- DNA sequencing is the process of determining the precise order of nucleotide bases in a DNA molecule. The four main DNA bases are adenine (A), thymine (T), cytosine (C), and guanine (G). Their specific arrangement carries the genetic information required for the development, functioning, reproduction, and inheritance of living organisms. DNA sequencing makes it possible to read this information and investigate genes, mutations, genetic variation, genome organization, evolutionary relationships, and biological functions. It is a fundamental technique in molecular biology, genetics, genomics, biotechnology, medicine, agriculture, and bioinformatics.
- The central principle of DNA sequencing is that the nucleotide sequence of a DNA molecule can be determined by detecting the order in which its bases are identified during a sequencing reaction. Different sequencing technologies use different biochemical and physical approaches to achieve this. Some methods identify bases through DNA synthesis and fluorescent signals, others detect chemical or electrical changes during nucleotide incorporation, and some measure signals generated as individual DNA molecules pass through specialized sensors. Although the technologies differ, they share the objective of converting biological DNA molecules into digital nucleotide sequences that can be analysed computationally.
- DNA sequencing is based on the structure and complementary base-pairing properties of DNA. DNA generally consists of two antiparallel strands held together by hydrogen bonds between complementary bases. Adenine pairs with thymine, while cytosine pairs with guanine. During DNA synthesis, DNA polymerase uses one strand as a template to construct a complementary strand by adding nucleotides in the 5′ to 3′ direction. This complementary relationship forms the basis of several sequencing methods because the sequence of a newly synthesized strand can reveal the sequence of the original template strand.
- The first major principle in many sequencing methods is the preparation of DNA as a suitable template. DNA is extracted from biological material such as blood, tissue, saliva, cultured cells, plants, microorganisms, or environmental samples. The extracted DNA must be sufficiently pure and intact for the intended sequencing technology. Contaminants such as proteins, RNA, salts, chemicals, and other substances can interfere with library preparation or sequencing reactions. DNA concentration, purity, and integrity are therefore assessed before sequencing. The required DNA quality depends on the platform and application; long-read sequencing often benefits from high-molecular-weight DNA, while some short-read workflows can use fragmented DNA.
- In many sequencing workflows, DNA is converted into a sequencing library. Library preparation involves modifying DNA molecules so that they can be recognized and processed by the sequencing instrument. Depending on the technology, this may include DNA fragmentation, end repair, adapter ligation, barcode addition, amplification, and size selection. Adapters are short, known DNA sequences attached to the ends of DNA fragments. They provide binding sites for sequencing primers or platform-specific components. Barcodes, also known as indexes, are short identifying sequences that allow reads from different samples to be distinguished when multiple libraries are sequenced together.
- DNA fragmentation is commonly used in short-read sequencing because many second-generation platforms are designed to analyse relatively short DNA fragments. Fragmentation may be performed mechanically, enzymatically, or through other methods. The resulting fragments are processed to prepare their ends for adapter attachment. In contrast, long-read sequencing often aims to preserve DNA molecules as intact as possible because longer molecules provide information about larger genomic regions. Some targeted sequencing approaches may also use amplification or capture methods to enrich particular genes or genomic regions before sequencing.
- A key principle of DNA sequencing is the detection of nucleotide identity. Each technology uses a specific signal to distinguish the bases. In fluorescent sequencing, different nucleotides or sequencing reactions generate distinguishable optical signals. In semiconductor sequencing, nucleotide incorporation is detected through changes in hydrogen-ion concentration. In nanopore sequencing, changes in electrical current are measured as DNA passes through a pore. The instrument records these signals, and computational algorithms interpret them to determine the nucleotide sequence.
- The Sanger chain-termination method is based on controlled interruption of DNA synthesis. In this method, a DNA template is copied using a primer, DNA polymerase, normal deoxynucleotide triphosphates (dNTPs), and small quantities of dideoxynucleotide triphosphates (ddNTPs). Normal dNTPs contain a 3′-hydroxyl group that allows the DNA chain to continue growing. Dideoxynucleotides lack this group, so their incorporation prevents the addition of further nucleotides.
- When a ddNTP is incorporated into a growing DNA strand, synthesis stops at that position. The reaction therefore produces DNA fragments of different lengths, each ending at a particular nucleotide. In automated Sanger sequencing, fluorescently labelled ddNTPs distinguish the terminal bases. The fragments are separated according to size using capillary electrophoresis, and a detector records their fluorescent signals. The sequence is reconstructed from the order in which the fragments pass through the detector, generally from the shortest to the longest.
- The output of Sanger sequencing is commonly displayed as a chromatogram containing peaks corresponding to detected nucleotides. Each peak represents a signal associated with a base. Clear, well-separated peaks generally indicate more reliable base identification, while overlapping or weak peaks may indicate uncertainty. Sequence-calling software interprets the signal and assigns nucleotide bases, sometimes using ambiguity codes when the identity of a base cannot be determined confidently. Sanger sequencing is particularly useful for sequencing individual DNA fragments and validating selected genetic variants.
- Next-generation sequencing uses a different principle: many DNA fragments are sequenced in parallel. Instead of analysing one DNA molecule at a time, second-generation platforms process large numbers of DNA fragments simultaneously. This parallel approach produces millions or billions of sequence reads, depending on the platform and run configuration. It has made large-scale genome sequencing, exome sequencing, targeted sequencing, transcriptomics, and population genomics much more accessible.
- One widely used second-generation approach is sequencing by synthesis, particularly the Illumina method. DNA fragments are attached to a flow cell through platform-specific adapter sequences. The fragments are amplified locally to produce clusters containing many copies of the same original DNA fragment. During sequencing, fluorescently labelled reversible-terminator nucleotides are incorporated into newly synthesized DNA strands. Each cycle generally allows one nucleotide to be added, after which the instrument records the fluorescent signal to identify the incorporated base. The fluorescent label and blocking group are removed, allowing the next cycle to proceed.
- The repeated cycles of nucleotide incorporation, signal detection, and chemical processing generate a sequence of bases for each cluster. The instrument’s software converts the recorded signals into nucleotide calls and assigns quality scores to the identified bases. These quality scores estimate the confidence of base identification. In paired-end sequencing, both ends of a DNA fragment are sequenced, providing two reads that originate from the same fragment. The relationship between these reads can improve mapping, genome assembly, and the identification of certain genomic changes.
- Other second-generation technologies use different signal-detection principles. Ion Torrent sequencing detects hydrogen ions released during nucleotide incorporation. The change in hydrogen-ion concentration produces a measurable change in pH, which is detected electronically. Pyrosequencing, used in the historical 454 platform, detects light generated through an enzymatic reaction associated with the release of pyrophosphate during nucleotide incorporation. Sequencing by ligation, used by the historical SOLiD platform, identifies sequence information through the ligation of fluorescently labelled oligonucleotide probes. These approaches illustrate that DNA sequencing can be based on different physical and chemical signals rather than one universal detection mechanism.
- Third-generation sequencing is based largely on the analysis of individual DNA molecules. Unlike many second-generation methods, these technologies do not depend on the same type of clonal amplification of DNA fragments before sequencing. They can produce long reads that span thousands or tens of thousands of bases, and in some cases substantially longer molecules. Long reads provide information about genomic regions that may be difficult to reconstruct using short reads, including repetitive sequences, complex structural variants, duplicated regions, and large insertions or deletions.
- PacBio single-molecule real-time sequencing is based on observing DNA synthesis by an individual DNA polymerase molecule. The polymerase is positioned in a specialized structure called a zero-mode waveguide. Fluorescently labelled nucleotides enter the observation area, and a signal is detected when a nucleotide is incorporated into the growing DNA strand. The fluorescent label is released during incorporation, allowing the polymerase to continue synthesis. The recorded signals are converted into nucleotide sequences using computational methods.
- PacBio also supports circular consensus sequencing, which generates highly accurate HiFi reads. In this approach, a DNA insert is placed in a circular template so that the polymerase can read the same molecule repeatedly. The repeated observations are combined to produce a consensus sequence. This approach helps reduce random errors and provides a combination of long read length and high sequence accuracy. HiFi sequencing is useful for genome assembly, variant detection, haplotype analysis, and the characterization of complex genomic regions.
- Oxford Nanopore sequencing uses a different physical principle. DNA molecules pass through biological nanopores embedded in a membrane, while a controlled electrical potential produces an ionic current through the pore. As DNA moves through the nanopore, groups of nucleotides influence the electrical current in characteristic ways. The resulting signal is recorded and interpreted by computational basecalling algorithms to determine the nucleotide sequence. A motor protein regulates the movement of DNA through the pore, helping control the speed at which the molecule is read.
- Nanopore sequencing can generate long reads and provide sequence information in real time. Depending on the platform and library preparation, it can be used for whole-genome sequencing, pathogen identification, metagenomics, structural variant analysis, and transcript characterization. It can also detect certain DNA modifications, including methylation, through changes in the electrical signal, provided that suitable models and analysis methods are used. Its performance depends on factors such as pore chemistry, read length, sample preparation, basecalling models, and data quality.
- An important principle shared by modern sequencing technologies is the conversion of raw signals into digital sequence data. The instrument does not necessarily produce a nucleotide sequence directly. Instead, it records measurements such as fluorescence intensity, light emission, pH changes, or electrical current. Computational algorithms analyse these signals and assign nucleotide identities. This process is called basecalling. Basecalling methods may use statistical models, signal-processing techniques, machine learning, or neural networks, depending on the platform and software.
- Basecalling also produces quality scores that describe the estimated confidence of individual nucleotide calls. A commonly used measure is the Phred quality score, which expresses the estimated probability of an incorrect base call on a logarithmic scale. A higher Phred score indicates a lower estimated probability of error. For example, a score of Q20 corresponds to an estimated error probability of 1 in 100, while Q30 corresponds to an estimated error probability of 1 in 1,000. These values are theoretical quality estimates and should not be confused with direct experimental confirmation of every nucleotide.
- DNA sequencing accuracy is influenced by several factors, including the chemistry of the sequencing reaction, signal quality, DNA damage, sample contamination, amplification bias, and computational interpretation. Different platforms have characteristic error profiles. Short-read sequencing may provide high per-base accuracy but can struggle with repetitive regions and complex structural variation. Long-read technologies can resolve many difficult genomic regions but may have different error patterns depending on the platform and chemistry. Consensus sequencing, improved basecalling, and complementary sequencing methods can help improve reliability.
- Read length is another important principle in sequencing. A read is a nucleotide sequence obtained from one sequencing event or from one sequenced DNA fragment, depending on the technology. Short reads commonly contain tens to several hundred bases, while long reads may contain thousands or tens of thousands of bases. Longer reads can span repetitive regions and connect genomic features that are separated by large distances. However, longer reads are not automatically more accurate or more informative for every research question. The most suitable read length depends on the genome structure, variant type, sequencing technology, and analysis objective.
- Sequencing depth, also called coverage, describes how many times a genomic position is represented by sequencing reads. Average depth is often expressed as a multiple of genome size. For example, 30× coverage means that the total number of aligned sequencing bases is approximately thirty times the length of the reference genome. Greater depth can improve confidence in many analyses, particularly when identifying low-frequency variants or reconstructing genomes. However, the required depth varies according to the biological application, sample complexity, sequencing technology, and analytical method. High coverage cannot fully compensate for systematic sequencing errors, poor sample preparation, or an unsuitable reference genome.
- The distinction between sequencing depth and breadth of coverage is also important. Depth describes the average number of reads covering a position, while breadth describes the proportion of the target genome or region covered by at least a specified number of reads. A sample may have high average depth but still contain poorly covered regions. Researchers therefore assess both the amount of sequencing data and how evenly that data is distributed across the genome or target regions.
- Another important principle is the use of reference sequences. In reference-based sequencing analysis, reads are aligned to an existing reference genome to identify their genomic positions. Differences between the sample and reference can then be investigated through variant calling. Reference-based analysis is widely used in human genomics, cancer research, microbial genomics, and population studies. However, results can be affected by reference bias when the selected reference does not adequately represent the genetic diversity of the samples being analysed.
- De novo sequencing and genome assembly follow a different approach. Instead of relying entirely on an existing reference genome, sequencing reads are computationally combined to reconstruct longer sequences called contigs. Contigs may be connected into scaffolds using additional information, such as paired-end relationships, long reads, or other linking data. Assembly algorithms use overlaps, sequence graphs, or related computational representations to determine how reads fit together. De novo assembly is particularly useful for organisms without suitable reference genomes and for identifying genomic sequences that may be absent from an existing reference.
- The choice between reference-based analysis and de novo assembly depends on the research objective. Reference-based mapping is often efficient when a high-quality and representative genome is available. De novo assembly is valuable when the objective is to reconstruct a genome, discover novel sequences, or identify structural features that cannot be adequately represented by the available reference. Some studies combine both approaches by assembling genomes and then comparing the resulting sequences with reference genomes.
- DNA sequencing is also based on the principle that different experimental designs reveal different parts of the genome. Whole-genome sequencing aims to examine nearly all genomic regions, including coding and noncoding sequences. Whole-exome sequencing focuses primarily on protein-coding regions, which are enriched before sequencing. Targeted sequencing examines selected genes or genomic regions using methods such as amplicon sequencing or hybridization capture. Each design has different requirements for coverage, cost, data volume, and variant detection. The appropriate strategy depends on the biological question and the types of genetic changes being investigated.
- Sequencing data must undergo quality control before downstream analysis. Raw reads are examined for base quality, adapter contamination, sequence composition, read length, duplication, and other indicators of data quality. FastQC is commonly used for short-read quality assessment, while MultiQC can combine reports from multiple samples and tools. Fastp supports trimming and filtering of short-read data. NanoPlot and other specialized tools can summarize long-read quality and length distributions. Quality control helps identify technical problems before they influence genome alignment, assembly, or variant analysis.
- After quality control, sequencing reads may be aligned to a reference genome or used for de novo assembly. Short-read aligners include BWA-MEM2 and Bowtie 2, while minimap2 is widely used for long-read alignment and several other sequence-mapping tasks. Alignment data are commonly stored in SAM, BAM, or CRAM formats. SAMtools supports sorting, indexing, filtering, and inspecting these files. For variant analysis, tools such as GATK, bcftools, and FreeBayes can identify different classes of sequence variants under appropriate workflows. Structural variant detection may require specialized tools such as Manta, Sniffles, or cuteSV.
- The main output of a sequencing experiment is not simply a list of bases; it is biological information that must be interpreted in context. Sequence reads can reveal single-nucleotide variants, insertions, deletions, copy-number changes, structural rearrangements, gene content, and other genomic features. These findings can be used to investigate inherited diseases, cancer, pathogen evolution, antimicrobial resistance, population diversity, crop improvement, livestock breeding, and conservation. However, the presence of a variant does not automatically establish its biological effect. Functional annotation, statistical analysis, population data, experimental validation, and relevant biological evidence are often required.
- DNA sequencing technologies have several important limitations. Sequencing errors may arise during DNA extraction, library preparation, amplification, signal detection, basecalling, alignment, and variant calling. Short reads can map ambiguously to repetitive or duplicated regions, while long reads may have platform-specific error patterns and require high-quality DNA for optimal performance. Low sequencing depth can result in missed variants, and uneven coverage can leave important regions inadequately analysed. Contamination, sample mix-ups, incomplete reference genomes, and inaccurate annotations can also affect the reliability of results.
- Reproducibility is an essential principle of sequencing research. Researchers should record the biological sample source, extraction method, library preparation protocol, sequencing platform, chemistry, read configuration, sequencing depth, quality-control procedures, software versions, reference genome assembly, and analysis parameters. Raw sequence reads, processed data, genome assemblies, and variant files should be stored in appropriate repositories when permitted. Major public sequence repositories include the Sequence Read Archive (SRA), European Nucleotide Archive (ENA), and DNA Data Bank of Japan (DDBJ), which participate in the International Nucleotide Sequence Database Collaboration. Human genomic data may require controlled access to protect participant privacy.
- DNA sequencing is based on the detection and computational interpretation of nucleotide-specific signals generated from DNA molecules. Sanger sequencing identifies bases through chain termination and fragment separation, second-generation technologies sequence large numbers of DNA fragments in parallel, and third-generation technologies analyse individual DNA molecules using real-time synthesis or nanopore signal detection. Despite their differences, these approaches share the fundamental goal of determining DNA sequence information accurately and converting it into data suitable for biological analysis. Understanding these principles is essential for selecting appropriate sequencing technologies, designing reliable experiments, evaluating sequence quality, and interpreting genomic results in bioinformatics.