Skip to main content
Premium Trial:

Request an Annual Quote

The Next-Gen Showdown


Next-gen sequencing is a constantly moving target. Within the three years since 454 offered the first improvement on Sanger sequencing, the market has grown quickly, as have the options for platforms. Researchers no longer need loads of grant funding to afford a new sequencer, and they're finding that instruments have never been more affordable or enabled such a broad range of research.

With Roche's 454 platform, Illumina's Genome Analyzer, and ABI's SOLiD system covering both long and short read technologies, many scientists have realized the value of diversification — and that each machine may well have its niche of applications it runs best. In this article, we'll take a look at how customers are using the machines everywhere from sequencing centers to core and individual labs, as well as check in with some of the newer, third-generation technologies that are promising to sweep Sanger into the history books. First, a look at each platform that's readily available on the market and what customers have to say.

Roche's 454 Bill Farmerie, associate director of the Interdisciplinary Center for Biotechnology Research at the University of Florida, moved from Sanger to next-gen technology before most, buying a 454 in May 2005. "We thought, 'This is something we need to pay really close attention to.' ... The lure of being able to produce tens of millions of bases of sequence information and do it in a very compressed time scale, with far less labor and far less infrastructure, was attractive." In August 2005, his 454 GS 20 arrived. He upgraded to the FLX model in March 2007. Farmerie says the platform gets the most use by agricultural scientists doing de novo transcriptome sequencing.

Over the next several months, 454 plans to roll out its new Titanium chemistry and software upgrade, which will allow the GS FLX to produce up to 500 megabases of data per run with an average read length of about 500 bases. Reagent costs will drop as much as 15 percent, and there will be no charge for the upgrade, says Michael Egholm, VP of research and development at 454. Early access labs for Titanium include Baylor's genome center, the Broad, and Washington University. The main advantage to the upgrade is the significantly longer reads, which will enable the FLX to cross over into new applications, such as de novo sequencing of mammalian genomes, transcriptome sequencing for splice variants, and metagenomics, says Egholm. "We're basically crossing [the] threshold where Sanger sequencing used to be."

Illumina's GA

Yale's Shrikant Mane owns a Genome Analyzer, which he bought about a year ago. At the time, he didn't think the SOLiD was mature enough, and "we were not very impressed with 454 because [of] cost." He said that while Illumina's customer service was helpful, there were the inherent challenges with any new technology and a lot of down time. Since then, he's had "many successful results," using it largely for transcriptome analysis. Once he overcomes problematic capture mechanics, he says he'll start using it for resequencing. Mane says he's in the market for another instrument and this time may buy a SOLiD, which he says will allow him to pool samples, making it more cost effective for applications like digital gene expression.

The Stowers Institute's Anoja Perera also runs a Genome Analyzer for gene expression, ChIP-seq, and genome sequencing, and says that of all the pipelines, ChIP-seq is the simplest. "Right now we are not doing that much de novo sequencing," she says of her choice over 454. "At the time we purchased the Illumina unit, there wasn't much data out about the SOLiD system, so it was an easy decision."

Perera bought her Illumina in September 2007, and when she received it after several months' delay due to a backlog of orders, it took only three days to get trained to use it. Illumina, she says, offers training on generating runs and analyzing the sequence data; for sample prep, though, you're on your own.

Charlie Nicolet, director of the UC Davis Genome Center DNA technology and expression analysis core, owns two Illumina machines, one upgraded to the Genome Analyzer II paired end module. On it, he performs plant genome assembly, RNA-seq, ChIP-seq, and mutation discovery. His reasoning for buying the Solexa over 454 or SOLiD was that he "wanted the most flexibility. If we were only going to do genome assembly, then we might have gone for 454," Nicolet says. "But for the variety of applications that we saw at UC Davis, we felt the Solexa would be the most versatile."

He bought the first Genome Analyzer in August 2007 and upgraded it in June 2008. He bought the second in October 2007. From order to delivery, the first machine took about two to three months and the second, about two months. Getting them up and running was "pretty fast."

At the University of Washington, Jay Shendure runs a single Illumina Genome Analyzer, which he'll soon upgrade. Shendure is working on developing new applications for the instrument, mostly focused around genomic capture for resequencing. His lab is also using it for de novo sequencing and RNA-seq. He received the machine in October 2007 and he says it took an inexperienced lab tech a couple of weeks to get the machine running to a 95 percent success rate. "It's required very little of my time in terms of the actual instrument," Shendure says.

In deciding to go with Illumina, PCR type was important to Shendure. "In my experience, emulsion PCR is not a very pleasant thing to do," he says. "Cluster PCR certainly has appeal from a technical perspective."

Shendure says that Illumina offers pretty good software to build sequencing reads, and that the community has grown a lot more knowledgeable in this area. For the most part, people are using vendor-provided software, but mapping is a different story. "Software that does the mapping is one of the key needs," he says; his lab routinely takes advantage of available programs such as Sanger's Maq and Crossmatch. Another is Alta-Cyclic, a base-calling software developed at Cold Spring Harbor Laboratory.

Illumina upgraded its Genome Analyzer system earlier this year from the "GA classic to GA II," says Adam Lowe, Illumina's associate director of product marketing for life sciences. Some of the enhancements included making it faster, adding temperature controls, and increasing throughput. While most of the installed base has migrated to the GA II, Lowe says, Illumina's working on scaling the instrument to increase its throughput by the end of this year, adding new changes to the chemistry and software. First release test sites have increased throughput to 10 GB per run and 50-base read lengths that might be stretched as far as 75 bases. By the end of the year, the team hopes to scale the machine even further, to 20 GB per run and a standard read length of 75 bases, with a possible stretch to 100 bases.

Lowe says the system has seen a lot of use for human resequencing — both whole human genome resequencing and targeted resequencing – as well as for mRNA sequencing and ChIP-seq. "What has been challenging for some customers is the massive amounts of data that the system generates are unlike anything they've seen before," he says. To that end, Illumina has created an inline data analysis system for data reduction. The goal is to have the system output bases with quality scores, not images, "so customers will have the option of storing all the data that they want to or just storing base calls," Lowe says.

New applications on the horizon include: an mRNA-seq kit to be offered this month; early access kits for long insert (up to 10 KB) paired end reads; sample indexing for multiple samples; and the launch of Genome Studio, software based on Bead Studio that will integrate sequencing with microarray visualization data.


Florida's Farmerie also runs a SOLiD, which he ordered in June 2007 and received in February 2008. "We could see that diversifying was important," he says. "It was obvious, even before we had the FLX, we needed to be thinking about the short read technologies." The two things that influenced him when he made the decision to go with SOLiD over Illumina was the raw accuracy of reads and the versatility of the mate pairs, he says. In his lab, SOLiD has opened doors for clinical researchers, with analyzing sequence variation, digital gene expression profiling, and ChIP-seq being the strongest apps for these newer, short-read instruments.

Ron Hart, of the W.M. Keck Center for Collaborative Neuroscience at Rutgers University, runs a SOLiD. He chose it for "the low error rate, the high density, and the development pathway," he says. "Having worked on data sets of small RNA reads from other platforms, I knew that error rates were a major concern when trying to align such short sequences to the genome." For the most part, he uses it to sequence small RNA libraries, and has recently moved to more multiplexed samples. He's just begun using it for ChIP-seq. "In both methods, the short reads of the SOLiD are appropriate, and the apparent low error rate is helpful," he says. He adds that the best apps for SOLiD will probably be gene expression analysis and sequencing mRNA and miRNA.

"SOLiD is clearly a work in progress," Hart says, citing challenges of getting the machine set up, higher than expected prices per sample, complicated library preparation, and dealing with data. "Data analysis requires much more commitment in hardware and personnel than we expected. But … we're learning to adapt."

ABI, which launched its instrument a year ago, recently upgraded its system to double throughput from 3 GB to 6 GB, simplify workflows, and decrease run time. The new system is also capable of multiplexing up to 256 samples.

ABI says that early next year it plans to launch a new platform, called the SOLiD HT System. This "third-generation sequencing platform … will give us a clear path to the $10,000 genome, and potentially even beyond," says Jason Liu, director of SOLiD commercial operations.

One big difference is that the machine will be able to run unattended. "The most distinguishing thing is that this is going to go up to 20 GB minimum on mappable, and that it's going to be supporting 35-mers and 50-mers," says Kevin McKernan, senior director of scientific operations at ABI. "[It] has a clear path to 50 or 60 GB runs," he says, pushing the throughput to 500 million reads per run. Internally, ABI currently obtains read lengths of 50 bases in paired-end sequencing and 75 bases in fragment-based sequencing.

The HT is expected to be widely used in areas such as whole genome sequencing, targeted resequencing, and whole transcriptome sequencing, eventually moving into cancer genome research and personalized medicine. "We've got longer reads, and they're more accurate," McKernan says of the new system. The raw accuracy of the reads will be improved to the point that "de novo sequencing applications will open up with this platform." The system is available as a hardware upgrade and will not cost more than the current SOLiD system.

Multiple platforms

Some labs are figuring out creative ways to maximize the potential of both long- and short-read platforms by using a combination of instruments. Over at the Joint Genome Institute, Feng Chen runs four 454 and four Illumina sequencers, using each "for very different purposes." Because of the read length and quality of the 454, Chen uses it for whole genome shotgun sequencing of small microbes, medium-sized genomes like fungi, and large plant genomes. JGI scientists also use 454 to perform cDNA sequencing for genome annotation. Chen relies on Illumina for various resequencing projects, such as detecting SNPs and genomic rearrangements. His team also uses Illumina data to supplement 454 in de novo sequencing. "Each instrument [tends] to make mistakes in different places; that's why we combine them," he says. He also uses Illumina for ChIP-seq and RNA-seq.

Chen says the 454 and Illumina platforms have come a long way in their analysis tools. "We use [454's] Newbler by default, but we also use others to double check if everything is correct," he says. For the shorter Illumina reads, "we are still looking for a better alignment tool for SNP detection. Right now we use Maq."

Eventually, JGI aims to phase out its Sanger sequencing, Chen says. "We want to eventually use 454's Titanium platform to replace our Sanger sequencing maybe in two years," he says. He's considering purchasing a SOLiD in six to nine months, and maybe a machine from Pacific Biosciences — "if they keep their promises," he says.

Michael Dorschner at the University of Washington's high-throughput genomics unit has two Illumina Genome Analyzers and an ABI SOLiD. He uses Illumina mostly for tag-based assays as well as RNA-seq and ChIP-seq. The SOLiD comes in handy for resequencing applications like mutation discovery. "For applications where you're looking for variants, it's important that the data quality is extraordinarily high, and I think the accuracy is probably better on the SOLiD system," he says. He opted away from 454, mostly because his work requires only short sequences.

He got the first Genome Analyzer in April 2007, the second in August 2007, and the SOLiD in March of this year. He says it took three to four months to receive both Illumina machines, and each was up and running within a month. It took three months before the SOLiD was delivered, and about two months to get it up and running "mainly because we just haven't been putting as much effort into it because we have the two other sequencers going full blast at the moment."

Dorschner says that sample prep has been a bit easier with Illumina than with ABI because of the emulsion PCR used on the SOLiD. "I think it's probably a little bit smoother for the Illumina platform, but I think as we get more experienced with the ABI platform, it will smooth it out as well," he adds.

Rod Wing, director of the Arizona Genomics Institute, optimizes the strengths of both the 454 and Illumina systems for sequencing plant genomes ranging in size from 400 MB to 15 GB. Since his lab is interested in de novo sequencing, members "either use one exclusively or use a blend to get a nice, high-quality sequence," he says. His team's BAC-by-BAC approach, pioneered for the International Rice Genome Sequencing project, allows scientists to sequence chunks — say, a chromosome arm — and then put them back together to form a complete genome. In a collaboration with Cold Spring Harbor's Dick McCombie and others, Wing hopes to "optimize a blend of 454, Sanger, and Solexa to sequence BAC pools."

Wing bought the 454 GS FLX a year and a half ago, and it was delivered within a month. He purchased the Illumina about six months ago, received it three to four months after that, and has already put in an order for the $40,000 upgrade to the Genome Analyzer II.

Dealing with data

Next-gen sequencing users have found that troubleshooting issues such as sample prep and protocol development are a blip on the radar screen compared to data analysis and storage. JGI's Chen says the biggest challenge is building out the informatics infrastructure to handle  the large amounts of data produced. "We have had to really build up our computing infrastructure to meet the new sequencing technology demands," he says.

Chen is lucky enough to have access to high-end data storage systems thanks to JGI's connection to the Lawrence Berkeley National Laboratory, but scientists at smaller labs are facing the unpleasant decision to discard files so they don't overload their storage capacity.

Washington's Dorschner warns potential customers to be prepared for the computing infrastructure changes. "It definitely requires a serious investment in computational power," he says. "We bought additional analysis and storage servers for the Genome Analyzers, and we are also going to be doing the same thing for the SOLiD system.

In fact, we're probably going to outfit it with another complete server to do the analysis offline rather than on the instrument itself."

Dorschner says his team tends to deep-six image files to keep data storage needs manageable. "We generally don't save images for very long because the terabytes of space that we'd have to have on site would be enormous," he says.

Wing says figuring out the data challenge is so important that it should actually be hammered out first. "My recommendation is to set up the computer infrastructure first and then get the instrument. You can always generate sequence, but being able to process the data efficiently and being able to analyze it is really the critical step."

Next-Gen Platforms on Deck

Even as current vendors divvy up marketshare, there are several up-and-coming sequencing companies getting ready to jockey for position when the time is right.

Scientists think there will be space for them, too. The success of future machines, Shendure says, "will come down to cost, accuracy, and robustness. I think companies like VisiGen, and more particularly Pacific [Biosciences], have the potential to redefine what's possible, if they are successful" with their extremely long reads and high quality genome assembly.

Pacific Biosciences is considered one of the most promising of the lot. Its SMRT (single molecule, real-time) sequencing uses real-time observation of DNA polymerase to call bases in a system that could lead to much faster turnaround and longer reads. CTO Steve Turner says the team has demonstrated speeds of up to 10 bases per second and proof-of-concept read lengths of up to 25,000 bases. "In the five-year timeframe, we believe it will have multi-thousand base reads, maybe even greater than 10,000-base reads," he says. The company, which has increased its employees from 100 in February to 185, has just closed $100 million in additional venture funding. With plans to commercialize by 2010, the company is targeting its price in the same range as ABI and Illumina's platforms.

Like PacBio's technology, VisiGen Biotechnologies is building long-read technology. In February, the company received a patent for real-time, single-molecule DNA sequencing, and while the platform is still in early stages, President Susan Hardin has said one goal was to generate 1 megabase of sequence per second. According to Hardin, who presented at the Association for Biomolecular Resource Facilities meeting this year, VisiGen hopes to launch a sequencing service by the end of 2009.

Just Crossing Over

A couple of so-called next-next-gen sequencers are just beginning to become available. Danaher Motion has commercialized the Polonator, an open platform system developed in George Church's lab. At a cost of $150,000, the platform will come with limited technical support, relative to other systems. The system was introduced by Kevin McCarthy, CTO of Dover, part of Danaher Motion, at the Advances in Genome Biology and Technology conference in February. Early adopters of the instrument include the Broad Institute and the Max Planck Institute of Molecular Genetics. Early this year, the machine supported paired-end sequencing with read lengths of 13 bases for each end. The sequencer comes with two computers, one for image acquisition and instrument control, and the other for image processing and base calling. It also has 3 terabytes of local storage.

With the launch of its HeliScope, which uses a single-molecule, melt-and-resequence strategy to cover each base twice, Helicos Biosciences officially hit the market with its first order this February. The instrument was shipped to Expression Analysis, a genotyping and gene expression company in Durham, NC, which plans to offer limited digital gene expression candidate gene sequencing, and genomic signature sequencing on the instrument to customers. "We liked the fact that the HeliScope did not require sample amplification, thereby removing amplification bias introduced in sample preparation," says Expression Analysis CEO Steve McPhail.

Internally, Helicos says it is working to stabilize some reagent issues, and has used the platform for digital gene expression and bacterial genome sequencing projects. As of March, Helicos has stated that the platform can produce 25 MB per hour for dual-pass applications and 90 megabases for single-pass, lower-accuracy apps. The list price is hefty, though — at $1.35 million, it's more than twice that of competitors 454, Illumina, and ABI.

The Scan

Tens of Millions Saved

The Associated Press writes that vaccines against COVID-19 saved an estimated 20 million lives in their first year.

Supersized Bacterium

NPR reports that researchers have found and characterized a bacterium that is visible to the naked eye.

Also Subvariants

Moderna says its bivalent SARS-CoV-2 vaccine leads to a strong immune response against Omicron subvariants, the Wall Street Journal reports.

Science Papers Present Gene-Edited Mouse Models of Liver Cancer, Hürthle Cell Carcinoma Analysis

In Science this week: a collection of mouse models of primary liver cancer, and more.