Scientists from Northwestern and Stanford Universities have improved Sanger sequencing by combining two types of polymers to accelerate the electrophoretic separation step in the channel of a microchip.
The new method, which appeared in last week’s PNAS, could help increase the throughput and decrease the cost of Sanger sequencing, both in microchip-based and other formats.
“It’s quite a good deal faster than the current electrophoresis technology, and faster than what anybody had ever demonstrated to be possible on a chip before,” said Annelise Barron, an associate professor of bioengineering at Stanford University, who led the research. She recently moved to Stanford from Northwestern University.
According to the PNAS paper, conventional capillary-based sequencers typically take about one to two hours to generate 500 to 700 bases of DNA sequence. Those instruments use gel-filled capillaries that are about 50 to 60 centimeters in length.
But earlier studies have shown that using channels in microfluidic chips instead of capillaries can increase the separation speed. With these findings as a backdrop, the Northwestern researchers used a microfluidic chip to make the step three times faster than earlier research had shown. Specifically, their method relied on a comparatively shorter microfluidics channel to sequence up to 600 bases in 6.5 minutes in a single channel.
“What distinguishes [our microchip] not only from the capillaries, but also from any microchip that’s been used for sequencing in the past, is that the total path length is only 7.5 centimeters,” Barron said. “Probably not many people would have thought that was enough separation distance to get 600 bases.”
The ability to use such a short channel lies in the matrix that filled the channel, in this case a combination of a wall-coating polymer, pHEA, and a very high molecular weight separation matrix, pDMA.
Using this mix, “all of a sudden, the extent of band broadening in the microfluidic chip is greatly reduced,” Barron said. “This greatly narrowed peak width allows us to get, essentially, a full read in those 7.5 centimeters.”
The researchers are now planning to use an improved matrix or a slightly longer channel to try to increase the read length beyond 600 bases. “We would like to get it out to 800 or 900 bases in something less than eight or nine minutes, and I think we can do that,” Barron said.
In order to integrate the technology into a commercial sequencer, it would also need to be parallelized, but Barron said her group is not planning to work on this. “What we have demonstrated at this time is the separation in a single channel,” Barron said, adding that she is not an instrument developer.
The technology is available for licensing, according to Barron, and “we are not, at this time, under contract with anybody.”
Barron’s lab has collaborated in the past with Microchip Biotechnologies of Dublin, Calif., which has been developing a microchip-based sequencing system that integrates sample preparation, sequencing reactions, sample cleanup, and electrophoretic separation in a single device (see GenomeWeb Daily News, In Sequence’s sister publication, 4/24/2006).
“If you want to compete against the throughput of the other technologies that have recently come into the limelight, you really need thousands upon thousands of channels, not just 96.”
In 2004, MBI won a $6.1 million grant from the National Human Genome Research Institute to develop the platform in collaboration with Barron; Richard Mathies, a professor at the University of California, Berkeley, and a co-founder of Microchip Bio; and Jingyue Ju, a professor at Columbia University.
In theory, accelerating the separation speed could decrease the cost of Sanger sequencing by increasing the throughput and, thus, improving instrument amortization.
“The capitalization for the equipment is one of the major expenses in running genome centers,” Barron said. “If your throughput is 10 times higher because the separation is 10 times faster, that’s clearly going to impact that cost greatly.”
But in order to be competitive with next-generation sequencing technologies, speeding up existing Sanger sequencers, which run 96 capillaries in parallel, or replacing them with microchip sequencers will not be enough. “If you want to compete against the throughput of the other technologies that have recently come into the limelight, you really need 1,000s upon 1,000s of channels, not just 96,” Barron said. “That’s challenging.”
But because of its long reads, she said, Sanger sequencing is still attractive for applications such as de novo genome sequencing. “900-base reads, you can’t beat it,” she said.
One company that is betting on low-cost, ultra-high-throughput Sanger sequencing is Genome Corp., a Providence, RI-based based startup, whose scientific advisory board Barron recently joined (see Paired Ends in this issue).
According to Kevin Ulmer, co-founder and CSO of Genome Corp., Barron’s approach may help reduce the cost of Sanger sequencing both by increasing the throughput and by shrinking the physical area that is undergoing electrophoresis.
By using shorter channels, “the total area required to do the separation gets reduced, so the cost gets reduced,” he said. “We see [this improvement] as a key step in the further reduction of one of the components of the overall Sanger sequencing cost,” Ulmer told In Sequence this week.
However, Genome Corp. is not planning to employ the microchip format that Barron and her colleagues used in their study because it does not scale enough “in terms of achieving the cost and throughput that we are targeting” for human genome sequencing projects, he said.
Genome Corp., which recently raised $250,000 from the Slater Technology Fund (see Short Reads in this issue), a state-backed venture-capital fund that already invested $250,000 in the company last year, has not disclosed how exactly it wants to soup up Sanger sequencing.
Last year, Ulmer told In Sequence that he plans to build a sequencing “factory” that transforms electrophoresis and imaging into “an industrial-scale continuous process” and increases the read length to beyond 1,200 base pairs.