Overview of Technical Aspects and Chemistries of Next-Generation Sequencing

Ian S. Hagemann,    Departments of Pathology and Immunology and of Obstetrics and Gynecology, Washington University School of Medicine, St. Louis, MO, USA

Clinical genomic testing has been made possible by the development of "next-generation" sequencing (NGS) technologies that allow large quantities of nucleic acid sequence to be obtained in a clinically meaningful time frame. This chapter reviews the sequencing platforms that are currently in most common use, including Illumina, Ion Torrent, SOLiD, and Roche 454. Sanger sequencing is discussed as a "first-generation" technology that retains an important place in clinical genomics for orthogonal validation of NGS findings and for coverage of areas not amenable to NGS.

Keywords

Massively parallel sequencing; medical laboratory technologies; genomics

Outline

Clinical Molecular Testing: Finer and Finer Resolution 3

Sanger Sequencing 4

Chemistry of Sanger Sequencing, Electrophoresis, Detection 4

Applications in Clinical Genomics 5

Technical Constraints 6

Read Length and Input Requirements 6

Pooled Input DNA Puts a Limit on Sensitivity 7

Cyclic Array Sequencing 7

Illumina Sequencing 8

Library Prep and Sequencing Chemistry 9

Choice of Platforms 10

Phasing 10

SOLiD Sequencing 11

Ion Torrent Sequencing 14

AmpliSeq Library Preparation 16

Roche 454 Genome Sequencers 16

Third-Generation Sequencing Platforms 18

References 18

Clinical Molecular Testing: Finer and Finer Resolution

Progress in applying genetic knowledge to clinical medicine has always been tightly linked to the nature of the genetic information that was available for individual patients.

Classical cytogenetics provides pan-genomic information at the level of whole chromosomes and sub-chromosomal structures on the scale of megabases. The availability of clinical cytogenetics made it possible to establish genotype-phenotype correlations for major developmental disabilities, including +21 in Down syndrome, the "fragile" X site in Fragile X syndrome, monosomy X in Turner syndrome, and the frequent occurrence of trisomies, particularly +13, +17, and +14, in spontaneous abortions.

Over time, new experimental techniques have allowed knowledge to be accumulated at finer and finer levels of resolution, such that genotype-phenotype correlations are now routinely established at the single-nucleotide level. Thus it is now well known that germline F5 p.R506Q mutation is responsible for the factor V Leiden phenotype [1] and that loss of imprinting at the SNRPN locus is responsible for Prader-Willi syndrome [1], to cite examples of two different types of molecular lesions. Clinical advances have been closely paralleled by progress in research testing, since the underlying technologies tend to be similar.

Historically, much clinical molecular testing has taken an indirect approach to determining gene sequences. Although the sequence was fundamentally the analyte of interest, indirect approaches such as restriction fragment length polymorphism (RFLP) analysis, allele-specific polymerase chain reaction (PCR), multiplex ligation-dependent probe amplification (MLPA), and invader chemistry assays have proven easier to implement in the clinical laboratory-easier and more cost-effective to standardize, to perform, and to interpret [2].

Technological advances in the past two decades have begun to change this paradigm by vastly facilitating the acquisition of gene sequence data. Famously, the human genome project required an investment of 10 years and about 10 billion dollars to determine the genomic sequence of a single reference individual. While the technology used for that project was innovative at the time, the effort and cost were clearly monumental and the project could never have been translated directly into a clinical testing modality. Fundamental technical advances, broadly described as next-generation sequencing (NGS), have lowered the cost and difficulty of genomic sequencing by orders of magnitude, so that it is now practical to consider implementing these methods for clinical testing.

The first section of this book is a survey of the technologies used for NGS today. The present chapter focuses on the lowest-level building blocks of NGS: the chemical and technological basis of the methods used to convert nucleic acids into sequence. Subsequent chapters deal with methods for selecting the molecules to be sequenced (whole genome, exome, or gene panels) as well as different approaches for enriching the reagent pool for these molecules (capture and amplification) (Chapters 2-4). The section closes with a chapter on emerging "third-generation" methods, which promise to eventually allow single-molecule sequencing (Chapter 5), as well as a chapter on RNA-based methods which allow NGS technology to be used for expression profiling (Chapter 6).

Sanger Sequencing

Chemistry of Sanger Sequencing, Electrophoresis, Detection

In Sanger sequencing [3], DNA polymerase is used to synthesize numerous copies of the sequence of interest in a single primer extension step, using single-stranded DNA as a template. Chain-terminating 2´,3´-dideoxynucleotide triphosphates (ddNTPs) are spiked into the reaction. At each nucleotide incorporation event, there is chance that a ddNTP will be added in place of a dNTP, in which case, in the absence of a 3´ hydroxyl group, the growing DNA chain will be terminated. The endpoint of the reaction is therefore a collection of DNA molecules of varying lengths, each terminated by a dideoxynucleotide [4].

The original Sanger sequencing method consists of two steps. In the "labeling and termination" step, primer extension is performed in four parallel reactions, each reaction containing a different ddNTP in addition to [a-35S]dATP and dNTPs. A "chase" step is then performed with abundant unlabeled dNTPs. Any molecules that have not incorporated a ddNTP will be extended so that they do not interfere with detection. The products are then separated by polyacrylamide gel electrophoresis in four parallel lanes representing ddA, ddT, ddC, and ddG terminators. The DNA sequence is read off of an autoradiograph of the resulting gel by calling peaks in each of the four lanes (Figure 1.1A).

Historically, Sanger sequencing employed the Klenow fragment of Escherichia coli DNA polymerase I. The Klenow fragment has 5´3´ polymerase and 3´5´ exonuclease activity, but lacks 5´3´ exonuclease activity [5], thus preventing degradation of desired DNA polymerase products. Klenow fragment is only moderately processive and discriminates against incorporation of ddNTPs, a tendency which can be reduced by including Mn2+ in the reaction [6]. Sequenase, which was also commonly used, is a modified T7 DNA polymerase with enhanced processivity over Klenow fragment, a high elongation rate, decreased exonuclease activity, and minimal discrimination between dNTPs and ddNTPs [6,7].

Several variants of Sanger sequencing have been developed. In one of these, thermal cycle sequencing, 20-30 denaturation-annealing-extension cycles are carried out, so that small numbers of template molecules can be repeatedly utilized; since only a single sequencing primer is present, the result is linear amplification of the signal, rather than exponential amplification as would be the case in a PCR [4,8]. The high-temperature steps present in thermal cycle sequencing protocols have the advantage of melting double-stranded templates and disrupting secondary structures that may form in the template. A high-temperature polymerase, such as Taq, is required. Taq polymerase discriminates against ddNTPs, requiring adjustment of the relative concentration of dNTPs and ddNTPs in these reactions. Native Taq polymerase also possesses undesirable 5´3´ exonuclease activity, but this has been engineered out of commercially available recombinant Taq [4].

Other variant approaches consist of different detection methods:

 When radioisotope detection was in use, the original [a-32P]dATP protocol was modified to allow use of [a-33P]dATP and [a-35S]dATP, lower-energy emitters producing sharper bands on the autoradiogram [9].

 Chemiluminescent detection was also reported using biotinylated primers, streptavidin, and biotinylated alkaline phosphatase [10].

 5´-end labeling of the...