From Sanger to Next-Generation Sequencing

DNA is a long chain built from four chemical letters called bases: adenine (A), cytosine (C), guanine (G), and thymine (T). DNA sequencing means reading the exact order of those bases along a strand. This lesson traces how sequencing grew from reading a few hundred bases at a time to reading billions of them at once.

Sanger Sequencing

Sanger sequencing, invented by Frederick Sanger in 1977, is known as the chain-termination method. It copies a DNA strand and every so often adds a special base, a dideoxynucleotide, that has no attachment point for the next base and therefore stops the copying. By measuring the lengths at which copying stopped, the machine reconstructs the original sequence one fragment at a time.

The first human genome, completed by the Human Genome Project in 2003, was read almost entirely with Sanger machines and took about 13 years and close to three billion dollars. Each continuous stretch of sequence a machine reports is called a read, and Sanger reads are long, up to roughly 1,000 bases, and extremely accurate. Its weakness is throughput, meaning how much it can process at once: the machine handles only about 96 fragments at a time, so reading a whole genome this way is painfully slow and expensive.

Massively Parallel Sequencing

Next-generation sequencing, or NGS, breaks the genome into millions of small fragments and reads all of them at the same time, an approach called massively parallel sequencing. Because so many fragments are read in a single run, the cost of sequencing a human genome fell from billions of dollars to just a few hundred. The trade-off is that each individual read is much shorter than a Sanger read.

# echo prints text to the screen.
# The $(( )) syntax tells the bash shell to do the math inside it.
# One Illumina run can produce about 20 billion reads.
# One Sanger machine produces only about 96 reads per run.
echo "$(( 20000000000 / 96 )) Sanger runs equal one Illumina run"

208333333 Sanger runs equal one Illumina run

Short and Long Reads

The dominant NGS platform, made by Illumina, produces short reads of roughly 50 to 300 bases each. These reads are cheap and highly accurate, which makes them ideal for measuring how active genes are or spotting single-letter differences between genomes. Their shortness, however, makes it hard to piece together long repetitive stretches of DNA, where the same pattern appears again and again.

Long-read platforms from PacBio and Oxford Nanopore take the opposite approach, reading whole single molecules instead of tiny fragments. A PacBio read is typically tens of thousands of bases, while Oxford Nanopore can occasionally read a single molecule millions of bases long. PacBio watches a polymerase add fluorescent bases inside a tiny well, while Oxford Nanopore threads a DNA strand through a protein pore and measures small changes in electrical current. Long reads span repetitive regions and reveal large structural changes, so short and long reads are often combined to get both accuracy and completeness.