By Laurel Oldach
September 4, 2014 | The year of the thousand-dollar
genome just got a little more interesting as Edico Genome, which makes a data
processor called the DRAGEN for next generation sequencing (NGS) data, made its
first sale to Sequenom, Inc.
its designers at Edico, the DRAGEN, short for Dynamic Read Analysis of GENomes,
reduces the time to map a complete human genome to reference by sixty-fold. At present
the mapping, sorting, deduplication and variant calling that are required to
turn 200 GB of raw massively parallel sequences into an interpretable genome
takes, by Edico’s estimate, about twenty hours on an industry standard server.
The marketing team at Edico Genome likes to compare the task to piecing
together a telephone book that’s been put through an industrial shredder. This poses
a considerable threat of backlog given that a high-throughput sequencer like Illumina’s
HiSeq X Ten can turn out a comparable amount of sequence every thirty minutes.
Sequenom, a leader in noninvasive prenatal testing (NIPT) with about 40% market
share, faster analysis at lower cost is of the essence. In contrast to more
established methods of prenatal testing, which draw DNA directly from the placenta
and pose some risk of miscarriage, NIPT uses maternal blood, which contains
short fragments of circulating fetal DNA. If mapping a single genome is like
piecing together a telephone book, then NIPT is like putting together two telephone books, which share at
least half of the constituent phone numbers, when there are a dozen copies of
Mom’s phone book for every one of Baby’s. Nonetheless, Sequenom has found ways
to piece together a low-coverage sequence of the fetal genome, enough to identify
aneuploidy as well as some microdeletions. The company advertises diagnostic
capability for a panel of fourteen genetic disorders, from Down syndrome (trisomy
21) to Prader-Willi/Angelman syndrome (deletion in 15q).
flagship MaterniT21 Plus test currently has a five-day turnaround time from
receipt of a maternal blood sample. To speed its data processing, Sequenom
plans to use the mapping, sorting and deduplication algorithms built into the
DRAGEN; although the DRAGEN pipeline also includes an algorithm for variant
calling, Sequenom will continue to use its own proprietary software at this
step. Neither party will comment specifically on the time or cost savings Sequenom
gains from adding the DRAGEN to its pipeline, but both are reportedly significant.
The sale was
sealed after Sequenom took part in an early iteration of Edico Genome’s availability
program, which allowed participants to test the DRAGEN against their existing pipelines.
Both parties are close-mouthed about the specifics of these trial runs, citing
an upcoming poster to be presented at the American Society for Human Genetics
annual meeting in October. “We’re very excited specifically because Sequenom is
one of the leaders in clinical genomics. It’s also a big endorsement of our
product,” says Edico CEO Pieter van Rooyen, adding, “we’ve got a fairly significant number of
other customers involved in the early access program already.”
increase in processing speed is achieved by an application-specific integrated
circuit, hardware that sacrifices the flexibility of a general-purpose server
for efficiency in the specific task of arranging segments of a genome. “We
have the cheapest solution because we come to it from a consumer electronic
perspective,” says van Rooyen, an electrical engineer by training. The concept
of using reconfigurable hardware (the DRAGEN uses a field-programmable gate
array, or FPGA) to speed up computationally intensive alignment is not new; however,
engineering at Edico Genome has enabled them to shrink the analytical platform
to just one FPGA compared to the multiple parallel FPGA chips required for tools
such as the GSC XpressAlign. The DRAGEN is an external chip that can be
integrated into a sequencer or server, with what Edico describes as a
push-button workflow once installed.
It has been
an eventful summer for Edico, which raised $10 million in venture capital in July.
For now, van Rooyen says, the company will focus on scaling up, courting new customers
in subfields of clinical genomics such as tumor profiling, rare disease
diagnosis, and pharmacogenomics. “We’ve got awesome technology, and we’re happy
to get it out into the hands of customers,” says van Rooyen. “We would like to
make sequencing cheap and affordable, and get to the bottom of all these
different diseases… as was promised [with the initial draft of the human
genome] fourteen years ago.”