NEW YORK (GenomeWeb) – After rising from the ashes earlier this year, electronic DNA mapping company Nabsys has been busy generating its first long-range maps of whole human genomes and using them to call structural variants, work done in collaboration with the Genome in a Bottle (GIAB) consortium.
The GIAB consortium has been developing reference samples of well-characterized human genomes, with a recent focus on structural variant calls. At a workshop last month held by the consortium at the National Institute of Standards and Technology, Nabsys presented summary data for its maps for the first time, and last week it uploaded individual calls to the consortium's site.
The new human genome maps are the first fruits of the company's labor since it restructured financially and restarted earlier this year with several members of its original management team, after going out of business in 2015. The initial goal of the revived team was to "demonstrate that we could generate these maps and that these maps would be very good at analyzing structural variation at high sensitivity and specificity across whole human genomes," Barrett Bready, Nabsys' original and current CEO, told GenomeWeb. "From this point, the plan is to scale the company and allow this [technology] to be available commercially," he said. "We think this is an important inflection step in the company."
A little more than a year ago, Nabsys, a spinout from Brown University, closed its doors, following a change in management and focus in 2014. Bready said that under the interim leadership, the company had moved from analyzing long DNA molecules to studying shorter DNA fragments of cancer genomes, an approach that the firm said was ultimately unsuccessful.
After restructuring the company financially earlier this year, Bready reassembled some of the core team, including Chief Technology Officer John Oliver, and refocused the company on working with very long pieces of DNA in order to "demonstrate that we could look at structural variation in a systematic way across genomes," he said.
Barrett declined to comment on Nabsys' financial restructuring, its current funding, or its plans for scaling and commercializing its technology. "We're trying to get this to the market as quickly as we can," he said.
The firm currently has 12 employees and is still housed at its original location in Providence, he said. Local news outlets reported in late 2015 that Bready had purchased the company for $500,000 and assumed $100,000 of its debt. The new legal entity is called Nabsys 2.0, but the firm still operates under the name Nabsys.
Nabsys' approach uses single DNA molecules with sequence-specific probes bound to them. Using an electric field, the molecules are driven through nanodetectors at high speed, which pick up signals from the probes. The company analyzes these signals to measure the distance between probes and call structural variants.
For the GIAB project, Nabsys researchers created electronic maps for two samples, NA12878, a well-characterized HapMap sample originating from Utah, and NA24385, the child of an extensively studied parent-son trio of Ashkenazi Jewish origin. "We were able to see what sensitivity and specificity we have as a function of [variant] size, and how that compares to other methodologies," Bready said.
Justin Zook, a biomedical engineer at NIST and one of the leaders of the GIAB consortium, said that the group plans to make high-confidence structural variant calls for the parent-son trio by integrating data from various platforms, including Nabsys, BioNano Genomics, Illumina paired-end and long mate-pair sequencing, 10X Genomics, BGI's Complete Genomics, and Pacific Biosciences. "Nabsys has been active in developing and optimizing methods to use their mapping data to corroborate calls found by other methods," he said.
Overall, the Nabsys team was able to call structural variants 300 base pairs or larger on long DNA molecules on the order of 100 kilobases, which Bready said is a range that short-read sequencing technologies in particular have trouble with. "The SNV and small indel work is much further along, there is much better agreement among the different callers," he said, "whereas with structural variation, it is as much art as science at this point, there is a lot less agreement." Going forward, the company plans to further improve its analytics for making variant calls, he added.
The advantage of Nabsys' approach of electronic signal acquisition, as opposed to optical maps, is twofold, he said: the resolution is not limited by the wavelength of light, and electronic detection is very fast, with DNA molecules passing through the nanodetector at a speed of one megabase per second. "At that velocity, we are able to get all the information we need to create these maps," he said, and scaling the solid-state semiconductor-based technology to larger numbers of detectors would enable very high throughput.
Also, the data analysis is automated and "statistically rigorous" across the genome, and does not require manual visual inspection. "We feel pretty good about the sensitivity and specificity of our maps and the software we generated to date to systematically look across the genome," he said.
While a number of other technologies, including those used by the GIAB consortium, also go after intermediate-size and large variants, Bready still believes there is a niche for Nabsys' technology. "In terms of market need, I don't see anything at the moment that obviates the need for this type of information," he said.
Researchers interested in looking at Nabsys' data more closely can join the GIAB consortium or can request the data directly from the company, Bready said.