Reading the Genome: Platforms & Modalities
Short-read, long-read, optical mapping, single-cell, methylation — what each modality actually sees.
What's covered
- T01Illumina short-read SBS: chemistry, error model, throughput economics
- T02PacBio HiFi (CCS): consensus accuracy, methylation calling
- T03Oxford Nanopore: ultra-long reads, native modifications, real-time
- T04Optical mapping (Bionano), Hi-C, Strand-seq for structural variation
- T05Single-cell genomics: scRNA-seq, scATAC-seq, multiomics
- T06Targeted approaches: exome, panels, adaptive sampling
By the end of this module you will be able to
- L01Match a clinical question (germline SNV, repeat expansion, methylation, structural variant) to the appropriate sequencing modality.
- L02Quantify the cost-vs-resolution trade-offs between WGS, WES, and gene panels in 2026.
- L03Describe why long reads are required for accurate structural-variant detection and pseudogene resolution.
What you should walk away believing
- Short-read WGS misses ~30–50% of structural variants and most repeat expansions; long-read WGS now finds them at clinical sensitivity.
- PacBio HiFi and ONT both call 5mC and 6mA natively without bisulfite conversion — methylome and genome from a single library.
- Adaptive sampling on ONT lets you target arbitrary regions in silico with no library redesign.
- Single-cell multiomics (10x Multiome, Parse, BD Rhapsody) is now standard for tissue heterogeneity studies but rarely clinical.
What this means at your level
Different sequencing technologies see different parts of biology. Short-read sequencing (Illumina) is cheap and accurate for single-base changes. Long-read sequencing (PacBio, Oxford Nanopore) is required to resolve repeat expansions, large structural variants, and DNA methylation natively. The right test depends on the clinical question.
Default to a gene panel for known-disease workups; default to exome (WES) when the phenotype is broad; reserve genome (WGS) for unsolved cases or when structural variation, intronic variants, or mitochondrial DNA are likely. For Fragile X, Huntington, FSHD, repeat-expansion ALS — short-read is inadequate; order long-read or a targeted repeat-primed PCR.
HiFi error rates (~Q30) match short-read while preserving 15–25 kb read length, making it the de facto standard for de novo assembly. ONT's PromethION + Dorado basecalling now achieves modal Q30 simplex and Q35+ duplex with native 5mC/6mA. Adaptive sampling enables programmable enrichment without wet-lab capture.
Whole-genome sequencing finds all variants.
Standard short-read 30x WGS systematically misses repeat expansions, large insertions, complex SVs, and many regions of segmental duplication. Diagnostic yield improvements from short-read WGS over WES are modest (~5–10%); long-read WGS adds a meaningful additional yield in unsolved Mendelian cases.
What the data say
Test yourself
Key terms & abbreviations
- WGS / WES
- Whole-genome / whole-exome sequencing. Exome captures ~1–2% of the genome (coding regions plus splice sites).
- HiFi
- PacBio circular-consensus reads, typically 15–25 kb at >Q30 — currently the gold standard for de novo assembly.
- Adaptive sampling
- Oxford Nanopore feature that ejects molecules in real time based on early-read mapping, enabling target enrichment without library prep changes.
- Hi-C
- Proximity-ligation method to map 3D genome organization; resolves structural variants, scaffolds assemblies, and identifies TAD boundaries.
Anchor references
- Long-read human genome sequencing and its applications — Logsdon, Vollger & Eichler, Nat Rev Genet 2020
- Nanopore sequencing technology, bioinformatics and applications — Wang et al., Nat Biotechnol 2021