Which sequencing platform should you use? It sounds like a simple question, and it never is. Sanger, Illumina and Oxford Nanopore all read DNA, yet they go about it in completely different ways, and those differences decide which one is right for your question.
In this post I walk through the three, one at a time. Every technique gets a one-line explanation the first time it appears, so nobody is left behind, and at the end I suggest a practical way to choose between them. If you work with Plasmodium falciparum, I also point out where each platform helps and where it trips.
Sanger sequencing: the careful pioneer
Let's start where it all began. In 1977, Frederick Sanger, Steve Nicklen and Alan Coulson published a method for reading DNA that was faster and more accurate than anything before it, and they proved it on the genome of a small virus called bacteriophage φX174 (Sanger et al., 1977). Sanger shared the 1980 Nobel Prize in Chemistry for this body of work, and his method stayed the backbone of sequencing for about thirty years, right through the Human Genome Project.
How it works
Everything begins with a primer, a short piece of DNA that sticks to one defined spot on your template (the DNA you want to read). A polymerase, the enzyme that copies DNA, starts extending from that spot. This detail matters more than it looks. Because the primer binds at one known place, every new strand begins at exactly the same position, and only the far end differs from strand to strand. That is how we know where we are, and which direction we are reading: away from the primer, from the 5′ end toward the 3′ end of the new strand.
Now comes the clever part, called chain termination. Mixed in with the normal building blocks (dNTPs) is a small amount of special ones called dideoxynucleotides, or ddNTPs. A ddNTP is missing the 3′ hydroxyl group that the polymerase needs to attach the next base, so when one gets incorporated, copying stops dead. Because ddNTPs are scarce compared with normal dNTPs, the stop happens at a random position in each copy. Over millions of copies you end up with a fragment ending at every possible position: a nested set, all sharing the same start (the primer), each ending at a known base.
One tube or four? A little history
Here is something that confused me. In the original 1977 method, Sanger ran four separate reactions, one for each type of terminator, with radioactive labels, and loaded the products into four neighbouring lanes of a gel. (A fun detail: the cytosine terminator in that first paper was an arabinose analogue, araCTP, rather than ddCTP.) Many textbooks still draw it this way, and having that four-tube picture in my memory made the modern version confusing at first.
Today the reaction happens in a single tube. All four ddNTPs are present together, and each carries its own fluorescent dye, one colour per base. Since the last base of every fragment is colour-coded, the whole mixture can be loaded into one capillary and read in a single run. Same idea, a fraction of the pipetting.
Reading the result
The fragments are sorted by size using capillary electrophoresis. That sounds intimidating, but the idea is simple. DNA carries a negative charge, so when you apply an electric field across a hair-thin tube filled with gel, the DNA moves, and shorter pieces slip through faster than longer ones. As each fragment passes a detector, a laser excites its dye and the colour is recorded. The shortest fragment comes first, so the first colour is the base right after the primer, the second colour is the next base, and so on. The output is the chromatogram, a trace of coloured peaks with one peak per base.
Strengths and limits
Sanger reads run to roughly 700 to 1,000 bases, and accuracy is the best of the three platforms, around 99.99% in the clean middle of a read. The first 30 bases or so, and the tail end, are messier. Throughput is the catch: one fragment per reaction, or 96 or 384 in parallel at most. Think scalpel, not combine harvester.
That is why I still love it for jobs with a single target. Checking a plasmid (a small circular piece of DNA used as a cloning vector), confirming a CRISPR edit, genotyping one locus, verifying an amplicon, or validating a variant that another method flagged. For all of these it remains the gold standard.
Its blind spot is mixed templates. If two different sequences are in the tube, their peaks overlap and the chromatogram turns into soup. In malaria work, where one patient can carry several parasite genotypes at once, that is a real headache.
Illumina sequencing: the workhorse
If Sanger is the scalpel, Illumina is the factory. The technology began at a company called Solexa, and a landmark 2008 paper from that team showed the approach could produce an accurate human genome (Bentley et al., 2008). They sequenced a male Yoruba individual from Ibadan, Nigeria, reading short paired 35-base fragments to an average depth of more than 30x, and from that they characterised about four million single-nucleotide variants. It is rather lovely that a West African genome was the showpiece.
And here is a small twist of history. Two of the authors, John Milton and Clive Brown, list Oxford Nanopore Technologies as their present address in that very paper. Hold that thought, because we will meet their company again shortly.
How it works
Illumina uses sequencing by synthesis, which means you watch a polymerase build a new strand and record each base as it is added. First, the DNA is broken into fragments and given adapters, short pieces of known sequence on each end. The adapters let the fragments stick to a flow cell, a glass slide with tiny lanes whose surface is coated with matching anchor sequences.
One molecule on its own is far too faint to detect, so each fragment is copied in place into a tight cluster of identical molecules, using bridge amplification (or ExAmp on patterned flow cells). Thousands of identical strands shining together give the camera something it can see.
Then the cycles begin. Each cycle adds fluorescent, reversibly terminated nucleotides: bases that carry a dye and a removable block, so only one base can be added per round. The instrument images every cluster, cleaves off the dye and the block, and repeats. Newer instruments such as the NextSeq and NovaSeq use two-colour chemistry to speed this up. Reads are typically 50 to 300 bases long and are usually paired-end, meaning both ends of each fragment are sequenced, which makes it much easier to work out where a read belongs in the genome.
What it is good for
Accuracy is very high: better than Q30, which means fewer than 1 error in 1,000 bases (about 99.9%), and the errors are mostly substitutions, where one base is swapped for another. Output ranges from gigabases on a small MiSeq to multiple terabases per run on a NovaSeq X, at the lowest cost per base of the three platforms.
That scale powers whole genome sequencing (WGS) and exome sequencing, which reads only the protein-coding portion, roughly 1 to 2% of the genome. It also powers RNA-seq (reading which genes are active by sequencing RNA after converting it to DNA), ChIP-seq (finding where proteins bind DNA), amplicon panels (deep sequencing of PCR-amplified target regions), variant calling (finding where a sample differs from a reference) and population genomics.
Where it struggles
Short reads have blind spots. Repeats and structural variants (large rearrangements such as deletions, inversions and duplications) are hard to resolve, because a short piece of sequence cannot tell you where in a repetitive region it belongs. The same goes for highly variable gene families.
The chemistry brings its own gremlins. Amplification can cause GC bias (uneven coverage of GC-rich and AT-rich regions) and duplicates (PCR copies of the same original molecule). This is rough on AT-rich genomes like P. falciparum, where PCR-free or low-cycle library protocols help. Index hopping happens when the barcodes that label each sample swap onto the wrong one, causing crosstalk. Phasing errors occur when some molecules in a cluster fall out of step, so the signal blurs in later cycles.
Oxford Nanopore sequencing: the free spirit
Remember John Milton and Clive Brown? Oxford Nanopore Technologies took a completely different route. Its best-known moment came in 2015, when a team packed a genome surveillance laboratory into less than 50 kg of ordinary airline luggage and took it to Guinea during the West African Ebola epidemic. The sequencer at the heart of it, the MinION, weighs less than 100 grams. They began sequencing within two days of arriving, and they could produce results less than 24 hours after receiving a positive sample, with the sequencing itself taking as little as 15 to 60 minutes (Quick et al., 2016). Carrying the sequencer to the outbreak, instead of shipping samples to the sequencer, still strikes me as a wonderful idea.
How it works
Nanopore sequencing is different in kind. There is no synthesis step, and no amplification is needed: it reads a single native molecule of DNA or RNA. A motor protein, an enzyme that ratchets the DNA along, feeds a single strand through a protein nanopore embedded in a membrane, while a voltage drives ionic current through the pore. Each k-mer, a short stretch of a few bases sitting in the pore at that moment, disrupts the current in its own characteristic way. Neural-network basecallers, software that translates the current trace into A, C, G and T, turn that signal into sequence.
Reads, accuracy and scale
Read length is limited by your DNA, not by the chemistry. Reads run from a few hundred bases to well over a megabase if you can extract DNA that long. The Flongle and MinION are portable, the GridION is mid-scale, and the PromethION is built for high throughput. Per gigabase, it is still generally more expensive than Illumina at large scale.
Accuracy used to be the weak spot, and homopolymers (runs of the same base, such as AAAAAA) were especially hard. The Ebola team noted exactly this limitation in 2016 (Quick et al., 2016). Things have moved on. With the R10.4.1 pore and Kit 14 chemistry, Oxford Nanopore reports simplex reads (one strand read once) at Q20 or better, and duplex reads (both strands read and combined into one consensus) at about Q30 or higher (Oxford Nanopore Technologies, n.d.). Homopolymers and some systematic errors remain the hardest cases, and the platform is sensitive to DNA quality and contaminants, so extraction matters.
The superpowers
- Ultra-long reads that span repeats and structural variants, which is why genomes assemble so well.
- Direct RNA sequencing, which reads the RNA itself with no cDNA conversion and gives full-length isoforms (the different versions of a transcript made from a single gene).
- Native detection of base modifications such as 5mC and 6mA, methyl tags on DNA that can influence gene regulation. The older approach, bisulfite conversion (a harsh chemical treatment), is not needed, which is welcome news because bisulfite damages DNA.
- Real-time output, so you can stop the run when you have enough data.
- Adaptive sampling, where the pore decides as it reads whether to keep or reject a molecule, enriching the targets you care about.
- Portability, as the Ebola story shows.
Side by side
| Sanger | Illumina | Oxford Nanopore | |
|---|---|---|---|
| Read length | About 700 to 1,000 bases | 50 to 300 bases | A few hundred bases to over 1 Mb |
| Per-base accuracy | Highest (about 99.99%) | Very high (above Q30) | Good and improving (Q20+ simplex, about Q30 duplex) |
| Throughput | Very low | Very high | Low to high, scalable |
| Cost per base | Highest | Lowest | Generally higher than Illumina at scale |
| Best for | Validating one target | Variant calling, expression, population data | Assembly, structural variants, modifications, field work |
| Main weakness | Throughput, mixed templates | Short reads, GC bias | Raw error rate, sensitivity to DNA quality |
Which one should you choose?
Ask what your question really needs. If you want to confirm one variant or one clone, reach for Sanger. If you need depth and breadth, such as a whole genome, an exome, or expression across thousands of genes, reach for Illumina. If you need to span repeats, find structural variants, read base modifications, or sequence in the field, reach for Nanopore.
Better still, think of them as a team rather than rivals. A common strategy is hybrid. Long reads (from Nanopore, or from PacBio HiFi, another long-read platform) scaffold the genome, meaning they order and join assembled pieces, and they resolve repeats. Illumina polishes the accuracy. Sanger confirms the key variants.
What this means for Plasmodium falciparum work
The P. falciparum genome is extremely AT-rich, at roughly 80% A and T overall, which makes it a stress test for any method that relies on amplification. That is why PCR-free or low-cycle Illumina libraries, and increasingly native nanopore reads, attract so much interest in malaria genomics. Mixed-genotype infections, common in high-transmission settings, will muddy a Sanger chromatogram but can be picked apart with deep short-read coverage. Highly variable, repetitive gene families such as the var genes are exactly where long reads earn their keep. And Sanger still has a place: confirming a single resistance-associated variant once another platform has pointed to it.
Watch the video
Prefer to watch rather than read? The video version will be on the Adwoa Biotech YouTube channel on Dec 17, 2026.
Related Content
- Sanger sequencing: https://adwoabiotech.blogspot.com/2026/06/sanger-sequencing-explained-how-one.html
- Next-generation sequencing (Illumina): https://adwoabiotech.blogspot.com/2026/06/what-is-next-generation-sequencing-ngs.html
- Oxford Nanopore sequencing: https://adwoabiotech.blogspot.com/2026/06/what-is-oxford-nanopore-sequencing-how.html
- PacBio vs Nanopore: https://adwoabiotech.blogspot.com/2026/06/pacbio-smrt-sequencing-vs-oxford.html
- Ion Torrent sequencing: https://adwoabiotech.blogspot.com/2026/07/ion-torrent-sequencing-dna-reader-that.html
References
Bentley, D. R., Balasubramanian, S., Swerdlow, H. P., Smith, G. P., Milton, J., Brown, C. G., Hall, K. P., Evers, D. J., Barnes, C. L., Bignell, H. R., Boutell, J. M., Bryant, J., Carter, R. J., Cheetham, R. K., Cox, A. J., Ellis, D. J., Flatbush, M. R., Gormley, N. A., Humphray, S. J., . . . Smith, A. J. (2008). Accurate whole human genome sequencing using reversible terminator chemistry. Nature, 456(7218), 53–59. https://doi.org/10.1038/nature07517
Oxford Nanopore Technologies. (n.d.). Q20+ chemistry. https://nanoporetech.com/q20plus-chemistry
Quick, J., Loman, N. J., Duraffour, S., Simpson, J. T., Severi, E., Cowley, L., Bore, J. A., Koundouno, R., Dudas, G., Mikhail, A., Ouédraogo, N., Afrough, B., Bah, A., Baum, J. H. J., Becker-Ziaja, B., Boettcher, J. P., Cabeza-Cabrerizo, M., Camino-Sánchez, Á., Carter, L. L., . . . Carroll, M. W. (2016). Real-time, portable genome sequencing for Ebola surveillance. Nature, 530(7589), 228–232. https://doi.org/10.1038/nature16996
Sanger, F., Nicklen, S., & Coulson, A. R. (1977). DNA sequencing with chain-terminating inhibitors. Proceedings of the National Academy of Sciences, 74(12), 5463–5467. https://doi.org/10.1073/pnas.74.12.5463
Subscribe by Email
Follow Updates Articles from This Blog via Email





No Comments