Intermediatemetabolism
DNA Sequencing: From Sanger to Next-Generation
Overview
DNA sequencing — reading the exact order of bases — underlies modern genetics, from the Human Genome Project to personalized medicine. This guide compares the two eras: Sanger sequencing, the chain-termination method that read the first genomes with its elegant dideoxy chemistry and fluorescent colors, and next-generation sequencing (NGS), which parallelizes millions of reactions to read billions of bases cheaply. You will learn how each works, what each is best for, and how sequencing costs fell faster than Moore's law.
Learning objectives
- Explain Sanger sequencing: dideoxy termination and fragment separation.
- Explain next-generation sequencing by synthesis (Illumina).
- Compare Sanger vs NGS in read length, throughput, cost, and uses.
- Describe how the Human Genome Project used these technologies.
Structure and function
| Part | Structure | Function |
|---|---|---|
| Dideoxynucleotide (ddNTP) | Nucleotide lacking the 3'-OH group; its incorporation terminates chain growth. | The Sanger trick: random termination at each base creates nested fragments revealing the sequence. |
| Sanger sequencing | Chain-termination method: ddNTPs randomly stop synthesis; fragments separated by size, terminal base identified. | First-generation sequencing: ~800 bp reads, gold standard for accuracy, ideal for single genes. |
| Sequencing by synthesis (Illumina) | NGS method: millions of clusters amplified on a flow cell, fluorescent reversible terminators added one base at a time. | Massively parallel: billions of short reads per run — the engine of modern genomics. |
| Read | A single sequenced DNA fragment's base string; Sanger reads are long (~800 bp), NGS reads short (~150 bp). | The data unit: genomes are assembled by overlapping many reads. |
| Coverage | Average number of reads covering each genome position. | Quality metric: 30× coverage means each base read ~30 times, suppressing random errors. |
| Human Genome Project | International effort (1990-2003) producing the first human genome sequence. | Sanger-based, ~$3 billion; proved large-scale sequencing possible and launched the genomic era. |
Diagram: two eras side by side. SANGER: a template with primer, polymerase extending; ddNTPs (colored: ddATP green, ddTTP red, ddCTP blue, ddGTP yellow) randomly terminate chains → nested fragments of every length → capillary electrophoresis separates by size → chromatogram peaks in four colors read 5'→3' as the sequence. NGS: a flow-cell surface with millions of DNA clusters; each cycle adds one fluorescent reversible-terminator base to all clusters simultaneously, camera photographs colors, terminator cleaved, repeat → billions of short reads feeding into a computer assembling a genome. A cost curve plunges from 1000s.
Sanger sequencing: elegant termination
Frederick Sanger's method (1977) is conceptually beautiful. A primer is extended on the template in the presence of normal dNTPs plus a small fraction of dideoxynucleotides (ddNTPs) — each labeled with a base-specific fluorescent color. Whenever a ddNTP (lacking the 3'-OH needed for extension) is incorporated, that chain stops. The reaction yields nested fragments terminating at every position; capillary electrophoresis separates them by size (single-base resolution), and reading the colors from shortest to longest spells the sequence. ~800 high-quality bases per run — the gold standard that sequenced the first genomes.
Normal dNTPs extend; ddNTPs terminate. Random incorporation at low frequency guarantees fragments ending at every template position — the complete nested set that reveals the sequence when size-ordered.
Next-generation sequencing: parallelism at scale
NGS (notably Illumina sequencing by synthesis) trades read length for staggering parallelism. DNA is fragmented, adapters ligated, and millions of fragments amplified into clusters on a flow cell. Then, each cycle adds one fluorescent reversible-terminator nucleotide to every cluster simultaneously — the color photographed, the terminator cleaved, repeat ~150 times. One run yields billions of short reads (~150 bp). Short reads are individually modest, but overlapping them computationally reconstructs entire genomes — shotgun assembly at industrial scale.
Unlike Sanger's permanent ddNTPs, Illumina's terminators block extension for one cycle then are chemically removed — allowing stepwise, photographed synthesis. Reversibility is what makes cyclic reading possible.
Head-to-head: when to use which
Sanger: reads ~800 bp, extremely accurate per base, cheap per reaction but expensive per megabase — ideal for verifying single genes, plasmids, and clinical confirmation of variants. NGS: reads ~150 bp × billions, cheap per megabase, requires heavy computation — ideal for whole genomes, exomes, tumor profiling, and microbiomes. Modern practice combines them: NGS discovers, Sanger confirms. Third-generation long-read methods (PacBio, Oxford Nanopore) now add 10,000+ bp reads, resolving repeats and structural variants the short reads miss.
The eternal trade-off: Sanger reads far but few; NGS reads short but billions. No single method optimizes both — method choice follows the question: one gene or one genome?
The Human Genome Project and the cost collapse
The HGP (1990-2003) used Sanger sequencing across international centers to read the ~3 billion base human genome — 13 years, roughly 3 billion per genome toward ~$1,000. That collapse enabled cancer genome profiling, rare-disease diagnosis, pandemic variant tracking, and direct-to-consumer genetics — the genomic era is fundamentally a sequencing-cost story.
Reading bases was step one; interpreting them — finding genes, regulatory elements, disease variants — is the ongoing work of genomics and bioinformatics. The sequence is the text; functional genomics writes the annotations.
Key points
- Sanger: dideoxy (ddNTP) chain termination → nested fragments → size separation → color-read sequence; ~800 bp, gold-standard accuracy.
- NGS (Illumina): millions of clusters, cyclic fluorescent reversible-terminator synthesis → billions of ~150 bp reads, assembled computationally.
- Sanger for single genes/verification; NGS for genomes/exomes/tumor profiling; long-read third generation resolves repeats.
- HGP (1990-2003): Sanger-based, ~$3B, 13 years → NGS collapsed costs toward ~$1,000/genome, enabling modern genomics.
Common misconceptions
✓ Truth: Sanger remains the clinical gold standard for verifying variants and sequencing single genes — unmatched per-base reliability for targeted questions. NGS dominates discovery and scale; Sanger dominates confirmation. They complement, not replace.
✓ Truth: Sequencing produces the letter string; interpretation — gene functions, regulatory logic, variant effects — is the harder, ongoing science of genomics. The HGP's sequence was the beginning of understanding, not the end.
✓ Truth: Every method reads fragments (reads) that are computationally assembled. Sanger reads ~800 bp pieces; NGS reads ~150 bp pieces by the billions. 'The genome sequence' is always a reconstruction from overlapping fragments.
Practice questions
Q1. Sanger sequencing determines the base sequence by:
- Measuring DNA mass directly
- Generating nested fragments terminated by ddNTPs and reading their terminal colors by size
- Photographing the double helix
- Counting hydrogen bonds
Random ddNTP incorporation creates fragments ending at every position; size separation orders them shortest-to-longest, and each fragment's fluorescent terminal ddNTP reveals that position's base — the sequence read directly.
Q2. The main advantage of NGS over Sanger is:
- Longer individual reads
- Massively parallel throughput: billions of reads per run at low cost per base
- No need for computers
- Higher per-base accuracy
NGS parallelizes millions of sequencing reactions, collapsing cost per megabase — enabling whole genomes. Sanger still wins on read length and per-base simplicity; NGS wins on scale.
Q3. In the Human Genome Project, the ~$3 billion cost primarily reflected:
- Sanger sequencing's serial, labor-intensive scale across 3 billion bases
- Computer hardware prices
- Patent licensing fees
- Sample collection from millions of people
Reading 3 billion bases with ~800-bp Sanger reads required enormous serial effort across years and centers. NGS's parallel chemistry later collapsed this cost by orders of magnitude — the defining technology shift in genomics.
Click a question to reveal the answer and explanation.
Frequently asked questions
Why do ddNTPs terminate synthesis instead of just slowing it?
DNA polymerase requires the 3'-OH group to attack the next incoming nucleotide and form the phosphodiester bond. ddNTPs lack this 3'-OH (both 2' and 3' positions are hydrogen), so no further extension is chemically possible — the chain is permanently capped at that point.
How can 150-bp NGS reads reconstruct a 3-billion-bp genome?
Through massive overlap: billions of reads sampled randomly cover each position ~30× (coverage). Assembly algorithms find overlaps between reads and stitch them into contigs, like reconstructing a shredded book from billions of overlapping snippets — redundancy resolves ambiguity.
Why did sequencing costs fall faster than Moore's law?
The shift from serial (Sanger capillaries, one reaction at a time) to massively parallel chemistry (millions of clusters imaged simultaneously) changed the scaling physics — throughput per run exploded while reagent costs per base plummeted. Parallelism, not miniaturization alone, drove the collapse.
When would you still choose Sanger over NGS?
For targeted verification: confirming a single variant, checking a plasmid construct, or clinically validating an NGS finding. Sanger's long accurate reads and simple interpretation beat NGS's industrial overkill when the question is small and precision matters.