Sequencing Technologies 101: How Machines Read DNA
Under scientific review
This introductory page is under scientific review. Use it for orientation, not clinical decision-making.
DNA sequencing digitizes biological code into digital text. However, no single hardware platform excels at all parameters: read length, accuracy, throughput, cost, and speed. For software engineers, choosing a sequencing technology is analogous to selecting a database or network protocol — balancing trade-offs between throughput, latency, and reliability.
1. The Generation Evolution
1st Gen (Sanger) --> 2nd Gen (Illumina SBS) --> 3rd Gen (Nanopore / PacBio)
Low throughput, exact High throughput, short reads Long reads, real-time streamingSanger Sequencing (1st Generation) — The Debugger
- Mechanism: Chain termination using dideoxynucleotides (ddNTPs).
- Read Length: ~800 - 1000 bp.
- Accuracy: Extremely high (>99.99%).
- CS Analogy: Single-threaded step-through debugging. Excellent for validating single point mutations or small plasmids, but unfeasible for whole human genomes ($3 \times 10^9$ bases).
Illumina / Short-Read (2nd Generation) — Massively Parallel Batch Packets
- Mechanism: Sequencing-by-Synthesis (SBS). Fluorescently labeled reversible terminators are added one base at a time. High-resolution cameras capture the color signal at each spot across millions of clusters.
- Read Length: Short (150 bp to 300 bp paired-end).
- Accuracy: Very high (Q30 standard = 99.9% accuracy).
- CS Analogy: Splitting a 3-billion-character document into millions of 150-character micro-chunks, reading them all in parallel, and relying on high-performance alignment software to reconstruct the document.
- Role in Oncology: The dominant workhorse for clinical panels (FoundationOne), Whole-Exome Sequencing (WES), and liquid biopsy ctDNA detection.
Oxford Nanopore Technologies (3rd Generation) — Network Packet Streaming
- Mechanism: A synthetic lipid membrane embeds protein nanopores. An electric voltage drives a single DNA strand through the pore. As nucleotides pass through, they characteristic disrupt the ionic current.
- Read Length: Ultra-long (10 kb to >1 Megabase per read).
- Accuracy: Historically lower (~95-98%), now reaching >99% (Q20+) with R10.4 pores and duplex sequencing.
- CS Analogy: Deep Packet Inspection (DPI) on a continuous byte stream over a network socket. Basecalling (using Transformer/RNN neural networks) runs in real-time as the strand passes through the hardware.
- Role in Oncology: Resolving complex structural variants, chromosomal fusions, repetitive genomic regions, and real-time intraoperative methylome profiling (e.g., rapid brain tumor subtyping during surgery).
PacBio HiFi (3rd Generation) — Redundant Checksum Long Reads
- Mechanism: Single-Molecule Real-Time (SMRT) sequencing. A circularized DNA molecule is read repeatedly by a DNA polymerase in a micro-well (ZMW).
- Read Length: Long (15 - 20 kb).
- Accuracy: Extremely high (>99.9% / Q30+), because the machine reads the same single molecule 10-20 times in a circle and builds a consensus sequence (HiFi reads).
- CS Analogy: Sending data packets with high-redundancy forward error correction (FEC), yielding both long context and zero noise.
- Role in Oncology: Phasing haplotypes (distinguishing maternal vs. paternal alleles), resolving complex gene duplications (e.g., CYP2D6 drug metabolism genes), and structural variant discovery.
2. Platform Comparison Matrix
| Parameter | Sanger (1st Gen) | Illumina SBS (2nd Gen) | Oxford Nanopore (3rd Gen) | PacBio HiFi (3rd Gen) |
|---|---|---|---|---|
| Read Length | 800 - 1,000 bp | 150 - 300 bp | 10 kb - 1+ Mb | 15 - 20 kb |
| Primary Error Type | Rare | Substitution (low) | Insertion/Deletion (Indel) | Rare |
| Throughput / Run | Single target | Up to 6 Terabases (NovaSeq X) | Up to 150 Gigabases (PromethION) | Up to 500 Gigabases (Revio) |
| Data Output | Chromatogram (.ab1) | FASTQ / BAM | FASTQ / POD5 / BAM | FASTQ / BAM |
| Primary Use Case | Validation | Clinical SNV/Indel panels | Structural Variants & Real-time | De novo assembly & Phasing |
3. Selecting the Right Platform in Oncology
- Detecting low-frequency somatic mutations (VAF < 1%) in liquid biopsy: Choose Illumina (ultra-deep short-read sequencing at 10,000x coverage).
- Identifying large chromosomal translocations or gene fusions: Choose Nanopore or PacBio (short reads fail to span repetitive fusion breakpoints).
- Rapid intraoperative tumor diagnosis: Choose Nanopore MinION (portable USB sequencer delivering results in < 3 hours).
References
- Goodwin, S., McPherson, J. D., & McCombie, W. R. (2016). Coming of age: ten years of next-generation sequencing technologies. Nature Reviews Genetics, 17(6), 333-351.
- Logsdon, G. A., Vollger, M. R., & Eichler, E. E. (2020). Long-read sequencing interfaces for human genomics. Nature Reviews Genetics, 21(10), 597-614.
- Wenger, A. M., et al. (2019). Accurate circular consensus long-read sequencing improves variant detection and assembly of a human genome. Nature Biotechnology, 37(10), 1155-1162.