In recent years there have been tremendous advances in our ability to rapidly and cost-effectively sequence DNA. This has revolutionized the fields of genetics and biology, leading to a deeper understanding of the molecular events in life processes. The rapid technological advances have enormously expanded sequencing opportunities and applications, but also imposed strains and challenges on steps prior to sequencing and in the downstream process of handling and analysis of these massive amounts of sequence data. Traditionally, sequencing has been limited to small DNA fragments of approximately one thousand bases (derived from the organism’s genome) due to issues in maintaining a high sequence quality and accuracy for longer read lengths. Although many technological breakthroughs have been made, currently the commercially available massively parallel sequencing methods have not been able to resolve this issue. However, recent announcements in nanopore sequencing hold the promise of removing this read-length limitation, enabling sequencing of larger intact DNA fragments. The ability to sequence longer intact DNA with high accuracy is a major stepping stone towards greatly simplifying the downstream analysis and increasing the power of sequencing compared to today. This review covers some of the technical advances in sequencing that have opened up new frontiers in genomics.
Fully understanding the language of DNA requires the complete determination of the order of bases in the genome of humans (or other organisms of interest). Acquiring that knowledge promises to yield more complete insights into biological variations and etiology of diseases. In the dawn of sequencing, reading the order of the four bases of DNA was a cumbersome process. Sequencing was made possible, although still difficult, by the introduction of the chemical degradation method of Maxam and Gilbert  and Sanger’s chain-termination sequencing method  at the end of the 1970s. The latter proved to be more useful and was to be the dominant DNA sequencing technique for almost three decades, propelled by the Human Genome Project (HGP), and is still considered by many the “gold standard”. The commercial launch of massively parallel DNA sequencing instruments in 2005 initiated a paradigm shift powered by new DNA sequencing techniques that inspired researchers to address bolder questions in genome-wide experiments. Recent years have seen many new contenders in the field of massively parallel sequencing. Notably, innovative novel sequencing methods using single molecules of DNA and real-time detection have emerged. Currently, these novel approaches are complementing existing sequencing platforms, but they have a long way to go before replacing massively parallel sequencing.
The most commonly used DNA sequencing technologies, and their challenges and limitations, are described below. A summary table of the features of each technology is provided in Table 1 while the number of sequence reads and read lengths are shown in Fig. 1.
2 Sanger sequencing – chain-termination sequencing
The ingenuity of the chain-termination technique lies in the use of chain-terminating nucleotides, dideoxy - nucleotides that lack a 3’-hydroxyl group, restricting further extension of the copied DNA chain. Early Sanger sequencing required the sequencing process to be split into four separate reactions. Each reaction involves a singlestranded DNA template, a DNA primer and a DNA polymerase in the presence of a mixture of the four unmodified nucleotides, one of which is labeled, and a type of modified chain-terminating nucleotide. Fragments of varying length are synthesized after primer hybridization and polymerase extension, all having the same 5’ end but terminated by a chain-terminating nucleotide at the 3’ end. Adding only a fraction of the terminating nucleotide ensures the random incorporation of dideoxynucleotides in only a small subset of molecules. Since different chain- terminating nucleotides are used in the four sequencing reactions, all combinations of termination can be produced. The generated 3’-terminated DNA templates are then heat-denatured and fractionated by gel-electro pho - resis, running products of all four sequencing reactions in parallel . Using radioactive or, more recently, fluorescent labeling to visualize the bands enables the sequence of the original DNA template to be determined by following the migration order of successively larger fragments in the gel.
Several enhancements have been made to the original method developed by Sanger. These include: labeling the chain-terminating nucleotides with spectrally distinct fluorescent dyes enabling: a single tube and lane to be used in the fragment generation and fractionation steps, respectively [3, 4]; elimination of the need to cast gels by using capillary gel electrophoresis [5, 6]; and automation of the protocol, leading to increases in parallelization, reproducibility and throughputs . Sanger sequencing is still widely used today for many applications, particularly validation of genetic variants and in cases where high quality reads of 300–900 bases are needed. However, the major advances in sequencing technology in recent years have not been related to the mature Sanger sequencing method, but to the rapidly evolving massively parallel sequencing methods.
3 Massively parallel sequencing – consensus sequencing
The most widely used sequencing platforms in genetic research today are the massively parallel sequencing platforms, which have inherited many features from Sanger sequencing, such as the use of polymerases for synthesis, modified nucleotides and fluorescent detection. Another feature is that they require the DNA to be clonally amplified forming a consensus template prior to sequencing.
These sequencing methods have been called next generation sequencing, high-throughput sequencing methods, or second-generation sequencing. However, as the sequencing technologies continue to develop, it is hard to keep referring to old and novel emerging sequencing technologies as belonging to a certain generation. Therefore, it is preferable to categorize them by their most prominent common feature, such as being massively parallel or single-molecule sequencing methods. Currently, there are five competing massively parallel sequencing technologies, each with specific strengths and weaknesses. These are described and discussed below.
Melamede originally outlined the concept of sequencingby-synthesis (SBS) in 1985, in a report of efforts to detect nucleotide incorporation events by measuring nucleotide absorbance . Unaware of Melamede’s findings, Nyrén conceived another SBS approach using bioluminescence instead of absorbance in 1986 , which led (after another 12 years of experiments) to the introduction of pyro - sequencing . Pyrosequencing technology was able to complement Sanger sequencing and was further developed by 454 Life Sciences, founded by Jonathan Rothberg. In 2005, Rothberg and colleagues  released the first proof-of-concept paper demonstrating a massively parallel pyrosequencing approach, yielding a tremendous increase in sequence capacity, and thereby transforming the way sequencing was conducted.
In SBS, the event that is detected is the incorporation of nucleotides into growing DNA strands. As the nucleotides are incorporated by the polymerase, pyrophosphate (PPi) and protons are generated. In pyrosequencing, the PPi is used in an enzymatic cascade to generate a light burst. The PPi is converted by ATP sulfurylase to ATP, which is then used by the firefly enzyme luciferase to generate photons, providing a light signal that is proportional to the number of nucleotides being incorporated. The challenge is to detect the flashes of light from each unique DNA template while sequencing numerous templates in parallel. Spatially separating each sequencing reaction on beads deposited in small wells on a pico - titer plate elegantly solved this problem.
Massively parallel pyrosequencing begins with fragmentation of the DNA and adaptor ligation. Single-stranded DNA templates are then bound on beads and emulsion PCR is performed, clonally amplifying each DNA template in aqueous microreactors isolated by oil. The emulsion is then broken and the beads carrying DNA are separated from empty beads in a process called enrichment. The enriched beads are deposited in the small wells on the picotiter plate together with primer and DNA polymerase. Ideally, only one of these beads will fit into each well. Smaller beads are also added, carrying the enzymes responsible for the light generation using PPi. Nucleotides are then passed over the substrate in a laminar flow of solutions applied in a predetermined order, and the flashes of light in each well representing incorporation events are recorded. Efficient removal of reaction by-products, which could otherwise perturb the sequencing reaction, is facilitated by the laminar flow. Hence, the sequence of the DNA templates is determined from the knowledge of their location, the order of the flow of nucleotides and records of each flash of light from each well [10–12]. The major drawback using this approach are difficulties in sequencing stretches of identical nucleotides (homopolymeric regions) longer than approximately five nucleotides due to the nonlinear light response they generate .