Start-up aims to sequence human genomes for $30, in just a few hours
CAMBRIDGE — The race to sequence genomes faster and cheaper has a new entrant — a start-up spun out of a Harvard University laboratory that focuses on emulsions, or mixtures of liquids like those found in mayonnaise and salad dressings.
Deciphering the first human genome, a massive technical feat, took more than a decade and cost about $3 billion, but the price and time have been dropping rapidly in the 10 years since — down to about $20,000, powered by new technologies that take days or weeks.
The new company, GnuBio, is in the very early stages of its development, but it said last week that its technology could sequence a human genome in hours and for just $30.
The key, said David Weitz, a physics professor at Harvard University, is to make almost infinitesimally tiny droplets, each of which can serve as a “test tube’’ for deciphering a short fragment of DNA.
“It’s the size of the test tube’’ that makes the difference, Weitz said. “We’re just doing tiny volumes . . . We use known methods, proven methods. We just scale it down and use a lot less material and [do it] a lot faster.’’
Scientists say the ability to cheaply and rapidly sequence the genomes of many people will provide an unprecedented tool to seek the genetic causes of common diseases, understand human biology, and allow medicine to be tailored more closely to an individual’s needs.
GnuBio does not yet have a machine to make a $30 genome, and Weitz has not used the technology to sequence a full genome.
A prototype in a crowded Harvard laboratory is being tested as tiny droplets flow through a microscopic tube at the rate of about a million drops per second. The idea is that each drop will act as a test tube where a tiny fragment of DNA will be inserted and undergo a reaction. If the small strand matches a specific sequence, the drop will light up and be read by a camera.
Initially, the company intends to focus not on whole genomes, but on smaller sequencing projects so it can launch its first product quickly.
Weitz said that in December, the company plans to send machines to two clients: Ignite Institute, a nonprofit that focuses on personalized medical research, and the University of Montreal’s Beaulieu-Saucier Pharamacogenemics Centre. The company estimates it could sell an instrument capable of sequencing a full genome for about $50,000.
George Church, a genetics professor at Harvard Medical School who has developed gene-sequencing technology and sits on the scientific advisory boards of more than a dozen biotechnology companies, including GnuBio, said the estimates of the low cost are plausible.
“Most technologies go through a phase of making fairly broad estimates, and I don’t see anything particularly wrong with that estimate,’’ Church said. “A lot of this is about downsizing everything.’’
GnuBio unveiled its technology last week at the Consumer Genetics Show in Boston, at a panel that showcased next-generation sequencing technologies.
“I think what we’re seeing now is the most breathtaking advances . . . that we’ve seen in genomics at least in the time period I’ve been involved in the field,’’ said Eric Green, director of the National Human Genome Research Institute.
“What’s remarkable about it is not so much that there’s one new technology . . . [There are] technologies some that are on the ground, some that are about to land, and some that are out there a couple years away — each one exciting for different reasons.’’
At the conference, other companies that are racing to sequence genetic data more efficiently talked about their programs, too.
“We want to make full-genome sequencing a one-day experiment,’’ said Mostafa Ronaghi, chief technology officer for Illumina.
Even as sequencing has gotten dramatically faster and cheaper, new bottlenecks have emerged.
“Data is spewing out of these instruments so fast and so large in terms of amount . . . it is creating great challenges for how to deal with that data,’’ Green said.
Carolyn Y. Johnson can be reached at firstname.lastname@example.org.