NEW YORK (GenomeWeb) – Scripps Research Institute spinout iGenomX aims to improve human genome assembly from linked DNA reads by employing a new sequencing library construction chemistry it has developed that promises to be easier and less biased than existing approaches.
Over the next half year, the La Jolla, California-based company plans to make data and analysis software available, start a beta testing program, and roll out a commercial platform, CEO Keith Brown told GenomeWeb. "The main feature is our ability to provide uniform long-range information at a cost on par with what's been done today," he said.
While the company's library prep approach could benefit a number of next-gen sequencing applications, iGenomX plans to develop it first for generating linked sequence reads from long isolated DNA fragments to improve de novo assemblies of human genomes. For that, the iGenomX chemistry would work well in combination with platforms for isolating DNA molecules, for example 10X Genomics' microdroplet-based DNA partitioning system, Brown said.
IGenomX was founded in 2012 by researchers at TSRI and exclusively licenses patent-pending library prep technology developed by the laboratory of Daniel Salomon, a professor at Scripps. Salomon's research focuses on transplant immunology, and he directs the Laboratory for Functional Genomics.
Originally, the company, which has its own space within TSRI and currently has seven employees, had planned to offer clinical sequencing services for recessive genetic disorders in the context of immunology. But the firm found that current sequencing technology was missing a large percentage of pathogenic variants, including translocations, insertions, and deletions, so instead it decided to commercially develop sequencing library preparation chemistry from Salomon's lab.
IGenomX became operational in 2014, and has to date raised approximately $3 million in funding from a San Diego-based venture capital firm and from angel investors. According to a filing with the Securities and Exchange Commission last year, the funding is part of a $4 million round but Brown said the firm is not looking for additional funding at the moment.
The company's chemistry relies on generating copies of the target DNA, using random primers carrying an adaptor and a DNA polymerase. The length of the copies is restricted by including a small amount of dideoxy terminator nucleotides in the polymerase reaction. The terminators are labeled with biotin, so the single-stranded DNA products can be captured on streptavidin beads, where a second round of random priming and extension introduces another adaptor. Primers binding closest to the bead will displace all previously generated shorter DNA products in that process. After further amplifying the dual-adaptor library molecules, they are ready for sequencing.
Limiting the length of the polymerase products with terminators helps prevent bias, Brown explained, because in longer copies, certain regions are preferentially synthesized over others.
The terminator approach, which usually results in DNA copies shorter than 1,000 bases, also helps to prevent chimeric artifacts, he said, because there are fewer chances for self-priming and for cross-hybridization to another template molecule.
The method does not require the DNA to be fragmented — either mechanically or enzymatically — which prevents loss of starting material, Brown said. In most other library prep methods, only about 10 percent of template molecules end up being used in the library, he said. "When you go down to single molecules, you need more precision manipulation of those molecules, or else you get enormous bias, which adds to your cost."
Also, by generating many copies of the template DNA from random start points, rather than making clonal copies of template fragments, the approach minimizes duplicate reads and errors from mistakes early in the amplification process, he said.
Finally, iGenomX's chemistry uses a single enzyme, DNA polymerase, whereas other library prep methods require several enzymes and a number of steps, such as end repair, A-tailing, and ligation of adaptors, some of which are difficult to automate. "By eliminating all of those and reducing the loss of material in each step, we get more information and more uniform information of every molecule that gets sequenced," Brown said.
The library prep process can be adapted to genomes that are high or low in GC content, he said, by changing the composition of the random primers and terminators used.
For the linked read application iGenomX is initially developing, it uses a microdroplet-based approach to dilute and isolate long DNA molecules into separate chambers. So far, it has been using an undisclosed commercial microdroplet system but it is not necessarily wedded to that particular device, Brown said. The library prep chemistry is performed inside the droplets, using primers with barcodes unique for each DNA fragment.
10X Genomics has been developing a system that can be used to partition DNA fragments and prepare them for sequencing, and Brown said that system would work well with iGenomX's approach. "Our chemistry would be ideally suited for the 10X droplet systems," he said. "Our key is what goes on inside the droplet — that's the main differentiator. It would deliver the same results, just with more uniform coverage per molecule."
IGenomX is already working with undisclosed academic partners on generating data and plans to roll out its technology in a stepwise fashion over the next half year or so.
Within the next couple of months, the plan is to release data generated for the NA12878 human genome standard, along with software that integrates into standard pipelines and will allow researchers to replicate the analysis. Those data, the company hopes, will demonstrate the advantages of its approach. "When you reduce your bias and get uniform coverage of the molecule, the data improves," Brown said. "You can start to ask questions you simply could not ask in any other way."
Following this, early testers will be able to submit samples for analysis to be able to look at results for their genomes of interest.
Early this summer, iGenomX plans to deploy the technology, including an unspecified microdroplet generator, at a handful of early-access sites, followed by a full commercial release at the American Society of Human Genetics annual meeting in October.
The company is seeking commercial partnerships but Brown declined to provide further details at this point.