In their quests to grab a bigger portion of the sequencing market, the three leading vendors of post-Sanger sequencing platforms continue to push the performance of their instruments.
This week, Applied Biosystems is announcing the second update for its SOLiD platform this year, following the 2.0 upgrade in May. SOLiD 3, which will be available to customers early next year, increases throughput and read length and decreases cost per base and run time. It also promises to simplify workflows and decrease DNA input requirements, ABI said.
Also this week, Roche’s 454 Life Sciences is launching its Titanium reagents and plates for its Genome Sequencer FLX. The new consumables, which have been available to early-access customers for several months, increase the system’s read length and throughput, and reduce the cost per base.
Meanwhile, Illumina said at a conference last week that owing to a new reagent that is part of its sequencing chemistry, along with software improvements, read lengths and output of its Genome Analyzer II will go up by the end of the year, and run time and cost per base will decrease.
ABI: SOLiD 3
ABI’s president and chief operating officer Mark Stevenson said during a presentation at the UBS Global Life Sciences conference in New York last week that the company plans to announce an update for its SOLiD system — version 3 — this week.
The new version, he said, will enable customers to generate 20 gigabases of data per run, though internally, ABI researchers have already achieved a higher throughput.
The current version of the instrument, SOLiD 2.0, which ABI launched in May (see In Sequence 4/29/2008), has a nominal throughput of 6 gigabases per run, although some customers have achieved almost 15 gigabases of data. ABI has internally obtained more than 20 gigabases from a single run (see In Sequence 7/29/2008).
The upgrade to version 3, which will be available to customers early next year, will involve changes in hardware, reagents, protocols, and software, ABI officials told In Sequence last week.
The increased throughput is mainly achieved by packing more beads on the slides, according to Shaf Yousaf, president of ABI’s molecular and cell biology genomic analysis division.
The system’s read length for fragment runs is expected to increase from 35 to 50 base pairs, and for mate pair runs from 2x25 to 2x50 base pairs. Internally, ABI researchers “have demonstrated” read length of 75 base pairs for fragment runs and plan to “roll that out as quickly as it is ready,” according to Jason Liu, director of SOLiD commercial operations.
The company has also reduced the run time of the instrument, bringing the length of a 50-base-pair fragment run down 40 percent to about three and a half days, according to Liu.
The main hardware change of SOLiD 3 is the addition of a refrigeration unit to chill reagents during the run. This will allow the system to run unattended for 72 hours, according to Yousaf. Up until now, users had to change reagents once a day during a run.
ABI also plans to “streamline” workflows for emulsion PCR, which is part of the sample-preparation process for the SOLiD system. Liu said the company has “many R&D projects ongoing” to simplify the workflow and to “improve user experience.” ABI is also “working with an automation supplier to explore the potential for automating certain steps,” he said. Improvements in the ePCR workflow will definitely be part of version 3, but could be released earlier than that, he said.
In order to improve other workflows for the system besides ePCR, ABI plans to automate certain protocols and to develop a “master mix solution to cut down the number of steps of tube mixing the customer has to do,” Liu said.
According to Liu, the company also has a “very active R&D program” for reducing the amount of input DNA needed.
“We do think we have now displaced Sanger sequencing for a lot of sequencing applications.”
In order to cater to customers that do not need a throughput of 20 gigabases for a single experiment, ABI is working on increasing the multiplexing capability of its system. Along with the SOLiD 3 launch, it plans to release barcodes that will enable customers to run up to 256 samples in parallel. Those barcodes are already in early-access testing as part of a small RNA analysis kit, Yousaf said.
Regarding new application kits for the SOLiD system, he said that ABI plans to release “in the near future” a whole-transcriptome analysis kit that is already available to early access customers.
The company is also working “with different vendors” — among them Agilent Technologies — on methods for multiplexed targeted resequencing. Several companies, including Roche/NimbleGen, Agilent Technologies, RainDance Technologies, and Febit, are working on such sample-prep methods.
SOLiD 3 will be available to new and existing customers in the first quarter of 2009. The cost of upgrading an existing instrument will be “minimum” and the list price of the new instrument will be “not dissimilar to the current system,” according to Yousaf.
The new version will eventually enable ABI to “get to the $10,000 genome,” according to Yousaf, defined as reagent costs a typical customer would incur for generating 12-fold coverage of a human genome, or 36 gigabases. This would translate to a cost of approximately $280 per gigabase of sequence data.
Throughout next year, ABI plans to increase the throughput of SOLiD 3 even more. For example, further optimization of the bead packing could bring the output to 50 to 60 gigabases per run, the “theoretical limit,” Yousaf said. Internally, ABI has already achieved runs “at the 40-gigabase level,” he added. Stevenson said during his talk that ABI’s goal is to get up to 50 gigabases per run sometime in 2009.
In order to go beyond that, the company will have to array the beads on the slide, Stevenson said. He did not provide further details on the technology the company plans to use. However, earlier this year at the Association for Biomolecular Resource Facilities annual conference, Xiaohua Huang, a professor of bioengineering at the University of California, San Diego, presented a patent-pending technology for generating ordered arrays of “DNA particles.”
At the time, Huang said that at least one company was interested in licensing the technology for use with its next-generation sequencing platform (see In Sequence 2/19/2008).
Roche’s 454 Life Sciences this week is officially launching its new Titanium reagents and plates for its Genome Sequencer FLX.
The company talked about the upcoming changes, which have been available to early-access customers for several months, at a meeting in Marco Island, Fla., in February (see In Sequence 2/19/2008). It has also mentioned the planned upgrade to customers since at least early 2007 (see In Sequence 1/30/2007). With this week’s launch, the upgrade, which involves no hardware changes, is now available to all customers.
Titanium — a series of reagent kits and consumables — quintuples the output per run on the GS FLX from 100 megabases to between 400 to 600 megabases, according to Jason Affourtit, director of advanced technologies at 454. The average read length has increased from 200 to 300 base pairs to 350 to 400 base pairs, though many reads are 500 bases pairs in length or longer, said Affourtit during a presentation on the technology at the Cambridge Healthtech Institute Exploring Next-Generation Sequencing meeting in Providence, RI, last week.
Besides read length, the greater throughput is made possible by a denser picotiter plate with 3.6 million wells, where a metal coating prevents crosstalk between neighboring wells.
Affourtit also mentioned that 454 will soon release a new protocol for generating paired end libraries with 20-kilobase inserts and 2x150 base-pair reads, and has updated its existing protocol for paired-end libraries with 3-kilobase inserts. Another protocol using medium-sized inserts is in progress, he said.
Since 454 will not charge more for its reagent kits, the cost per base is effectively decreasing by 80 percent, 454 CEO Chris McLeod told In Sequence this week. In fact, reagent kits will now include a polymerase mix at no additional cost that customers had to purchase separately before.
“We do think we have now displaced Sanger sequencing for a lot of sequencing applications,” McLeod said. While the previous read length of 200 to 300 base pairs meant that customers were still using the Sanger technology for certain applications, this has now changed, he said, and early-access Titanium users have started replacing Sanger with 454. “It’s a real milestone for us.”
Affourtit mentioned several recent examples of applications for the Titanium upgrade. One of its customers, for example, has sequenced two microbial genomes in a single run on a two-part picotiter plate, obtaining 660 megabases of high-quality data and average read lengths above 400 base pairs.
In-house, 454 researchers have been working on a finished de novo assembly of the bacterial Campylobacter jejuni using only 454 reads, generating 38-fold coverage of the genome.
The company and its customers have also used the technology to sequence mid-sized eukaryotic genomes de novo, such as Arabidopsis, Drosophila, and two strains of apple, he said.
Affourtit also mentioned that researchers at Baylor College of Medicine have used the Titanium chemistry to sequence a 3-megabase region of the human genome as a pilot project for the 1,000 Genomes Project, obtaining better results with Titanium than with the old GS FLX chemistry.
Commenting on future improvements, McLeod said that the last two upgrades — made over the last three years — each resulted in a five-fold improvement in throughput, both through improvements in read length and density, and “certainly, you can extrapolate that another five-fold improvement in 18 to 24 months would be feasible.”
Illumina: New Reagent, Longer Reads
Like its competitors, Illumina has been working on increasing the performance of its system, the Genome Analyzer II.
During a talk at the CHI conference in Providence last week, Gary Schroth, senior director at Illumina in Hayward, Calif., told the audience that by the end of this year, the company expects to increase the number of reads, read length, and accuracy; and to decrease the sequencing cycle time for the GAII, the update to its original system that it began shipping in February (see In Sequence 4/29/2008).
He said that read lengths for paired-end reads will increase to 2x75 base pairs. Users will be able to add additional cycles, though, to achieve a read length of 2x100 base pairs.
What made this increase in read length possible was a “significant breakthrough in one reagent,” according to Schroth.
In the spring, he said, Illumina attempted to increase the read length beyond 50 base pairs internally and found towards the end of the read, T bases were overcalled and the error rate increased significantly. The company solved that problem by devising a new reagent that is used during the deblocking step of the sequencing cycle, he reported.
The new chemistry, which “can go even beyond 100 bases,” is in beta-testing with customers, he said, and yields an improved sequencing error rate.
The system’s output, he said, is expected to increase to more than 15 gigabases for a paired-end run by the end of the year. Using additional sequencing cycles, customers can even go beyond 20 gigabases, according to Schroth.
As a result, the reagent cost per gigabase — based on list prices — will decrease to $450 for a 2x75 base-pair run, and to $340 for a 2x100 base-pair run.
What enables this increase in throughput, besides longer reads, is an improvement in certain algorithms that allow the system to analyze 30 percent more DNA clusters than before, he said. The chemistry also uses a “new enzyme,” he said.
Along with the increased throughput, the run time for the GAII will increase to about seven days for a 2x50 base pair paired-end run. Each sequencing cycle, however — during which one base gets incorporated into the growing DNA strand — will be shorter than before, according to Schroth, owing to an optimized chemistry and new protocols.
In terms of new application kits for its system, Illumina plans to launch an mRNA-sequencing kit, which is currently in beta-testing, next month, Schroth said. Illumina developed the method in collaboration with Barbara Wold’s group at Caltech, which published an RNA-Seq-based transcriptome study this spring.
Schroth mentioned that Illumina is now applying its 100-base pair fragment reads in RNA-Seq studies to map exon-exon relationships.
By the end of the year, Illumina also plans to launch a new software package, called GenomeStudio, which will contain several data analysis programs for different sequencing applications, such as ChIP-Seq, mRNA-Seq, and DNA sequencing for SNP calling.