BIF604 Handouts

Next-generation DNA · sequencing techniques · Wilhelm J. Ansorge

Ecole Polytechnique Federal Lausanne, EPFL, Switzerland

Next-generation high-throughput DNA sequencing techniques are opening fascinating opportunities in the life sciences. Novel fields and applications in biology and medicine are becoming a reality, beyond the genomic sequencing which was original development goal and application. Serving as examples are: personal genomics with detailed analysis of individual genome stretches; precise analysis of RNA transcripts for gene expression, surpassing and replacing in several respects analysis by various microarray platforms, for instance in reliable and precise quantification of transcripts and as a tool for identification and analysis of DNA regions interacting with regulatory proteins in functional regulation of gene expression. The next-generation sequencing technologies offer novel and rapid ways for genome-wide characterisation and profiling of mRNAs, small RNAs, transcription factor regions, structure of chromatin and DNA methylation patterns, microbiology and metagenomics. In this article, development of commercial sequencing devices is reviewed and some European contributions to the field are mentioned. Presently commercially available very high-throughput DNA sequencing platforms, as well as techniques under development, are described and their applications in bio-medical fields discussed.

Introduction Next-generation high-throughput DNA sequencing techniques, which are opening fascinating new opportunities in biomedicine, were selected by Nature Methods as the method of the year in 2007 [1]. However, the path to gaining acceptance of the novel technology was not an easy one. Until a few years ago the methods used for the sequencing were the Sanger enzymatic dideoxy technique first described in 1977 [2] and the Maxam and Gilbert chemical degradation method described in the same year [3], which was used in sequence cases which could not easily be resolved with the Sanger technique. The two laboratories where the first automated DNA sequencers were produced, simultaneously, were those of Leroy Hood at Caltech [4], commercialised by Applied Biosystems, and Wilhelm Ansorge at the European Molecular Biology Laboratory EMBL [5,6] and commercialised by Pharmacia-Amersham, later General Electric (GE) Healthcare. The Sanger method was used in the first automated fluorescent project for sequencing of a genome region, in which sequence determination of the complete gene locus for the HPRT gene was performed using the EMBL technique; in that project the important concept of paired-end sequencing was also introduced for the first time [7]. The achievement of successful and unambiguous sequencing of a real genomic DNA region, loaded with many sequence pitfalls like Alu sequences in both directions of the HPRT gene locus, demonstrated the feasibility of using an automated fluorescence-based technique for the sequencing of entire genomes, and in principle the feasibility of the technical sequencing part of the Human Genome project. Whentheinternationalcommunitydecided ondeterminationof the whole human genome sequence, the goal triggered the development of techniques allowing higher sequencing throughput. In Japan, the work on fluorescent DNA sequencing technology by the team of H. Kambara ( a.html) in the Hitachi laboratories resulted in the developmentafter 1996 of a high-throughput capillary array DNA sequencer. Two

1871-6784/$ - see front matter  2009 Published by Elsevier B.V. doi:10.1016/j.nbt.2008.12.009 companies, ABI (commercialising the Kambara system) and Amersham(takingoveranddevelopingfurtherthesystemsetupintheUS by the Molecular Dynamics company), commercialised automated sequencing using parallel analysis in systems of up to 384 capillaries at that time. Together with partial miniaturisation of the robotic sample preparation, large efforts in automation of laboratory processes and advances in new enzymes and biochemicals, the Sanger technique made possible the determination of the sequence of the human genome by two consortia working in parallel. It was the unique method used for DNA sequencing, with innumerable applications in biology and medicine. As the users and developers of the DNA sequencing techniques realised, the great limitations of the Sanger sequencing protocols for even larger sequence output were the need for gels or polymers used as sieving separation media for the fluorescently labelled DNA fragments, the relatively low number of samples which could be analysed in parallel and the difficulty of total automation of the sample preparation methods. These limitations initiated efforts to develop techniques without gels, which would allow sequence determination on very large numbers (i.e. millions) of samples in parallel. One ofthefirstdevelopmentsofsuchatechniquewasattheEMBL(atthat time one of the two world leaders in DNA sequencing technology) from 1988 to 1990. A patent application by EMBL [8] described a large-scale DNA sequencing technique without gels, extending primers in ‘sequencing-by-synthesis, addition and detection of the incorporatedbase’,proposingand describing the use ofthe so-called ‘reversible terminators’ for speed and efficiency [8]. The first step of the technique consisted in detecting the next added fluorescently labelled base (reversible terminator) in the growing DNA chain by means of a sensitive CCD camera. This was performed on a large number of DNA samples in parallel, attached either to a planar support or to beads, on DNA chips, minimising reaction volumes in a miniaturised microsystem. In the next step the terminator was convertedtoastandardnucleotideandthedyeremovedfromit. This cycleandtheprocesswererepeatedtodeterminethenextbaseinthe sequence. Theprincipledescribed in thepatent application isin part very similar to that used today in the so-called next-generation devices, with many additional original developments commercialised by Illumina-Solexa, Helicos and other companies. Since 2000, focused developments have continued in several groups. Various institutions, particularly European laboratories, considered the capillary systems as the high point and in a less visionary decision ceased developments of even the most promising novel sequencing techniques, turning their attention exclusively to arrays. By contrast, in the US, funding for development and testing of novel, non-gel-based high-throughput sequencing technologies were provided by the large granting agencies and private companies. Efforts to bring the platforms to maturity were under way. The resulting devices and platforms available on the market in mid-2008, as well as some interesting parallel developments, are described in more detail below. The EU has recently initiated significant support for the development of novel highthroughput DNA sequencing technologies, among others the

Next-generation DNA sequencing platforms Novel DNA sequencing techniques provide high speed and throughput, such that genome sequencing projects that took several years with the Sanger technique can now be completed in a matter of weeks. The advantage of these platforms is the determination of the sequence data from amplified single DNA fragments, avoiding the need for cloning of DNA fragments. A limiting factor of the new technology remains the overall high cost for generating the sequence with very high-throughput, even though compared with Sanger sequencing the cost per base is lower by several orders of magnitude. Reduction of sequencing errors is another factor; in this respect the Sanger sequencing technique remains competitive in the immediate future. Other limitations in some applications are short read lengths, non-uniform confidence in base calling in sequence reads, particularly deteriorating 30-sequence quality in technologies with short read lengths and generally lower reading accuracy in homopolar stretches of identical bases. The huge amount of data generated by these systems (over a gigabase per run) in the form of short reads presents another challenge to developers of software and more efficient computer algorithms.

The 454 GenomeSequencer FLX instrument (Roche Applied Science) The principle of pyrophosphate detection, the basis of this device, was described in 1985 [9], and a system using this principle in a new method for DNA sequencing was reported in 1988 [10]. The technique was further developed into a routinely functioning method by the teams of M. Ronaghi, M. Uhlen, and P. Nyren in Stockholm [11], leading to a technique commercialised for the analysis of 96 samples in parallel in a microtiter plate. TheGSinstrumentwasintroducedin 2005, developed by454 Life Sciences, as the first next-generation system on the market. In this system (Fig. 1),DNAfragmentsareligatedwithspecific adaptersthat causethe binding ofone fragmentto a bead. EmulsionPCR iscarried out for fragment amplification, with water droplets containing one bead and PCR reagents immersed in oil. The amplification is necessarytoobtainsufficientlightsignalintensityforreliabledetectionin the sequencing-by-synthesis reaction steps. When PCR amplification cycles are completed and after denaturation, each bead with its one amplified fragment is placed at the top end of an etched fibre in anoptical fibre chip, created from glassfibre bundles. Theindividual glass fibres are excellent light guides, with the other end facing a sensitive CCD camera, enabling positional detection of emitted light. Each bead thus sits on an addressable position in the light guide chip, containing several hundred thousand fibres with attached beads. In the next step polymerase enzyme and primer are added to the beads, and one unlabelled nucleotide only is supplied to the reaction mixture to all beads on the chip, so that synthesis of the complementary strand can start. Incorporation of a following base by the polymerase enzyme in the growing chain releases a pyrophosphate group, which can be detected as emitted light. Knowing the identity of the nucleotide supplied in each step, the presence of a light signal indicates the next base incorporated into the sequence of the growing DNA strand. The method has recently increased the achieved reading length to the 400–500 base range, with paired-end reads, and as such is being applied to genome (bacterial, animal, human) sequencing. One spectacular application of the system was the identification of the culprit in the recent honey-bee disease epidemics (see company web pages below). A relatively high cost of operation and generally lower reading accuracy in homopolar stretches of identical bases are mentioned presently as the few drawbacks of the method. The next upgrade 454 FLX Titanium will quintuple the data output from 100 Mb to about 500 Mb, and the new picotiter plate in the device uses smaller beads about 1 mm diameter. The device, schema of operation, its further developments and list of publications with applications can be found at

The Illumina (Solexa) Genome Analyzer The Solexa sequencing platform was commercialised in 2006, with Illumina acquiring Solexa in early 2007. The principle (Fig. 2) is on the basis of sequencing-by-synthesis chemistry, with novel reversible terminator nucleotides for the four bases each labelled with a different fluorescent dye, and a special DNA polymerase enzyme able to incorporate them. DNA fragments are ligated at both ends to adapters and, after denaturation, immobilised at one end on a solid support. The surface of the support is coated densely with the adapters and the complementary adapters. Each single-stranded fragment, immobilised at one end on the surface, creates a ‘bridge’ structure by hybridising with its free end to the complementary adapter on the surface of the support. In the mixture containing the PCR amplification reagents, the adapters on the surface act as primers for the following PCR amplification. Again, amplification is needed to obtain sufficient light signal intensity for reliable detection of the added bases. After several PCR cycles, random clusters of about 1000 copies of single-stranded DNA fragments (termed DNA ‘polonies’, resembling cell colonies after polymerase amplification) are created on the surface. The reaction mixture for the sequencing reactions and DNA synthesis is supplied onto the surface and contains primers, four reversible terminator nucleotides each labelled with a different fluorescent dye and the DNA polymerase. After incorporation into the DNA strand, the terminator nucleotide, as well as its position on the support surface, is detected and identified via its fluorescent dye by the CCD camera. The terminator group at the 30-end of the base and the fluorescent dye are then removed from the base and the synthesis cycle is repeated. The sequence read length achieved in the repetitive reactions is about 35 nucleotides. The sequence of at least 40 million polonies can be simultaneously determined in parallel, resulting in a very high sequence throughput, on the order of Gigabases per support. In 2008 Illumina introduced an upgrade, the Genome Analyzer II that triples output compared to the previous Genome Analyzer instrument. A paired-end module for the sequencer was introduced, and with new optics and camera components that allow the system to image DNA clusters more efficiently over larger areas, the new instrument triples the output per paired-end run from 1 to 3 Gb. The system generates at least 1.5 Gb of single-read data per run, at least 3 Gb of data in a paired-end run, recording data from more than 50 million reads per flow cell. The run time for a 36cycle run was decreased to two days for a single-read run, and four days for a paired-end run. Information on the Genome Analyzer system can be found at and in [1].

The Applied Biosystems ABI SOLiD system The ABI SOLiD sequencing system, a platform using chemistry based upon ligation, was introduced in Autumn 2007. The generation of a DNA fragment library and the sequencing process by subsequent ligation steps are shown schematically in Figs 3,4. In this technique, DNA fragments are ligated to adapters then bound to beads. A water droplet in oil emulsion contains the amplification reagents and only one fragment bound per bead; DNA fragments on the beads are amplified by the emulsion PCR. After DNA denaturation, the beads are deposited onto a glass support surface. In a first step, a primer is hybridised to the adapter. Next, a mixture of oligonucleotide octamers is also hybridised to the DNA fragments and ligation mixture added. In these octamers, the doublet FIGURE 1 (A) Outline of the GS 454 DNA sequencer workflow. Library construction (I) ligates 454-specific adapters to DNA fragments (indicated as A and B) and couples amplification beads with DNA in an emulsion PCR to amplify fragments before sequencing (II). The beads are loaded into the picotiter plate (III). (B) Schematic illustration of the pyrosequencing reaction which occurs on nucleotide incorporation to report sequencing-by-synthesis. (Adapted from .)