Illumina is most often chosen for projects requiring a large number of highly accurate short reads, such as genome resequencing, RNA-seq or SNP identification. PacBio works well when accurate long HiFi reads are needed, while Oxford Nanopore Technologies - ONT is particularly useful for generating very long and ultra-long reads, analysing DNA modifications and performing direct RNA sequencing.
You do not always need to choose just one platform. For de novo genome assembly, the best results often come from combining PacBio or ONT long reads with Illumina short reads and data from Hi-C.
Technology comparison
| Criterion | Illumina | PacBio HiFi | Oxford Nanopore - ONT |
|---|---|---|---|
| Read type | Short | Long, highly accurate | Long and ultra-long |
| Typical read length | Most often 2 × 150 bp | Usually a dozen or so to several tens of kb | From a few kb to over 100 kb for ultra-long libraries |
| Single-read accuracy | Very high | Very high - HiFi | High, depending on sequencing chemistry, DNA quality and basecalling |
| Throughput | Very high | High | Scalable depending on the instrument and flow cell |
| SNPs and small InDel variants | Very well suited | Very well suited | Possible, usually with appropriate coverage and a suitable pipeline |
| Structural variants | Limited capabilities | Very well suited | Very well suited |
| Repetitive regions | More difficult to resolve | Very well suited | Particularly well suited thanks to very long reads |
| Haplotyping | Limited by short read length | Yes | Yes |
| De novo genome assembly | Usually insufficient as the only data source | Yes | Yes |
| DNA methylation analysis | Requires an additional protocol, e.g. WGBS | Possible with native DNA | Possible with native DNA |
| Full-length transcripts | Not in a single read | Yes - Iso-Seq | Yes - cDNA or direct RNA |
| Direct RNA sequencing | No | Not in the standard workflow | Yes |
| Real-time analysis | No | No | Yes |
| Main advantage | Accuracy, throughput and unit cost | Combination of length and accuracy | Longest reads and flexibility |
When should you choose Illumina?
Choose Illumina if:
- you are planning genome resequencing against a good-quality reference,
- your main objective is to identify SNPs and small InDel variants,
- you are analysing gene expression using mRNA-seq or Total RNA-seq,
- you are performing WES, WGBS, ChIP-seq or ATAC-seq,
- you are analysing a large number of samples,
- you need high throughput and a low cost per base,
- you are performing Amplicon Sequencing or shotgun metagenomics,
- you do not need reads that span long repetitive regions.
Illumina uses sequencing by synthesis technology. It most often generates 2 × 150 bp paired-end reads. Short reads are of high quality, but may be insufficient to assemble repetitive regions unambiguously, identify complex structural variants or phase haplotypes.
When should you choose PacBio HiFi?
Choose PacBio if:
- your objective is de novo genome assembly,
- you need accurate long reads,
- you are analysing complex structural variants,
- you want to phase haplotypes,
- the genome under study is polyploid, heterozygous or rich in repetitive sequences,
- you are planning to analyse full-length transcripts using Iso-Seq,
- you want to analyse DNA sequences and modifications simultaneously,
- you need high consensus accuracy without correction using short reads.
PacBio HiFi technology is based on reading the same circular molecule multiple times. These reads are used to generate a consensus sequence that combines read length with very high accuracy.
PacBio is a particularly good choice for generating high-quality reference genomes and analysing structural variants. HiFi data can be combined with Hi-C to obtain a chromosome-scale genome assembly.
When should you choose Oxford Nanopore Technologies?
Choose ONT if:
- you need very long or ultra-long reads,
- your objective is to assemble a genome de novo and link distant regions,
- you are analysing complex rearrangements and structural variants,
- you want to perform sequencing and data analysis in real time,
- you are analysing methylation without prior bisulfite conversion,
- you want to sequence full-length RNA molecules,
- you are interested in direct RNA sequencing.
ONT read length depends primarily on the length and quality of the input molecules. The technology can generate reads exceeding 100 kb, and much longer reads under suitable conditions.
ONT allows direct analysis of native DNA and RNA. It is therefore of particular interest in projects investigating base modifications, full-length transcripts and the heterogeneity of RNA molecules.
Sample Requirements
The exact requirements depend on the application, organism, selected protocol and expected read length.
Illumina
The most common requirements are:
- purified DNA or RNA in an appropriate amount and at an appropriate concentration,
- no inhibitors of enzymatic reactions,
- good purity confirmed by OD260/280 and OD260/230 measurements,
- appropriate RNA integrity for RNA-seq projects,
- DNA may be fragmented during library preparation.
Illumina does not require very long DNA molecules to be kept intact, so material integrity can be lower than in long-read projects. However, the degree of degradation can still affect library quality and the scope of possible analyses.
PacBio HiFi
The most common requirements are:
- high-molecular-weight DNA - HMW DNA,
- minimal mechanical fragmentation of the material,
- high purity and no inhibitors,
- an amount of DNA appropriate for the selected protocol and genome size,
- reliable measurement of concentration by fluorometry,
- fragment length assessment.
Iso-Seq analysis requires good-quality RNA, as degradation may prevent the recovery of complete transcripts.
Oxford Nanopore Technologies
The most common requirements are:
- high-molecular-weight or ultra-high-molecular-weight DNA,
- gentle isolation that limits DNA shearing,
- high sample purity,
- no residual salts, alcohol, detergents or other inhibitors,
- an appropriate quantity of material for the selected kit and flow cell.
If the objective is to obtain ultra-long reads, the way DNA is isolated and handled is one of the most important aspects of the entire project. Vigorous pipetting, vortexing, repeated freezing or improper transport can significantly shorten the molecules.
Direct RNA sequencing requires native RNA that meets the requirements of the specific ONT protocol.
Approximate data output
Amount of data should be determined based on genome size, ploidy, heterozygosity, the type of analysis and the expected coverage.
Illumina
Indicative values:
- genome resequencing: usually approximately 20-30×,
- human WGS analysis: most often approximately 30×,
- assembly of a small microbial genome: often at least 50-100×,
- mRNA-seq: usually approximately 20-30 million read pairs per sample,
- projects that require analysis of lowly expressed transcripts or novel isoforms may require more data.
PacBio HiFi
Indicative values:
- de novo genome assembly: usually approximately 20-30×,
- heterozygous, polyploid or particularly complex genomes may require higher coverage,
- structural variant analysis and haplotyping: the range should be tailored to the organism and project objective,
- Iso-Seq: the amount of data depends on transcriptome complexity and the number of tissues or conditions studied.
Oxford Nanopore Technologies
Indicative values:
- de novo genome assembly: often approximately 30-40×,
- structural variant analysis: usually a minimum coverage of around a dozen-fold to several tens-fold, depending on the objective,
- ultra-long data to supplement assembly: the amount is determined individually,
- direct RNA or full-length transcriptome: the amount of data depends on the number of samples, transcriptome complexity and the expected sensitivity.
In de novo projects, the amount of data should be calculated in gigabases based on the estimated genome size:
amount of data = genome size × planned coverage
Is it worth combining technologies?
Yes. A hybrid approach can provide more complete information than using a single platform.
The most commonly used combinations are:
The decision to use such an approach should be driven by a specific objective, rather than the assumption that more technologies always improve the result.
Common mistakes when choosing a technology
Choosing a platform before defining the project objective
The choice of technology should be guided by the research question. The optimal platform for analysing SNPs in a large population will differ from that for assembling a new reference genome.
Trying to assemble a complex genome using only short reads
Illumina may not resolve long repeats and complex genomic regions. Long reads are usually needed for de novo assembly of large eukaryotic genomes.
Providing fragmented DNA for long-read sequencing
With PacBio and ONT, the resulting read length depends on the integrity of the input material. Damaged DNA cannot be restored during library preparation.
Relying solely on the number of gigabases
The analytical value of the same amount of data can vary depending on read length, accuracy and distribution. You should take genome coverage and the analysis objective into account.
Assuming all long-read sequencing produces the same result
PacBio HiFi and ONT offer different benefits. PacBio combines long reads with very high accuracy, while ONT can provide much longer reads and real-time data.
Failing to account for ploidy and heterozygosity
Genome size is not the only parameter that determines the amount of data required. Polyploid, hybrid and highly heterozygous genomes may require higher coverage and an appropriate bioinformatics strategy.
Assuming long-read sequencing replaces Hi-C
Long reads enable contig assembly, but Hi-C data are often needed to order and orient sequences at the chromosome level.
Project planning questions
Before choosing a technology, consider the following questions:
Plan your project
Describe your project - we will recommend a method and amount of data.
Based on the organism under study, the quality of the material and the analysis objective, we will help you choose Illumina, PacBio, Oxford Nanopore or a hybrid approach. We will also determine the required coverage, sample preparation method and appropriate scope of bioinformatics analysis.