Long-Read Sequencing Technologies

Core Idea

Long-read sequencing platforms (PacBio and Oxford Nanopore) produce reads of 10,000 to over 100,000 base pairs, overcoming the fundamental limitation of short-read technologies in resolving repetitive regions, structural variants, and complex genomic rearrangements. PacBio's HiFi mode generates highly accurate long reads (~99.9% at 15-20 kb) through circular consensus sequencing. Oxford Nanopore sequences single DNA or RNA molecules in real time by measuring current changes as they pass through a protein nanopore, enabling ultra-long reads (>1 Mb) and direct detection of base modifications without bisulfite conversion. These technologies have enabled telomere-to-telomere genome assemblies, comprehensive structural variant detection, and full-length transcript sequencing.

How It's Best Learned

Compare assemblies of the same genome using short reads alone versus long reads alone versus a hybrid approach. Examine how repeat-rich regions (centromeres, segmental duplications) that were gaps in the short-read assembly become resolved with long reads. Then visualize structural variant calls from long reads and see how many were invisible to short-read analysis.

Common Misconceptions

• Long-read sequencing is not universally better than short-read sequencing — short reads remain more cost-effective for applications like RNA-seq quantification and variant calling in non-repetitive regions where read length does not matter.
• Oxford Nanopore's raw error rate (~5-10%) does not mean the final results are inaccurate — consensus from multiple reads or correction with short reads brings accuracy to 99.99%+.

Explainer

The Illumina sequencing revolution made genomics affordable, but it introduced a fundamental limitation: read lengths of 150-300 bp cannot resolve genomic features longer than themselves. Repetitive elements (which comprise half the human genome), structural variants, full-length transcript isoforms, and base modifications all require longer reads to study comprehensively. Third-generation long-read sequencing, pioneered by Pacific Biosciences and Oxford Nanopore Technologies, addresses these limitations with fundamentally different approaches to reading DNA.

PacBio sequencing uses single-molecule real-time (SMRT) sequencing: a DNA polymerase is fixed at the bottom of a tiny well (zero-mode waveguide), and fluorescently labeled nucleotides are incorporated in real time, with each incorporation producing a light pulse that identifies the base. The original continuous long-read (CLR) mode produced reads averaging 10-20 kb but with 10-15% error rate. The breakthrough came with HiFi (high-fidelity) reads: the template DNA is circularized, and the polymerase reads around the circle multiple times. Consensus across multiple passes of the same molecule reduces the error rate to ~0.1% (Q30), while maintaining read lengths of 15-20 kb. HiFi reads combine the two properties that were previously mutually exclusive: long length and high accuracy.

Oxford Nanopore takes a radically different approach. A single-stranded DNA (or RNA) molecule is ratcheted through a protein nanopore embedded in a synthetic membrane. As each base passes through the pore, it modulates the ionic current flowing through the pore. A neural network base-caller translates the raw current signal into a nucleotide sequence. Read length is limited only by the input DNA fragment length — reads of 1 Mb+ have been demonstrated, and typical reads are 10-100 kb. The raw error rate is higher (~5-10%) than HiFi, but newer base-callers and consensus approaches are rapidly improving accuracy. A unique advantage is direct modification detection: because the current signal is affected by base modifications (5-methylcytosine, 6-methyladenine), Nanopore can detect epigenetic marks on native DNA without bisulfite treatment or antibody enrichment.

These technologies have transformed several areas of genomics. De novo assembly benefits most dramatically: HiFi reads produce assemblies with N50s of tens of megabases, compared to tens of kilobases for short-read assemblies of the same genome. The T2T Consortium used long reads to complete the first gapless human genome, adding 200 Mb of sequence (including centromeres and short arms of acrocentric chromosomes) that were missing from GRCh38. Structural variant calling reveals the full spectrum of genomic variation, with long reads detecting thousands of SVs invisible to short-read methods. Full-length transcript sequencing (PacBio Iso-Seq, Nanopore direct RNA) captures complete isoforms without assembly, resolving alternative splicing and gene fusion events at single-molecule resolution. The tradeoff remains cost: per-base, long-read sequencing is more expensive than Illumina, though the gap narrows continuously.