instance, he said the company compared two different strains of Enterococcus faecalis, bacteria that can cause urinary tract infections, meningitis, and other infections.
The team compared strain 49477, which has not been sequenced, to strain v583, which is vancomycin resistant. Constructing an optical map of 49477 allowed the researchers to identify the portions of the genome that differed from v583.
"We found a large area of DNA sequence in the v583 strain that is different from the strain we mapped," Wagner said. It was the portion that contained the genes that confer vancomycin resistance, "so 49477 is probably not resistant because it doesn't contain those genes," he added.
Indeed, the company recently put its technology to work to help sequence the genome of Escherichia coli O104 — the strain linked to the recent outbreak in Europe. OpGen worked with the University of Münster to do an assembly of the genome based on data from Life Technologies' Ion Torrent PGM. The company used the Argus system to produce whole-genome optical maps of six isolates and then used the maps to orient contigs to produce a complete, whole-genome sequence of the pathogen (IS 6/21/2011).
The whole process took just 48 hours from the time the company received the samples, to producing de novo, whole genome maps of all six isolates.
OpGen has also tested its technology on sequenced genomes from the Human Microbiome Project, to help in closing gaps in assemblies. In the process, it unintentionally found that some of the finished genomes, sequenced next-gen technology contained assembly errors.
Wagner said the OpGen team produced maps for seven organisms that were sequenced as part of the microbiome project, and found that three of them had at least one misassembled contig. "We observed this by seeing that a contig mapped to two different places in the genome," he said.
In order to confirm that the findings were real, Wagner said the team generated optical maps using a completely different restriction enzyme, and still obtained the same results.
"Optical mapping is a sequence-independent method," he added, so it could "aid in validation of sequence assembly, finishing, and improving the quality of assemblies."
While the genomes tested from the Human Microbiome Project were "unfinished," Wagener said the company also found many "discrepancies across published, finished genomes."
For example, using optical maps, OpGen researchers found 375 kilobases more data in the Saccharomyces cerevisiae genome than what was found in the reference, corresponding to a repetitive region.
In total, he said the company looked at 13 "finished" genomes sequenced with next-gen platforms and found discrepancies in eight of them. "Errors are being propagated because of resequencing against reference genomes that are wrong," he said.
While OpGen's technology has been primarily used on microbial genomes, Wagner said the company has also been moving into larger genomes. Last fall, the company announced it was collaborating with BGI to use its optical mapping technology for de novo sequence finishing in human, plant, and animal genomes (IS 11/9/2010).
Recently, in collaboration with BGI, the company tested its ability to close gaps in the human genome. It introduced gaps into the human reference genome of up to 200 kilobases and found that it could close 91 percent of them with optical data.
Additionally, optical maps helped to improve the de novo assembly of the goat genome. BGI sequenced and assembled the goat genome, using its SOAPdenovo sequencer, producing 1,236 scaffolds with an N50 of 2.29 megabases. Optical mapping reduced the number of scaffolds to 181 and increased the N50 to 16.89 megabases.
For de novo assemblies of these larger genomes, OpGen is using its technology along with next-gen sequencing to help "grow" scaffolds to "generate super scaffolds," Wagner said, describing it as a "hybrid approach" to assembly.
In this process, the company generates single-molecule restriction maps and matches them to scaffolds produced from next-gen sequencing data. The restriction maps should provide more data, which will allow scaffolds to be stretched out or bridged, Wagner said.
While Wagner did not mention other competing technologies, BioNanomatrix is developing technology for similar purposes, with its first expected application to be in sequence assembly. Its technology analyzes single molecules of fluorescently labeled DNA with a CCD camera as it flows through nananochannels. While the company has not yet commercially launched its system, it is testing prototypes at four early test sites and is planning for a commercial launch in 2012 (IS 6/21/2011).
Additionally, OpGen may also face competition from Pacific Biosciences. Users of the PacBio RS have said that one potential application of the machine could be in de novo microbial sequencing and assembly, due to the machine's long reads. Early users reported in May on the system's ability for whole-genome sequencing of bacteria genomes, as well as using the machine along with Illumina to perform a hybrid assembly (IS 5/17/2011).
Have topics you'd like to see covered by In Sequence? Contact the editor at mheger [at] genomeweb [.] com.