Sequencing: Sanger, NGS, and Long Reads
TL;DR
Sanger sequencing reads one targeted region at high per-read accuracy and remains useful for small validation tasks. NGS sequences millions of fragments in parallel and powers panels, exomes, genomes, RNA-seq, and single-cell assays. Long-read sequencing reads longer DNA/RNA molecules, helping with structural variants, repeats, phasing, isoforms, and methylation-aware workflows. Sources: [1], [2], [3]
Method comparison
| Method | Best for | Weakness |
|---|---|---|
| Sanger | single locus, plasmid, small validation | low throughput, poor for mixtures |
| Targeted NGS panel | known cancer genes, clinical depth | limited discovery outside panel |
| Exome | coding mutations | misses noncoding and many structural events |
| Whole genome | broadest DNA view | cost, storage, interpretation |
| RNA-seq | expression, fusions, splicing | RNA quality and expression bias |
| Long-read | SVs, repeats, phasing, isoforms | cost, input quality, platform-specific error profile |
What sequencing produces
| Wet-lab step | Data object |
|---|---|
| extraction | concentration, purity, integrity |
| library prep | insert size, adapter/index metadata |
| sequencing | FASTQ reads and quality scores |
| alignment | BAM/CRAM |
| variant calling | VCF/BCF |
| expression analysis | count matrix, TPM, differential expression |
Common failure modes
- degraded DNA/RNA
- low tumor purity
- formalin artifacts
- index hopping or sample swaps
- insufficient depth for low allele fraction
- capture bias
- PCR duplicates
- poor reference alignment in repetitive regions
- overinterpreting variants without orthogonal evidence
Developer notes
- FASTQ is not "the sequence"; it is reads plus quality scores.
- BAM/CRAM alignment depends on reference genome, aligner, parameters, and preprocessing.
- VCF records are hypotheses after filtering, not raw truth.
- Tumor-only sequencing needs special care because germline and somatic variants can be confused.
- Long-read and short-read calls should not be merged blindly; they have different error profiles.
- Every variant table should preserve sample, assay, genome build, caller, filtering, and depth metadata.
See also
References
- Sanger F, Nicklen S, Coulson AR. DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci U S A 1977;74:5463-5467. https://doi.org/10.1073/pnas.74.12.5463
- Behjati S, Tarpey PS. What is next generation sequencing? Arch Dis Child Educ Pract Ed 2013;98:236-238. PMID 23986538. https://doi.org/10.1136/archdischild-2013-304340
- Si HQ, Wang P, Long F, et al. Cancer liquid biopsies by Oxford Nanopore Technologies sequencing of cell-free DNA. Mol Cancer 2024;23:265. PMID 39614371. https://doi.org/10.1186/s12943-024-02178-6