For assembling large eukaryotic genomes, the best starting point is currently PacBio HiFi, which combines long reads with high accuracy. Oxford Nanopore is a good choice when very long reads and resolving repetitive regions are priorities, while Illumina is primarily suited to Genome Survey (initial genome characterisation), quality control and projects involving small, simple genomes.
If the aim is a high-quality chromosome-level reference genome assembly, a hybrid approach is most often recommended: PacBio HiFi + Hi-C, optionally supplemented with ONT ultra-long and Illumina.
Technology comparison
| Technology | Main advantage | Main limitation | Best application |
|---|---|---|---|
| Illumina | Highly accurate, low-cost short reads | Difficulties with repeats and haplotypes | Genome Survey, small genomes, additional validation |
| PacBio HiFi | Long and highly accurate reads | Requires high-quality HMW DNA | Reference-quality de novo assembly |
| Oxford Nanopore | Very long and ultra-long reads | The result depends heavily on HMW DNA quality | Repeats, centromeres, telomeres, T2T |
| Hybrid approach | Highest completeness and contiguity | Higher cost and more material | Chromosome-level genome assembly |
When should you choose Illumina?
Illumina is suitable when:
- you want to carry out an initial genome assessment - Genome Survey,
- you need to estimate genome size, heterozygosity and repeat content,
- you are analysing a small and simple bacterial genome,
- the data are intended for additional assembly validation or correction,
- you have DNA that does not meet the requirements for long-read sequencing.
Illumina may be sufficient for a draft assembly of a small genome. For large plant and animal genomes, short reads do not usually allow repetitive regions, duplications and haplotypes to be represented accurately.
When should you choose PacBio HiFi?
PacBio HiFi is recommended when:
- the aim is a high-quality de novo assembly,
- the genome is large, heterozygous or polyploid,
- haplotype resolution is needed,
- high base-level accuracy is important,
- structural variants will be analysed,
- the genome contains numerous repeats and duplications.
PacBio HiFi is often the best foundation for building a new reference genome (de novo). To scaffold contigs into chromosomes, HiFi data should be supplemented with Hi-C.
When should you choose Oxford Nanopore?
ONT is a good choice when:
- maximising read length is crucial,
- the aim is to resolve very long repeats,
- you are analysing centromeres, telomeres or segmental duplications,
- the project aims for a telomere-to-telomere assembly,
- identifying large-scale rearrangements is important,
- very high-quality ultra-HMW DNA is available.
ONT can be used on its own or to complement PacBio HiFi. Even moderate sequencing coverage with ultra-long reads may help bridge regions that shorter reads have not resolved.
When should you choose a hybrid approach?
A hybrid approach is best when the aim is to obtain a complete, gap-free (0-gap) genome assembly.
The most common combinations are:
With current PacBio HiFi data, additional polishing with Illumina is often not required. However, short reads remain useful for Genome Survey and independent assembly validation.
Sample Requirements
Illumina
- purified genomic DNA,
- usually at least approximately 1 µg,
- a concentration of approximately 30 ng/µl or above,
- preserving ultra-long fragments is not required.
PacBio HiFi
- high-molecular-weight DNA - HMW DNA,
- usually several to a dozen or so micrograms, depending on the protocol,
- high purity and integrity,
- minimal fragmentation caused by isolation and/or transport,
- fresh or properly frozen material is preferred.
Oxford Nanopore
- HMW or ultra-HMW DNA,
- usually several micrograms or more,
- very high purity,
- do not pipette samples vigorously or vortex them,
- an appropriate isolation method selected for the expected read length.
The exact requirements depend on genome size, the technology and the number of planned libraries. It is best to discuss sample collection and isolation methods before starting the experiment.
Approximate data output
| Data type | Indicative sequencing coverage |
|---|---|
| Illumina Genome Survey | approximately 30-50×, with even ≥50× recommended |
| PacBio HiFi de novo | approximately 25-30× |
| PacBio HiFi per haplotype | approximately 10-15× per haplotype |
| ONT long-read | approximately 30-40× |
| ONT ultra-long as a supplement | approximately 5-15× |
| Hi-C | selected according to genome size and complexity |
Amount of data is calculated using the formula:
genome size × target coverage = required data output
Example: a 1 Gb genome at 30× sequencing coverage requires approximately 30 Gb of data.
Higher sequencing coverage may be needed for polyploid, highly heterozygous or repeat-rich genomes.
Common mistakes
- choosing Illumina-only sequencing for a large genome rich in repetitive sequences,
- starting a project without estimating genome size and heterozygosity,
- providing degraded DNA for PacBio or ONT,
- pipetting HMW DNA vigorously and freezing it repeatedly,
- failing to account for ploidy and hybrid origin,
- expecting a chromosome-level assembly without Hi-C,
- sequencing DNA from multiple genetically distinct individuals.
The quality of the genome assembly should also be assessed in terms of completeness, accuracy, the number of gaps, the presence of contamination and correct haplotype resolution.