Quantcast

Saying goodbye to 454: how to choose your next NGS platform

The Rise and Fall of the 454 Sequencer 

The GS20 454 sequencer, released in 2005, was the first next-generation DNA sequencer to hit the market, and its feats quickly dazzled the scientific community. As new sequencing platforms proliferated, however, many researchers opted for less expensive options and 454 market share fell.

About a year ago, Roche announced that it was closing its 454 facility in Branford, Connecticut and phasing out the sequencers by mid-2016. Many researchers are discovering that their core labs are already retiring 454 sequencers, while other scientists have decided on their own to identify good alternatives in advance of the phase-out. Luckily, sequencing choices abound, but choosing the right platform can be tricky. Here, we offer a brief guide to your next gen sequencing choices.

Some 454 Alternatives

Looking for a low-cost, flexible platform?

Then you should check out Illumina’s HiSeq range. The HiSeq 2500 model is available for $700,000. Illumina’s next gen sequencing systems are the most popular by leaps and bounds. Of all the machines available on the market today, the HiSeq provides the greatest output at the lowest reagent cost ($0.07 per megabase, compared to $0.13 for the SOLiD and $10 for the 454*). In addition, it offers a lot of flexibility. Want to design runs with varying read length? Do both paired-end and single-end sequencing? Multiplex many samples? The HiSeq can do it all.

Need 99.9% accuracy?

Then you might want to consider Applied Biosystem’s SOLiD system, another fairly popular next gen platform that retails for roughly $500,000 (v4). One advantage of SOLiD sequencing is its accuracy: 99.9% vs the HiSeq’s 98%*. However, SOLiD does produce shorter reads and less data than the HiSeq (120 vs. 600 gigabases per run, respectively*). In addition, runs take longer (7–14 days depending on whether single end or paired-end sequencing is being performed vs. 3–10 days for the HiSeq*).

 Is long read length a must?

If the answer is yes, take a look at Pacific Bioscience’s RS II system, which has a list price of about $700,000. Although the HiSeq and SOLiD work fine for many research applications, their read length is an order of magnitude shorter than the 454’s. This may limit their utility for purposes such as genome assembly. For this reason, although 3rd generation sequencing platforms such as the RS have yet to really catch on, due to early problems with accuracy and expense ($7–$38 per megabase a few years ago, now a more competitive $1 per megabase), their very long read length is garnering attention. Recently, Sergey Koren and colleagues reported a combination next generation/3rd generation method that resulted in long reads and nearly perfect accuracy. This type of approach may gain popularity among researchers aiming to sequence novel genomes or perform structural variation studies, especially after the 454 disappears. An added bonus of the real-time RS system is its short run time: only 30 minutes to 3 hours (vs. days for the HiSeq and SOLiD).

 We’ve just described a few of the more popular alternatives to 454 sequencing here. However, these aren’t the only possibilities. Interested in your very own desktop next generation sequencer? Check out Illumina’s MiSeq (~$100,000 for the instrument; $0.10 per megabase for v3) or Applied Biosystem’s Ion Proton (~$225,000 for the instrument; the company estimates the cost per megabase is $0.01 for the Proton III). And don’t forget the still exotic 3rd generation sequencers, of which the RS is only one example. Oxford’s Nanopore systems, another example, are based on electronic, single molecule-sensing technology: no amplification, labeling, or optical instrumentation required. Although Oxford Nanopore’s MinION is not yet generally available, some researchers are participating in an early access program for the flash drive-sized sequencer; the cost of reagents per megabase is ~$1.

If you want to delve into this subject in more detail, we’ve listed some helpful resources below. They include more specifics on the systems highlighted here, as well as pros and cons of those we didn’t have room to feature. Happy hunting!

 

*based on comparison of HiSeq 2000, 454 GS FLX and SOLiDv4.

 

Some comparisons of high-throughput sequencing platforms

Frey KG, Herrera-Galeano JERedden CLLuu TV, Servetas SL, Mateczun AJ, et al. (2014). Comparison of three next-generation sequencing platforms for metagenomic sequencing and identification of pathogens in blood. BMC Genomics. 15:96.

Liu LLi Y Li SHu NHe YPong R, et al. (2012). Comparison of next-generation sequencing systems. Journal of Biomedicine and Biotechnology. Article ID 251364.

Quail MA, Smith M, Coupland P, Otto TD, Harris SR, Connor TR, et al. (2012). A tale of three next generation sequencing platforms: comparison of Ion Torrent, Pacific Biosciences, and Illumina MiSeq sequencers. BMC Genomics. 13:341.

The Molecular Ecologist’s 2014 NGS Field Guide.

Next generation sequencing practical course (free, online) from the European Bioinformatics Institute.

3 Comments

  1. Bitesize Bio on November 18, 2015 at 7:15 am

    […] goal of course, is to sequence your hard won genomic data. There are many types of sequencing methods currently available, and the rapidly decreasing cost of Next Generation Sequencing (NGS) are […]

  2. scbaker on December 19, 2014 at 6:02 pm

    I think you’re doing your readers a disservice by steering them away from one dying/dead platform (454) only to have them replace it with another dying/dead platform (SOLiD). Even Thermo/Life doesn’t talk about SOLiD anymore.

    Also, pretty much all of your numbers are out of date (not just the PacBio numbers). You’re linking to a table that doesn’t list the latest instruments. If you want to see the latest data all in one place, check out AllSeq’s NGS Knowledge Bank. We’ve got the latest outputs and costs per system, as well as a qualitative assessment of each platform for the various applications.

    AllSeq’s NGS Knowledge Bank: http://allseq.com/knowledgebank/sequencing-platforms/illumina

  3. DavidKonigsberg on December 18, 2014 at 5:14 am

    Thanks for the summary. The PacBio numbers are a bit out of date and we finally have a couple Nanopore publications to infer actual throughput.

    For PacBio, the current chemistry is claiming 10,000 to 15,000 average read length with average reads per cell about 55k. And on twitter folks appear to be confirming that (e.g. https://twitter.com/geesthc/status/545156720251899905 and https://twitter.com/lexnederbragt/status/542040161237073922) though no papers out yet with the latest chemistry. Probably some at PAG next month. Together that should lower your cost per base estimate by a factor of 7.6 (before based upon 30k reads at 3kbp versus 55k reads at 12.5kbp)

    And for nanopore, we don’t have many publications yet but one project with 30 flow cells had a mean of 50 Mb per cell (see http://schatzlab.cshl.edu/presentations/2014.09.20.Genome%20Informatics.NanoCorr2.pdf) so instead of that February projection of 900 Mb per cell looks like your cost per megabase was low by a factor of 18? Assumes those cost projections hold, but it’s not for sale yet so who knows?

    Also, if you’re going for accuracy you should consider PacBio again. Not raw read accuracy certainly but consensus accuracy. E.g. http://dazzlerblog.wordpress.com/2014/05/15/on-perfect-assembly/

Leave a Comment





Share14
Tweet
Share11
+11