NEW YORK (GenomeWeb) – San Diego-based Bionano Genomics has developed a method for improving its single-molecule optical genome maps — increasing the size of the sub-maps in some cases up to chromosome length — using a new Direct Labeling and Staining (DLS) chemistry it plans to release more broadly toward the end of this quarter.
In contrast to Bionano's current "Nick, Label, Repair, and Stain" (NLRS) strategy, which dings targeted motifs on individual DNA strands with nickase enzymes during the labeling process, the DLS chemistry directly labels targeted motifs without introducing break-prone sites in the DNA strand.
Broadly speaking, the firm's optical mapping method involves enzymatically labeling very long strands of native DNA at specific recurring sequence sites, flowing single DNA molecules into nanochannels on a chip, and detecting these labels after loading the chip into the Saphyr instrument.
"We detect those labels and also detect the base pair distance between those labels," explained Mark Borodkin, chief operating officer at Bionano. "We use that information to create true de novo maps across the whole genome. Those maps, then, are used to either anchor sequence information — sequencing contigs for genome assembly projects — or to detect large structural variations of all kinds."
Up until now, the NLRS chemistry has made it possible to generate maps spanning the genomes of many organisms, Borodkin said, though each genome map was typically made up of many shorter maps. And for many applications, investigators needed to run the mapping protocol twice, using two complementary NLRS enzymes.
He said the next-generation DLS chemistry "resolves some key limitations of NLRS," generating chromosome-arm-length maps with a single run using an enzyme that recognizes sequences that are commonly found in a range of genomes — from humans and other animals to crop plants such as maize.
"From the application point of view, the map contiguities on the scaffolds have increased substantially, over 50 times, and our [structural variant] detection has improved, especially for more complex events," Borodkin said, "and all of this with a single enzymatic reaction that costs less than $1,000, all-in cost, per sample."
More generally, Borodkin called 2017 "a key year" for Bionano, noting that the company's technology was used in genome assembly or structural variant detection in dozens of published studies that focused on everything from human health-related research to plant and animal traits.
Bionano Genomics introduced its initial Irys instrument in 2012, after publishing a proof-of-principle study with collaborators at the University of California, San Francisco, and the Genome Institute of Singapore on a nanochannel array-based genome mapping method for assembly and structural variant profiling. That instrument made its way into the European market in 2013, alongside research demonstrating its feasibility for human genome mapping.
Early last year, at the annual Advances in Genome Biology and Technology meeting, the firm introduced a speedier, higher-throughput optical mapping instrument known as Saphyr. That instrument retained the nanochannel-based approach introduced with Irys while incorporating faster scanning optics and a higher-capacity chip.
Now, Bionano is getting its DLS chemistry into the hands of early-access users working on a range of research projects. Borodkin said there are currently around 10 early-access DLS users at various stages of evaluating the approach, including investigators from the Vertebrate Genomes Laboratory (VGL) at Rockefeller University.
At the Plant and Animal Genomes conference in San Diego this month, researchers from the VGL reported that they are using the Saphyr instrument to provide optical mapping data that is incorporated into the center's current genome assembly pipeline, alongside Pacific Biosciences long sequence reads, 10x Genomics linked reads, and Arima Genomics Hi-C data.
The VGL is one of three sequencing hubs for the Vertebrate Genomes Project (VGP) — an ongoing effort to develop reference quality genome assemblies for essentially all of the 66,000 or so extant vertebrate species.
Before officially kicking off VGP, members of the G10K consortium-led project did extensive benchmarking of a range of the latest sequencing, genetic mapping, and assembly strategies using an Anna's hummingbird sample. The quality and contiguity metrics for that genome have informed the pipelines being used to produce assemblies for phase 1 of VGP, which aims to establish reference-quality genomes for representatives from 261 vertebrate orders.
In collaboration with Bionano Genomics, the VGL team has taken a crack at incorporating optical maps generated with the new DLS chemistry into at least two of the phase 1 assemblies, explained VGP co-director Olivier Fedrigo.
"We've been working with Bionano to test DLS," Fedrigo said, noting that the VGP team sent the firm DNA from the Kakapo (Strigops habroptilus), a critically endangered bird species from New Zealand, and a common blackbird (Turdus merula) representative from Germany.
In a poster presented at PAG, Fedrigo and co-authors from Rockefeller and the National Human Genome Research Institute reported that they achieved a contig N50 of 5.77 megabases and a scaffold N50 of 73.81 megabases for the Kakapo genome, based on PacBio long reads and Bionano optical mapping with DLS chemistry alone. The PacBio and Bionano DLS-based assembly for the blackbird currently has a contig N50 of 1.86 megabases and scaffold N50 of 41.7 megabases.
The authors noted that they plan to include additional 10x Genomics and Hi-C data into the Kakapo and blackbird assemblies. Still, Fedrigo said the genome mapping results they achieved with the DLS chemistry were "spectacular."
He noted that VGL researchers have just started to test the new kits in their own lab, and have not yet determined whether they will switch over to the DLS chemistry for the remaining phase 1 genomes being sequenced there.
"We're happy with the preliminary results, but I think we're going to test a bit more how robust this method is," Fedrigo said. "We have so many genomes, we want to not have to troubleshoot every time, so those are tests we're doing right now."
Another early-access team, from Genoscope in France, presented research on PAG that used Bionano Saphyr mapping with DLS chemistry as part of a study on large genomes for plants in Musa (banana) and Brassica genera.
In some plants and animals, Borodkin said, the DLS chemistry has helped to grow the sub-maps to chromosome length, raising the possibility that it might be feasible for some teams to opt out of a Hi-C mapping step to get long range information when assembling such genomes — a possibility that is being explored by early-access users.
"There's a certain financial motivation for them to be able to remove Hi-C, as it is a relatively expensive add-on," he said.
Borodkin cautioned, though, that the DLS chemistry still cannot get to full chromosome-length sub-maps for all organisms. In the human genome, for example, the approach does not span the centromere, leaving a gap that investigators seeking chromosome-level contiguity may want to tackle with Hi-C or other proximity ligation methods.
The firm has not yet released the price of DLS, but Borodkin noted that cutting out a nick-repair step needed for the NLRS protocol translates into a simpler and speedier sample labeling reaction. The DLS reagent kit will be compatible with Bionano's Saphyr instruments and will include related bioinformatics software.
"Workflow-wise, we're getting feedback from our early-access customers that it seems very compatible with what they've been doing so far — so, in line with what they're used to already. There's not much of a change in procedure," said Goran Pljevaljcic, Bionano's director of strategy and corporate development.
Early-access users have not yet come across genomes with label densities that are too low for successful mapping based on sites targeted by the DLS enzyme system, according to Borodkin, who explained that there appears to be a "wide range of label density that's acceptable within the system, so we're confident that it will fit most needs."
"We have run through an analysis of the common genomes that people work on," Pljevaljcic added, "and have a pretty extensive list of genomes where we have tested whether the label density is amenable … and found that that is the case."
Still, Bionano is aware that some genomes might turn up with exceptionally low label density using this first DLS enzyme and is continuing to assess additional enzymes. More broadly, the company is continuing to work on methods to improve the Saphyr instrument's throughput and to expand and improve bioinformatics tools for interpreting the mapping data.