By Julia Karow
This article was originally published June 2.
GnuBio, a newcomer among developers of DNA sequencing technology, is working towards reading a human genome for $30 in reagent costs, drop by drop.
Co-founded by David Weitz, a professor of physics and applied physics at Harvard University who is also a founder of microdroplet company RainDance Technologies, the startup is planning to develop an inexpensive, fast, and scalable microfluidics-based sequencer where reactions take place in picoliter-sized drops.
Besides selling sequencing systems, the firm wants to build a large-scale open-access database of anonymized genotype and phenotype data from service work and charge license fees for biomarkers discovered from the data.
The company, whose name alludes to the freely available GNU/Linux operating system, was founded around the end of last year by Weitz; Michael Weiner, chief scientific officer of antibody firm Affomix and a former vice president of 454 Life Sciences and RainDance Technologies; and John Boyce, former head of business development at Helicos BioSciences and co-founder of consulting firm Delphi Bio. It continues GnuSeq, which Weitz and Weiner founded less than two years ago to develop their sequencing idea. Harvard professor George Church and Stanford genetics chair Mike Snyder are among the firm's scientific advisors.
According to Boyce, the firm is currently comparing funding options from several venture capital firms for undisclosed amounts of financing and has applied for a Small Business Innovation Research grant from the National Institutes of Health. Once it has funding in hand, GnuBio plans to move into its own space and to hire a team of engineers in order to develop a commercial sequencing platform, with the goal to ship beta systems to early-access customers before the end of the year.
The sequencing technology, which uses amplified DNA template and an optical readout method, builds on almost a decade of microfluidics work in the Weitz lab, which is currently working on a proof of concept. In principle, it can use a variety of existing sequencing assays. "Any of the assays that are available now you can translate into this drop technology," Weitz told In Sequence. "The only innovation here is putting everything into drops where the volume is very small," which saves reagent costs.
Weitz shared some details of the technology with In Sequence last week and presented it at the Consumer Genetics Conference in Boston.
"What we have done in my lab is learned how to control and manipulate these drops, so we can fill them, let the reaction take place, and read them at rates of something like a million drops a second," he said.
Genomic DNA sheared to a size of about a kilobase is first PCR-amplified in nanoliter-sized drops and then transferred to barcoded picoliter-sized drops that have a diameter of about 10 microns. A single drop maker — a channel the size of a human hair that can easily be multiplexed — can generate about 10,000 drops per second. After adding labeled oligonucleotides, a camera reads out both the barcode and the sequencing signal from each drop at a speed of a million drops per second. Most likely, the system will use fluorescent labels, according to Weitz. A phenomenon called "spot glare" from each drop — akin to the glare from a window — helps the camera software recognize and track individual drops.
Right now, Weitz' team uses an assay that relies on a combination of hybridization and ligation and requires preformed libraries of oligonucleotides. That is "a perfectly fine assay" to sequence small amounts of DNA, he said, for example several genes, and the plan is to sequence templates of about 1,000 bases of DNA. "But as soon as we have a larger engineering team, we will go to a variety of different assays," for which he has already filed patents.
Sample preparation for the system "is not going to be more involved than what everybody is doing now," Weitz said, adding that it is "not something we have been able to work hard on" so far.
For proof of concept, the scientists plan to sequence an oligonucleotide in order "to show that we can sequence something, show where the pitfalls are, where the things are that have to be optimized," Weitz said. The goal is to have a paper submitted by the end of the year.
[ pagebreak ]
The technology has the potential to be used for both small-scale targeted sequencing and whole-genome sequencing. According to Weitz' estimates, it will cost about $30 in reagents and take approximately 10 hours to sequence a human genome at 30-fold coverage. One hundred genes could be sequenced at 30-fold coverage within 15 seconds at a cost of 2 cents. "It's totally scalable; it's not something where you have to run huge amounts of sample to take advantage of the speed and the cost. You can run small amounts and large amounts," he said.
A team of company engineers will focus on building a commercial instrument, and the plan is to deliver beta instruments, which will not be able to sequence entire genomes yet, to early-access customers by mid-December.
GnuBio already has commitments for instrument placements from the Beaulieu-Saucier Pharmacogenomics Centre at the University of Montreal and the Ignite Institute. The commercial instrument will cost on the order of $45,000, according to Boyce. Weitz said the company might decide to develop several versions, for example an inexpensive machine for a doctor's office and a research-grade instrument for whole-genome sequencing.
While the company's primary goal is to develop sequencing instruments for sale, he said, a future plan is to harness the machines in house for inexpensive large-scale genotyping services in exchange for anonymized genotype and phenotype data that will be entered into a large database. That database will be open access for qualifying researchers, though the company would collect license fees for biomarkers discovered from the data. GnuBio has already signed letters of intent from several undisclosed organizations interested in genotyping large numbers of samples, Weitz said.