DNA Sequencing: Definition, Methods, and Applications

DNA sequencing—determining the exact order of nucleotides in a DNA molecule—is the foundational technology of modern genomics. Since the first Sanger sequencing reactions in 1977, three generations of technology have emerged, each with distinct trade-offs in read length, throughput, accuracy, and cost. Selecting the right method for a specific biological question is no longer about which technology can sequence DNA, but which one provides the optimal balance of these parameters for the project at hand.

This guide provides a practical framework for that decision. It covers all major sequencing platforms—Sanger, Illumina NGS, PacBio SMRT, and Oxford Nanopore—with a focus on their performance characteristics, application fit, and the key parameters that determine whether a sequencing project succeeds or fails. The article is written for researchers who understand basic molecular biology and need a decision-oriented overview to plan their next sequencing experiment. The focus throughout is on actionable guidance: how many reads you need, which platform matches your application, and what practical steps ensure a successful outcome. By the end of this article, you will have the framework you need to design your sequencing project with confidence, regardless of your experience level.

What Is DNA Sequencing — And Why the Choice of Method Matters

DNA sequencing reads the order of the four nucleotide bases—adenine (A), cytosine (C), guanine (G), and thymine (T)—in a DNA molecule. The human genome contains approximately 3.2 billion base pairs; a typical bacterial genome contains 2-5 million base pairs. The sequencing method determines how much of that information can be read, at what accuracy, and at what cost.

The sequencing market is dominated by four technology families, each occupying a distinct niche. Illumina’s sequencing by synthesis (SBS) accounts for approximately 80% of published sequencing data due to its unmatched combination of throughput, scalability, and accuracy. Long-read technologies from PacBio and Oxford Nanopore are growing rapidly, driven by applications that short reads cannot resolve—structural variant detection, de novo genome assembly, and direct epigenetic modification detection. Sanger sequencing, while accounting for a small fraction of total bases sequenced, remains essential for targeted validation due to its >99.99% single-read accuracy.

The cost of sequencing has fallen by more than five orders of magnitude since the human genome project, enabling applications that were economically unthinkable a decade ago—from population-scale whole-genome studies to routine clinical diagnostics and agricultural genomics. However, lower costs have not eliminated the need for careful method selection. A poorly chosen sequencing strategy can waste budget on unnecessary depth or fail to detect the very variants the project aims to discover.

When to choose a sequencing service provider vs. building in-house capacity: For laboratories sequencing fewer than 50 human genomes per year or running fewer than 200 RNA-seq samples annually, outsourcing to a sequencing service provider is typically more cost-effective than purchasing and maintaining sequencing instruments. Service providers also offer access to multiple platforms, allowing method selection based on project needs rather than instrument availability. For high-throughput centers sequencing hundreds of samples monthly, in-house platforms may offer cost advantages and shorter turnaround times.

For researchers planning their first sequencing project or evaluating a new application, comprehensive NGS services provide expert guidance on method selection, experimental design, and data analysis.

The Core Parameters That Define Every Sequencing Method

Every sequencing technology can be described by six core parameters. Understanding these parameters is the foundation for comparing platforms and selecting the right one for a given application.

No platform excels at all six parameters simultaneously. The art of sequencing project design lies in identifying which parameters are most important for a specific application and selecting the platform that best matches those priorities.

How the six parameters interact in practice: Read length and throughput are inversely related on most platforms—longer reads mean fewer total reads per run. Accuracy and throughput also have a complex relationship: higher accuracy often requires more sequencing cycles or repeated measurements, which reduces net throughput. Input DNA requirement is frequently the binding constraint for precious clinical or environmental samples where only nanogram quantities are available. Understanding these trade-offs helps researchers avoid common planning mistakes, such as requesting 300 bp reads for a project that only needs 150 bp but would benefit from higher depth, or selecting a platform that requires 1 µg of DNA when only 10 ng is available.

Practical example of parameter trade-offs: Consider a researcher planning a human WGS project. They need high throughput (to sequence many samples cost-effectively), moderate read length (150 bp is sufficient for SNV detection), and high accuracy (Q30 or better). Illumina NGS is the obvious choice. Now consider a researcher assembling a novel plant genome. They need long reads (to span repetitive regions that may be >10 kb) and are willing to accept lower throughput and higher per-base cost. PacBio or Nanopore is the appropriate choice. The same parameters, weighed differently, lead to different platform conclusions. A third scenario: a researcher needs to validate 10 NGS-detected variants in a clinical context. They prioritize accuracy above all else and need results quickly. Sanger sequencing, despite its low throughput, is the correct choice.

Genomic data analysis services can help researchers navigate these trade-offs by providing computational resources that align with the selected platform’s output characteristics and by offering expert guidance on parameter optimization for specific project types.

Sanger Sequencing — The Gold Standard, Still Relevant for Targeted Work

The Sanger chain-termination method uses fluorescently labeled dideoxynucleotides (ddNTPs) to terminate DNA synthesis at specific positions, producing a ladder of fragments whose lengths reveal the sequence. Capillary electrophoresis separates these fragments with single-base resolution, generating a four-color electropherogram.

Where Sanger excels — and where it does not: Sanger sequencing delivers 600-1,000 bp per reaction at >99.99% accuracy. This makes it the method of choice for validating NGS-detected variants, confirming plasmid constructs, and sequencing small numbers of PCR products. However, Sanger cannot quantify allele frequencies in mixed samples, detect variants present at <15-20% frequency, or scale beyond a few hundred targets per project. For these applications, NGS is required despite its lower per-base accuracy.

For researchers needing high-accuracy confirmation of specific targets, Sanger sequencing services provide rapid and cost-effective validation.

Sanger’s role in the NGS era: Despite being a 40-year-old technology, Sanger sequencing has maintained its relevance through its unmatched accuracy. In clinical genomics, guidelines from organizations like the American College of Medical Genetics (ACMG) still recommend Sanger validation for certain classes of NGS-detected variants before reporting. For research laboratories, Sanger sequencing provides a rapid and inexpensive way to confirm that a library preparation or capture step worked correctly before scaling up to full NGS runs. The typical Sanger sequencing turnaround time—48-72 hours from sample submission to results—also makes it the method of choice for time-sensitive applications like plasmid verification in cloning workflows.

Figure 1. Sanger sequencing workflow — chain termination and capillary electrophoresis
Caption: Sanger sequencing workflow showing fluorescently labeled dideoxynucleotide chain termination, capillary electrophoresis fragment separation, and four-color electropherogram generation for 600-1,000 bp reads at >99.99% accuracy.

Illumina Next-Generation Sequencing — The High-Throughput Workhorse

Illumina’s sequencing by synthesis (SBS) technology performs millions of sequencing reactions in parallel on a single flow cell surface. DNA fragments are first ligated to platform-specific adapters, then amplified into clonal clusters on the flow cell surface through bridge amplification. In each sequencing cycle, a fluorescently labeled, reversibly terminated nucleotide is incorporated, imaged, and cleaved. The cycle repeats for the desired read length—typically 150 or 300 cycles.

Key performance characteristics: Illumina platforms span a 10,000-fold throughput range, from the MiniSeq (1 Gb per run) to the NovaSeq X Plus (16 Tb per run). Read lengths are 2 × 150 bp for most applications, up to 2 × 300 bp on MiSeq. Per-base accuracy exceeds 99.9% at Q30 for the latest XLEAP-SBS chemistry, which also reduces run time by 30-50% compared to standard SBS.

Coverage requirements by application — a practical guide: The required sequencing depth varies substantially by application. For human whole-genome sequencing (WGS), 30× coverage is standard for detecting germline single-nucleotide variants. For cancer somatic mutation detection, 60-100× is typically required to identify low-frequency variants. RNA-seq gene expression analysis requires 20-50 million reads per sample; isoform-level analysis requires 100+ million. For 16S amplicon sequencing, 10,000-50,000 reads per sample are sufficient for community profiling. These coverage targets directly affect platform selection and project budget.

How to calculate reads needed for a WGS project: The formula is straightforward: required reads = (genome size × desired coverage) ÷ read length. For a 3.2 Gb human genome at 30× with 2 × 150 bp reads (300 bp per fragment): (3.2 × 10⁹ × 30) ÷ 300 = 320 million reads per sample. Multiply by the number of samples and add 10-20% over-sequencing margin to account for failed samples or quality filters. A 100-sample WGS project therefore requires approximately 35-38 billion reads, placing it firmly on a NovaSeq-class platform.

Read type considerations: Paired-end sequencing (reading both ends of each fragment) provides more accurate alignment and enables detection of structural variants and fusion transcripts. Single-read sequencing is lower cost and may be sufficient for gene expression quantification where fragment orientation is not critical. Most modern sequencing projects use paired-end reads as the default, with single-read reserved for specific applications where the cost saving justifies the information loss.

The cluster density trap: One parameter that researchers frequently underestimate is the impact of cluster density on data quality. On a NovaSeq 6000 S4 flow cell, the optimal cluster density is 250-350 K clusters per mm². A 20% deviation in library loading concentration can shift cluster density by 30-50%, either underusing the flow cell or producing overlapping signals that reduce base-calling accuracy. Using qPCR-based library quantification—not Qubit—is the single most effective step to avoid this problem.

Illumina NGS is the most widely adopted platform for large-scale genomic projects. Whole genome sequencing services leverage the full Illumina platform range to match throughput to project requirements.

NGS platform selection within the Illumina family: The Illumina portfolio spans six active platforms with a 10,000-fold throughput range. The MiniSeq (1.2 Gb) and MiSeq (15 Gb) are suited for small-scale applications like targeted panels and 16S amplicon sequencing where read length and turnaround time are prioritized over total throughput. The NextSeq 2000 (330 Gb) is a mid-range workhorse for RNA-seq, exome sequencing, and small WGS projects. The NovaSeq 6000 (6 Tb) and NovaSeq X (16 Tb) are designed for large-scale WGS, population studies, and single-cell sequencing projects requiring billions of reads per run. Selecting the right platform within the family is as important as selecting between technology families.

Figure 2. Illumina SBS workflow — library preparation, cluster generation, and sequencing by synthesis
Caption: Illumina sequencing by synthesis workflow showing the three main stages—library preparation with adapter ligation, bridge amplification cluster generation on the flow cell surface, and the cyclic SBS process with fluorescent nucleotide incorporation and imaging.

Long-Read Sequencing — Resolving What Short Reads Cannot

Short-read NGS platforms produce reads of 150-300 bp, which are sufficient for most variant detection applications but cannot span repetitive regions, structural variants, or complex genomic rearrangements. Long-read sequencing technologies from PacBio and Oxford Nanopore address these limitations with reads exceeding 10 kb, and in some cases exceeding 2 Mb.

PacBio SMRT sequencing: Single Molecule, Real-Time sequencing uses zero-mode waveguides—nanoscale wells that confine light to a detection volume small enough to observe a single DNA polymerase. The instrument records fluorescence signals in real time as fluorescently labeled nucleotides are incorporated. Circular consensus sequencing (CCS) generates HiFi reads: 10-20 kb at >99.9% accuracy by sequencing the same molecule multiple times.

Oxford Nanopore sequencing: Nanopore sequencing passes a single DNA strand through a protein nanopore embedded in an electrically resistant membrane. Each nucleotide causes a characteristic change in ionic current, which is decoded by neural network base-callers. Read lengths routinely reach 10-100 kb, with ultra-long reads exceeding 2 Mb reported. Raw read accuracy is 90-97%, but consensus accuracy exceeds 99% with 30-50× coverage. Nanopore directly detects base modifications (5mC, 6mA) without chemical conversion—a unique capability.

Decision guide — when to choose long reads: Long-read sequencing is the method of choice for: de novo genome assembly (where short reads cannot resolve repetitive content), structural variant detection (deletions, insertions, and inversions spanning >50 bp), full-length transcript sequencing (capturing complete isoforms without computational assembly), and direct epigenetic modification detection (Nanopore only). For applications where short reads perform well—germline SNV detection, standard RNA-seq quantification—the higher per-base cost of long reads is not justified.

Researchers working on genome assembly or structural variant projects can access PacBio SMRT sequencing services and Nanopore sequencing services tailored to their specific requirements.

Figure 3. Long-read vs short-read sequencing — read length comparison and application suitability
Caption: Comparative visualization of short-read (150-300 bp) and long-read (10 kb – 2 Mb) sequencing technologies, showing how long reads span repetitive regions, structural variants, and full-length transcripts that short reads cannot resolve.

Platform Comparison — Head-to-Head by Key Metrics

How to read this table: The table shows that no single platform dominates all metrics. Illumina NGS leads in throughput and versatility for high-depth applications. PacBio HiFi leads in combined read length and accuracy. Nanopore leads in maximum read length and real-time capability. Sanger leads in per-base accuracy for short targets. The right choice depends entirely on which metrics are prioritized by the specific application.

For a more detailed comparison between the two long-read platforms, PacBio vs Oxford Nanopore provides an in-depth analysis.

Emerging applications that long reads unlock: Long-read sequencing is enabling research questions that were previously inaccessible. Telomere-to-telomere (T2T) genome assemblies, which require resolving complex repetitive regions near centromeres and telomeres, have been achieved for the human genome and multiple model organisms using Nanopore ultra-long reads combined with PacBio HiFi. Similarly, full-length RNA sequencing is revealing isoform diversity that was missed by short-read RNA-seq, with important implications for understanding alternative splicing in disease. Direct RNA sequencing on Nanopore platforms also detects RNA base modifications (m⁶A, pseudouridine) in native transcripts, providing a direct view of epitranscriptomic regulation.

Figure 4. Platform comparison radar chart — read length, throughput, accuracy, and application fit
Caption: Multi-dimensional radar chart comparing Sanger, Illumina NGS, PacBio HiFi, and Oxford Nanopore platforms across read length, throughput, per-base accuracy, consensus accuracy, input requirement, and run time.

Applications by Platform — Matching the Method to the Question

The following decision framework maps common sequencing applications to their optimal platform, based on the core parameters discussed above.

This framework is a starting point, not a rigid rule. Many projects benefit from hybrid approaches—for example, using Illumina short reads for depth and PacBio long reads for structural variant detection within the same study.

Hybrid sequencing strategies in practice: A growing trend in genomics is the use of complementary platforms within a single project. In de novo genome assembly, PacBio HiFi reads provide accurate long-range information for contig construction, while Illumina short reads fill gaps and polish the final assembly at lower cost than sequencing additional long reads alone. In structural variant studies, Nanopore long reads detect large deletions and insertions that short reads cannot span, while Illumina sequencing provides the depth needed for accurate breakpoint resolution. The cost of this dual-platform approach is higher than single-platform sequencing, but the data quality improvement is often substantial enough to justify the additional expense for high-priority projects. NGS services can help design integrated multi-platform strategies that balance data quality and budget.

Figure 5. Application-to-platform mapping — key sequencing applications and their optimal methods
Caption: Decision matrix mapping 11 common sequencing applications to their optimal platform (Illumina NovaSeq, NextSeq, MiSeq, PacBio HiFi, Oxford Nanopore, or Sanger) with recommended coverage depth and rationale for each selection.

The Sequencing Data Analysis Pipeline — From Signal to Variant

Raw sequencing data must be processed through a standardized analysis pipeline to produce interpretable results. The pipeline structure differs by platform but shares common stages:

1. Base calling: Raw instrument signals (fluorescence images for Illumina, ionic current for Nanopore, pulse data for PacBio) are converted into nucleotide sequences with quality scores. Illumina produces BCL files converted to FASTQ; Nanopore uses neural network base-callers like Dorado or Guppy.
2. Quality control: FASTQ files are assessed for per-base quality (Q-scores), GC content distribution, adapter contamination, and duplication rates. Tools like FastQC and MultiQC provide standardized summaries. Libraries with >30% duplication rates or <75% bases above Q30 typically indicate a problem in library preparation or sequencing run quality.
3. Alignment or assembly: Reads are either aligned to a reference genome or assembled de novo. Alignment is faster and more sensitive for variant detection in species with a high-quality reference. Assembly is necessary for novel genomes or when reference bias must be avoided. The choice between alignment and assembly also affects the required compute resources: aligning 30× human WGS data takes 6-12 hours on a 16-core server, while de novo assembly of the same data can take 24-72 hours and requires 128+ GB of RAM.
4. Variant calling and quantification: Aligned reads are analyzed for single-nucleotide variants (SNVs), insertions and deletions (indels), and structural variants. GATK is the standard for Illumina germline variant calling; FreeBayes and DeepVariant are alternatives. For RNA-seq, tools like Salmon, Kallisto, or STAR quantify transcript abundance.
5. Biological interpretation: Called variants are annotated with predicted functional effects, population frequencies, and literature associations. Final interpretation requires domain expertise to distinguish true biological signals from technical artifacts.

The computational resources required for this pipeline are substantial. A single human genome at 30× produces 100-200 GB of FASTQ data. A 100-sample WGS project requires 10-20 TB of storage and significant compute capacity. Bioinformatics analysis services can provide the necessary computational infrastructure and analysis expertise.

Figure 6. Sequencing data analysis pipeline — from raw signals to biological interpretation
Caption: Five-stage sequencing data analysis pipeline showing base calling from raw instrument signals, quality control of FASTQ files, alignment or de novo assembly, variant calling and quantification, and biological interpretation of annotated variants.

How to Plan a Sequencing Project — A Practical Checklist

Before starting a sequencing project, the following checklist helps identify the critical decisions that will determine success.

1. Define the question and the required data type: Single nucleotide variants, structural variants, gene expression levels, epigenetic modifications, or microbial community composition? Each requires a different sequencing strategy.
2. Calculate required coverage or read depth: Use published guidelines for the specific application. Add 20-30% margin to account for samples that may fail QC or produce lower-than-expected data.
3. Choose platform and flow cell type: Match total read requirement to platform capacity. A NovaSeq X 25B flow cell producing 12.5 billion reads is appropriate for a 250-sample WGS project at 30×; a MiSeq run producing 25 million reads is appropriate for a 96-sample 16S project.
4. Prepare samples with appropriate QC: Verify DNA concentration using fluorometric methods (Qubit), not UV spectrophotometry. Check integrity by gel electrophoresis or TapeStation. For RNA-seq, a RIN score ≥ 7 is generally required for mRNA libraries.
5. Plan data management before sequencing begins: Ensure sufficient storage, compute, and analysis pipeline are in place. Sequencing generates data faster than most laboratories expect—a NovaSeq run completing overnight produces data that must be processed the same day. A single run from a NovaSeq X 25B flow cell produces approximately 5-8 TB of raw data. For a 100-sample human WGS project, plan for 15-20 TB of total storage including FASTQ, BAM, VCF, and intermediate analysis files. Data transfer from the sequencing facility to the analysis environment must also be planned—uploading 10 TB of FASTQ files to a cloud compute environment can take 24-72 hours depending on network bandwidth.
6. Include controls: A positive control sample with known variants validates the workflow from extraction through analysis. A negative (no-template) control identifies contamination. For low-diversity libraries, add 5-20% PhiX control to calibrate the instrument’s optical system.

Following this checklist systematically minimizes the risk of expensive mid-project corrections. For research teams that prefer to outsource project management, end-to-end NGS services cover every step from experimental design through data delivery.

Figure 7. Sequencing project planning workflow — from experimental design through data delivery
Caption: Seven-step sequencing project planning checklist covering question definition, coverage calculation, platform selection, sample QC, data management planning, control inclusion, and the option to outsource project management to end-to-end NGS service providers.

FAQ

What is the difference between Sanger sequencing and NGS?

Sanger sequencing analyzes one DNA fragment at a time using chain-termination chemistry, producing 600-1,000 bp reads at >99.99% accuracy. NGS analyzes millions to billions of fragments simultaneously, producing shorter reads at slightly lower per-base accuracy but far higher throughput and lower per-base cost.

How do I choose between short-read and long-read sequencing?

Choose short-read sequencing (Illumina NGS) when the application requires high depth at low per-base cost and read length is adequate for the target—this covers most SNV detection, RNA-seq, and metagenomics projects. Choose long-read sequencing (PacBio or Nanopore) for de novo assembly, structural variant detection, and full-length transcript sequencing where short reads cannot span the regions of interest.

What coverage depth do I need for human whole-genome sequencing?

For germline SNV detection in human WGS, 30× coverage is the standard. For cancer somatic mutation detection, 60-100× is recommended to identify low-frequency variants. For rare variant detection, depths of 60× or higher may be needed.

How much does a sequencing project cost?

Costs depend on platform, coverage depth, sample number, and analysis requirements. Library preparation, sequencing reagents, platform access, data storage, and bioinformatics analysis are all components. Requesting a project-specific quote from a sequencing service provider is the most reliable way to estimate total project cost.

Can I combine data from different sequencing platforms in one project?

Yes. Hybrid approaches are common and often provide the best results. A typical strategy is Illumina short reads for depth and cost efficiency, supplemented by PacBio or Nanopore long reads for resolving complex genomic regions. Many published genome assemblies now use this combined approach.

What is the most common cause of sequencing project failure?

Inaccurate library quantification is the single most common cause. Libraries quantified by UV spectrophotometry frequently overestimate usable DNA concentration by 2-5×, leading to under-clustering or over-clustering on the flow cell. Using qPCR or fluorometric quantification avoids this problem.

How do I decide between PacBio and Nanopore for long-read sequencing?

Choose PacBio when high single-read accuracy is critical (HiFi reads at >99.9%) and the required read length is 10-20 kb. Choose Nanopore when ultra-long reads (>100 kb) are needed, when real-time data streaming is advantageous, or when direct detection of DNA base modifications is a primary goal.

What read length should I choose for RNA-seq?

For standard gene expression quantification by read counting, 2 × 50 bp or 2 × 75 bp is sufficient—the main requirement is that reads map uniquely to transcripts. For isoform detection or discovery of novel splice junctions, longer reads (2 × 150 bp) improve alignment accuracy. Full-length transcript sequencing requires long-read platforms such as PacBio Iso-Seq or Nanopore cDNA sequencing, which capture complete transcript isoforms in a single read without computational reconstruction.

How do I decide between whole-genome and whole-exome sequencing?

Whole-genome sequencing provides complete coverage of coding and non-coding regions. Whole-exome sequencing covers only the protein-coding regions (~1-2% of the genome) at higher depth. Choose WGS when non-coding variants, structural variants, or comprehensive coverage are needed. Choose WES when the focus is on coding variants and budget is constrained.

References:

1. DNA sequencing methods: from Sanger to NGS. INTEGRA Biosciences.
2. Next generation sequencing and beyond: a review. Functional & Integrative Genomics. 2025;25:1724.
3. The chemistry of next-generation sequencing. Nature Biotechnology. 2023;41:1709-1715.
4. Nanopore sequencing technology, bioinformatics and applications. Nature Biotechnology. 2021;39:1348-1365.
5. Accurate circular consensus long-read sequencing. Nature Biotechnology. 2019;37:1155-1162.