By Julia Karow
As it prepares to roll out its first single-molecule real-time sequencing instrument to early-access customers this summer, Pacific Biosciences is already working on a new, highly multiplexed version of its technology that integrates sequencing reactions, optical detection, and signal processing into a microchip, In Sequence has learned.
PacBio believes that its "version 2" or V2 technology "is going to completely replace any second-gen [sequencing] device," according to CEO Hugh Martin, whereas the current version 1 will likely be used in conjunction with high-throughput short-read instruments, for example in genome assemblies.
V2 will enable the company to multiplex several million zero-mode waveguide reaction wells and to use polymerase enzymes that synthesize DNA at a speed of up to 50 bases per second, resulting in a throughput of more than 100 megabases per second.
Starting in 2014, PacBio plans to launch two FDA-certified systems that are based on the new technology: a high-throughput instrument that will be able to sequence a human genome with several-fold coverage in 15 minutes or less; and a sub-$50,000 point-of-care instrument that will be "the size of a copier" and have a lower throughput.
While the high-throughput system will be geared at core labs and genome centers, the lower-throughput instrument — which will possibly be portable — will be aimed at small clinical labs "or even doctors' offices" and is designed to be "a delivery vehicle for some of the applications developed on V1 around diagnostics," Martin said.
Prior to that, in 2011 and 2012, the company plans to improve the performance of its "version 1" system by increasing the read length, polymerase speed, and number of productive ZMWs; lowering the cost per base; upgrading software; and adding new applications such as methylation sequencing, direct RNA sequencing, and protein-translation analysis. None of these improvements will require hardware changes, according to Martin.
The first commercial V1 system, scheduled for rollout later this year (see related story, this issue), will likely be used mostly in research. While "you are not going to see that in a doctor's office," Martin said, "you will see large pharma companies or diagnostic companies developing applications that are going to be deployed, eventually, on much higher-volume, lower-cost technology."
To that end, the company is developing its V2 technology, an integrated "detection core" the size of a shoe box or smaller that can be used as part of different types of commercial instruments geared at different markets.
The detection core, developed internally at PacBio in collaboration with partners from the semiconductor industry, will integrate chemical reactions with "electro-optic functions" into a micro-scale device with large numbers of so-called "Optode cells."
Each Optode cell consists of a ZMW connected to a microfluidic channel for reagent delivery, a layer underneath to illuminate the bottom of the ZMWs, and a "smart" pixel detector below. The signal from each cell "will be contained completely within the detector," Martin said, which has filtering and base-calling circuitry. Optodes operate individually and do not transmit their signals in synchrony.
The key difference from the V1 system is that version 2 will lack a camera, instead assigning each ZMW its own detector. PacBio will be able to pack up to several million Optode cells into an array, enabling it to generate a total of 100 megabases of sequence data per second, Martin said.
[ pagebreak ]
The DNA bases will be called directly on the chip. "If you are doing 100 million bases per second, you could not get that imaging data off of the chip, there is no technology that would support that," Martin said, but once the bases are called, the data can be transferred.
The arrays, he said, can be manufactured with commercially available MEMS and CMOS technology. By using an integrated fluid control system, the company expects to "dramatically lower" sample volume and reagent costs.
One of the "key areas" of the technology, developed internally and "highly proprietary" to the company, is a structure underneath the illumination layer that spatially separates the four frequencies of the light from the four-color nucleotide labels.
The scalable design of the Optode cells allows the firm to build instruments with different throughput. "We have tremendous flexibility in where we position these products, and all we have to do is stamp out as many Optodes as we need for each individual product," Martin said.
It will also enable PacBio to build more instruments more quickly, and to enter new markets beyond academic research. "Volume will be dramatically different from before," Martin said.
Development work on the detection core is headed by PacBio's chief technology officer, Steve Turner. The team is currently running experiments to make a number of "implementation choices," Martin said. The plan is to finalize the detection core within the next couple of years, and then to start developing "the actual real system," he added.
Early access for the V2 system might start in 2013, prior to the planned commercial launch in 2014. The company has not yet decided which of the two instruments — the high-throughput or the low-throughput version — it wants to launch first, or whether they will be launched simultaneously. "We need to see how the market evolves," Martin said.
PacBio is not the only group working on a chip device that integrates sequencing reactions and optical detection: a team at the Industrial Technology Research Institute in Hsinchu, Taiwan, said recently that it is working on similar technology (see In Sequence 12/22/2009).