- Single-molecule sequencing is a DNA or RNA sequencing approach that determines nucleotide sequences by analysing individual nucleic acid molecules rather than relying on large populations of amplified DNA fragments. It represents an important development in modern genomics because it can provide long sequence reads, reveal complex genomic structures, and preserve information about individual DNA or RNA molecules. Unlike many conventional short-read sequencing technologies, single-molecule methods can often sequence native DNA or RNA directly without requiring PCR amplification of the target molecules.
- The fundamental principle of single-molecule sequencing is the detection of molecular events associated with individual nucleotides as they are identified or incorporated into a nucleic acid strand. Different technologies use different detection mechanisms. Pacific Biosciences (PacBio) sequencing monitors DNA polymerase activity in real time, whereas Oxford Nanopore sequencing detects changes in electrical current as a DNA or RNA molecule passes through a nanopore. These approaches allow researchers to obtain long reads and investigate genomic regions that are difficult to analyse using conventional short-read sequencing.
- Traditional sequencing methods often require amplification to generate enough DNA material for detection. For example, Illumina sequencing uses bridge amplification to generate clusters of DNA fragments on a flow cell. Single-molecule sequencing instead aims to detect signals originating from individual DNA or RNA molecules. This can reduce amplification-associated biases and preserve certain forms of molecular information, although individual technologies may still use amplification or other preparatory steps for particular applications. The term single-molecule sequencing therefore describes the detection strategy rather than one specific laboratory procedure.
- One of the major advantages of single-molecule sequencing is its ability to produce long reads. Depending on the technology, library preparation, sample quality, and sequencing configuration, reads may extend from several thousand bases to tens or even hundreds of thousands of bases. These long reads can span repetitive regions, structural rearrangements, complex genomic loci, and multiple variants located on the same DNA molecule. This information can improve genome assembly, haplotype reconstruction, transcript analysis, and the investigation of genomic variation.
- Pacific Biosciences sequencing, commonly known as PacBio sequencing, is based on a technology called single-molecule real-time (SMRT) sequencing. It observes DNA synthesis by an individual DNA polymerase molecule in real time. The sequencing reaction takes place in extremely small observation chambers known as zero-mode waveguides (ZMWs). Each ZMW is designed to allow the detection of fluorescent signals generated when a DNA polymerase incorporates labelled nucleotides into a growing DNA strand.
- In PacBio SMRT sequencing, DNA molecules are prepared as circular templates using specially designed adapters. These circular templates are commonly called SMRTbell templates. The adapters connect the two ends of a DNA fragment, creating a circular DNA molecule. A DNA polymerase binds to the template and repeatedly moves around the circular molecule, synthesizing complementary DNA strands. The circular structure allows the polymerase to pass through the same insert multiple times, depending on the template design and polymerase performance.
- The SMRTbell library is loaded into a sequencing system containing ZMWs. A ZMW is a nanoscale optical structure that restricts the observation volume, allowing fluorescent signals from nucleotide incorporation to be detected against background fluorescence. A DNA polymerase molecule is positioned within the observation region, and fluorescently labelled nucleotides are supplied. When the polymerase incorporates a nucleotide into the growing DNA strand, a characteristic fluorescent signal is detected. The label is subsequently cleaved during incorporation, allowing the polymerase to continue DNA synthesis.
- The sequence of fluorescent signals identifies the nucleotides incorporated by the polymerase. Because DNA synthesis is monitored continuously, the method is called real-time sequencing. The instrument records the timing and identity of nucleotide incorporation events and converts these signals into a nucleotide sequence through computational base calling. The resulting data can provide long reads from individual DNA molecules without requiring the conventional cluster amplification used by Illumina sequencing.
- An important development in PacBio sequencing is circular consensus sequencing (CCS). In this approach, the DNA polymerase reads the same circular template multiple times, generating repeated observations of the original DNA insert. These observations are combined computationally to produce a consensus sequence. Repeated measurements help reduce random sequencing errors and improve base accuracy. High-quality consensus reads generated through this process are commonly known as HiFi reads.
- PacBio HiFi sequencing combines relatively long reads with high consensus accuracy. It is particularly useful for de novo genome assembly, haplotype-resolved genome reconstruction, detection of structural variants, and analysis of complex genomic regions. The number of passes through a circular template depends on the insert length, polymerase performance, and sequencing conditions. Longer inserts generally allow fewer complete passes within a given sequencing period, whereas shorter inserts may permit more passes and potentially stronger consensus accuracy.
- Another major single-molecule sequencing technology is Oxford Nanopore sequencing. It determines nucleotide sequences by measuring changes in electrical current as DNA or RNA molecules pass through nanoscale pores embedded in a membrane. These pores are biological or engineered protein structures that allow individual nucleic acid molecules to pass through under controlled conditions. A voltage is applied across the membrane, producing an ionic current through each nanopore.
- When a DNA or RNA molecule enters a nanopore, the nucleotides occupying the sensing region influence the flow of ions through the pore. Different nucleotide combinations produce characteristic changes in the measured current. The instrument continuously records these electrical signals, and computational algorithms interpret them to determine the nucleotide sequence. This process is known as base calling.
- Nanopore sequencing differs from PacBio sequencing in its signal-detection mechanism. PacBio detects fluorescent signals associated with DNA polymerase activity, while Oxford Nanopore measures changes in ionic current as nucleic acids move through a pore. Both technologies can generate long reads, but their sequencing chemistry, signal processing, library preparation, and error characteristics are different.
- Oxford Nanopore sequencing can analyse DNA and, using suitable protocols, native RNA molecules. Direct RNA sequencing is particularly valuable because it can investigate RNA molecules without first converting them into complementary DNA. It can preserve information about RNA modifications and transcript structure, although detection and interpretation depend on the sequencing chemistry, base-calling models, and analytical methods used. Nanopore sequencing can also detect certain modified DNA bases through changes in electrical current, allowing researchers to investigate epigenetic features alongside nucleotide sequence.
- Library preparation for single-molecule sequencing depends on the platform and the research objective. In PacBio workflows, DNA is prepared as SMRTbell templates by attaching circularizing adapters to the ends of DNA fragments. The resulting library is purified and prepared for loading into the sequencing system. In Oxford Nanopore workflows, sequencing adapters and motor proteins are attached to DNA or RNA molecules. The motor protein regulates the movement of the nucleic acid through the nanopore, allowing the instrument to record a sequence of electrical signals suitable for computational interpretation.
- The starting material must be of suitable quality for reliable sequencing. Long-read applications often benefit from high-molecular-weight DNA because extensive fragmentation can reduce the maximum achievable read length. DNA extraction procedures should therefore minimize mechanical damage and degradation when long intact molecules are required. RNA sequencing also requires careful handling because RNA is more susceptible to degradation. The appropriate extraction and library preparation methods depend on the organism, sample type, sequencing platform, and intended application.
- After sequencing, the instrument generates raw signal data. In PacBio systems, raw data include information about polymerase activity and fluorescent nucleotide incorporation. In Oxford Nanopore systems, raw data contain electrical current measurements recorded as molecules pass through nanopores. Base-calling software converts these signals into nucleotide sequences and associated quality information. The resulting reads are commonly stored in FASTQ format, while additional platform-specific files may preserve raw signals and processing metadata.
- A simplified FASTQ record is shown below:
@READ_001
ATGCTAGCTAGCTACGATCG
+
IIIIIIIIIIIIIIIIIIII
- The first line contains the read identifier, the second contains the nucleotide sequence, the third is a separator, and the fourth stores quality characters encoded using the Phred scale. Actual records contain platform-specific identifiers and quality values. Retaining the original signal data, where available and appropriate, can be important because improved base-calling models may allow researchers to reanalyse the signals in the future.
- Phred quality scores express the estimated probability of an incorrect base call using the relationship Q = −10 log10(P), where P represents the estimated probability of error. A Q20 score corresponds to an estimated error probability of 1%, while Q30 corresponds to an estimated error probability of 0.1%. Quality characteristics differ between sequencing platforms and technologies. PacBio HiFi reads can achieve high consensus accuracy through repeated observations of circular templates, while nanopore read accuracy depends on pore chemistry, signal quality, base-calling models, and the sequencing configuration.
- Bioinformatics quality control is an important step in single-molecule sequencing analysis. Researchers commonly assess read length distribution, read quality, sequencing yield, adapter contamination, and other platform-specific characteristics. NanoPlot and NanoStat are widely used for visualizing and summarizing Oxford Nanopore sequencing data. pycoQC can generate quality-control reports for suitable nanopore datasets. NanoFilt can filter reads according to quality or length. For PacBio data, platform-specific tools such as SMRT Link and pbtk support processing and analysis of sequencing output. MultiQC can combine compatible reports from multiple tools and samples.
- An example of generating a summary of Oxford Nanopore reads using NanoPlot is:
NanoPlot \
–fastq sample.fastq.gz \
–outdir nanoplot_results
NanoFilt can be used for quality-based filtering in suitable workflows:
gunzip -c sample.fastq.gz | \
NanoFilt -q 10 | \
gzip > sample.filtered.fastq.gz
- Filtering thresholds should be selected according to the sequencing technology and analytical objective. Removing low-quality reads may improve downstream analysis, but aggressive filtering can eliminate biologically valuable information, particularly when the data contain very long reads or low-abundance sequences.
- Single-molecule sequencing is especially valuable for de novo genome assembly. In de novo assembly, sequencing reads are reconstructed into longer genomic sequences without depending entirely on an existing reference genome. Long reads can span repetitive regions and connect genomic segments that short reads cannot resolve. This can produce more continuous genome assemblies and improve the reconstruction of complex regions, although assembly quality still depends on coverage, read accuracy, genome complexity, heterozygosity, and other factors.
- Common long-read assembly tools include Flye, Canu, hifiasm, and Shasta. Flye supports long-read genome assembly and can be used with suitable nanopore or PacBio datasets. Canu was developed for long-read assembly and includes read correction and assembly procedures. hifiasm is widely used for high-quality PacBio HiFi genome assembly and can support haplotype-resolved assembly. Shasta is designed for efficient assembly of suitable nanopore datasets.
- An illustrative Flye command for assembling Oxford Nanopore reads is:
flye \
–nano-raw sample.fastq.gz \
–out-dir flye_assembly \
–threads 16 \
–genome-size 5m
- The genome-size parameter should be adjusted to the expected genome size of the organism. Flye options and input modes depend on the read type and software version, so the documentation should be consulted before using a command in a production workflow. Assembly quality can subsequently be assessed using tools such as QUAST and BUSCO. QUAST reports assembly statistics, while BUSCO evaluates the presence of expected conserved genes for an appropriate lineage dataset.
- Another important application is structural variant detection. Structural variants include large insertions, deletions, inversions, duplications, and translocations. Short-read sequencing may have difficulty resolving these events because individual reads often cannot span the affected genomic regions. Long reads can provide direct evidence across structural changes, making them valuable for detecting complex genomic rearrangements.
- Tools such as Sniffles2, cuteSV, and SVIM support structural variant detection from suitable long-read alignments. Reads are generally aligned to a reference genome using long-read aligners such as minimap2. The resulting alignments are examined for unusual mapping patterns, split alignments, insertions, deletions, and other evidence of structural variation. Variant calls are often stored in VCF format and may require filtering, annotation, and validation.
- An example of mapping long reads to a reference genome using minimap2 is:
minimap2 \
-ax map-ont \
reference_genome.fa \
sample.fastq.gz \
> sample.sam
- The map-ont preset is intended for Oxford Nanopore genomic reads. For PacBio HiFi data, a different preset, such as map-hifi, should be used. The alignment strategy should match the platform and read type rather than treating all long reads as equivalent.
- Single-molecule sequencing also supports haplotype phasing. Haplotype phasing determines which genetic variants occur together on the same chromosome. Long reads can span multiple heterozygous variants, providing direct evidence that particular variants belong to the same DNA molecule. This is useful in studies of inherited diseases, population genetics, genome evolution, and complex genomic regions. Tools such as WhatsHap and LongPhase can support phasing using suitable sequencing data.
- Transcriptomics is another important application. Long-read sequencing can capture complete or near-complete RNA-derived transcript sequences, providing information about alternative splicing, transcript isoforms, alternative transcription start sites, and alternative polyadenylation. Short-read RNA sequencing often requires computational reconstruction of transcripts from many smaller fragments, whereas long reads can provide more direct evidence of transcript structure.
- PacBio Iso-Seq is a long-read transcript sequencing approach based on circular consensus sequencing of cDNA molecules. It can generate highly accurate full-length transcript reads and is used to characterize transcript isoforms and gene structures. Oxford Nanopore sequencing can also be used for cDNA sequencing and direct RNA sequencing. Tools such as Iso-Seq analysis software, FLAIR, TALON, StringTie2, and Bambu support different aspects of long-read transcript processing and analysis. The choice of tool depends on the sequencing platform, library type, and research objective.
- Single-molecule sequencing has important applications in epigenomics because some platforms can detect chemical modifications of DNA or RNA. PacBio polymerase kinetics can provide information about certain modified bases, while nanopore sequencing can detect changes in electrical current associated with modified nucleotides. These signals can be analysed using specialized computational models to investigate DNA methylation and other modifications. However, modification detection is dependent on the platform, chemistry, training data, and validation procedures. The presence and identity of a modification should not be inferred from raw signal changes without appropriate analytical support.
- In microbial genomics, long-read sequencing can improve the assembly of bacterial and archaeal genomes, including regions containing repetitive sequences, mobile genetic elements, and plasmids. It can help reconstruct complete or near-complete microbial chromosomes and investigate the organization of antimicrobial resistance genes. In metagenomics, long reads can improve the association of genes with microbial genomes and may help resolve complex microbial communities. However, low-abundance organisms, uneven sequencing depth, and contamination can still complicate metagenomic assembly and taxonomic interpretation.
- Single-molecule sequencing is also valuable in human genetics, cancer genomics, plant genomics, animal breeding, and evolutionary research. In human genetics, it can support the investigation of structural variants, repeat expansions, complex disease-associated regions, and difficult-to-sequence genes. In cancer research, it can help characterize genomic rearrangements and complex tumour genomes. In plants and animals, long reads can improve reference genome assemblies, investigate gene-family organization, identify structural variation, and support studies of genetic diversity and adaptation.
- Despite its advantages, single-molecule sequencing has limitations. The performance of individual technologies varies, and raw read accuracy, read length, throughput, sequencing cost, and sample requirements must be considered when selecting a platform. Although PacBio HiFi reads provide high consensus accuracy, the number of passes through a circular template and the resulting read length depend on insert size and sequencing conditions. Nanopore sequencing can produce very long reads, but raw signal interpretation and base-calling accuracy depend on the pore chemistry, sequencing system, and computational models.
- Another limitation is the requirement for high-quality DNA when very long reads are desired. DNA fragmentation during extraction or handling can reduce read lengths and limit the ability to resolve long genomic regions. Library preparation can also introduce biases, and insufficient sequencing coverage may result in incomplete assemblies or missed variants. In RNA sequencing, transcript degradation and incomplete reverse transcription can affect the representation of full-length transcripts.
- Single-molecule sequencing does not eliminate all sequencing errors or analytical uncertainty. Base-calling errors, alignment errors, repetitive sequences, sample contamination, and limitations in reference genomes can affect downstream results. Different platforms have different error profiles, so analysis pipelines should be selected and validated for the specific data type. Important variants may require confirmation using an independent method, particularly when results are used in clinical or other high-stakes settings.
- Hybrid sequencing approaches combine short-read and long-read data to improve genome analysis. Illumina short reads can provide high base accuracy and substantial sequencing depth, while PacBio or Oxford Nanopore long reads provide information about long-range genomic structure. Combining these data can improve genome assembly, variant detection, and error correction in suitable workflows. The best strategy depends on the biological question, sample quality, available resources, and required level of resolution.
- For reproducible single-molecule sequencing analysis, researchers should record the sample source, DNA or RNA extraction method, library preparation protocol, sequencing platform, chemistry, read configuration, and sequencing depth. They should preserve raw signal files when available, base-called FASTQ files, quality-control reports, reference genome accession and version, software versions, and analysis parameters. Genome assemblies and variant calls should be accompanied by appropriate quality assessments and metadata. These practices support independent evaluation, comparison between studies, and future reanalysis.
- Single-molecule sequencing has expanded the possibilities of modern bioinformatics by enabling the direct analysis of individual DNA and RNA molecules. PacBio SMRT sequencing uses real-time fluorescent detection of DNA polymerase activity, while Oxford Nanopore sequencing identifies nucleic acids through changes in electrical current as molecules pass through nanopores. Both approaches provide long-read information that is valuable for genome assembly, structural variant detection, haplotype phasing, transcriptomics, epigenomics, and microbial genomics. Understanding their principles, workflows, computational methods, applications, and limitations is essential for selecting suitable sequencing technologies and interpreting modern genomic data.