Whole-Genome Resequencing with Long-Read Sequencing — PacBio Revio & ONT PromethION
Whole-genome resequencing (WGRS) aligns an individual's genome against an existing reference to catalog single nucleotide variants (SNVs), small insertions and deletions (indels), and structural variants (SVs). Short-read WGRS handles SNVs and indels well, but struggles with SVs in repetitive or GC-extreme regions, where reads are too short to span the variant. CD Genomics' long-read whole-genome resequencing service, run on PacBio Revio and ONT PromethION, spans these regions directly, resolving variants that short reads routinely miss.
We provide WGRS for human, animal, plant, and microbial samples, and support downstream analysis—variant calling, GWAS, and population evolution studies—through dedicated services described below.
Why researchers choose our long-read WGRS service
- Long reads (median 15–20 kb) span repeats and structural variants that short reads cannot
- No PCR amplification required, avoiding amplification-related bias
- Direct access to GC-extreme and repeat-expansion regions
- Choice of PacBio Revio or ONT PromethION, matched to your project's needs
- Supports human, animal, plant, and microbial genomes on a single platform pair
At a glance:
- Introduction
- What Is Long-Read Whole-Genome Resequencing?
- Sequencing Platforms We Use
- Key Advantages
- Our WGRS Services
- Workflow – How It Works
- Variant Calling Toolchain
- Choosing Coverage and Platform
- Sample Requirements
- Why Choose CD Genomics
- Case Study: Benchmarking Long-Read Structural Variant Calling Methods
- FAQs
Introduction
Genomic variation ranges from single-base substitutions to structural rearrangements spanning thousands of base pairs. Short-read whole-genome resequencing captures SNVs and small indels efficiently, but structural variants—deletions, insertions, inversions, duplications, and translocations—are frequently missed or mislocalized when reads are too short to fully span the variant or the flanking repetitive sequence. Long-read sequencing on PacBio Revio and ONT PromethION generates reads long enough to span these regions directly, improving both detection sensitivity and breakpoint precision.
What Is Long-Read Whole-Genome Resequencing?
Long-read WGRS sequences an individual's whole genome and aligns the resulting reads to an existing reference genome, rather than assembling a new one from scratch. Because PacBio HiFi and ONT reads average well over 10 kb, they span repetitive elements, segmental duplications, and structurally complex loci that defeat short-read alignment, allowing variant callers to resolve SNVs, indels, and SVs from the same dataset.
This differs from de novo assembly, which builds a new reference rather than comparing to an existing one—see our Animal/Plant Whole Genome De Novo Sequencing service if no adequate reference exists for your organism.
Sequencing Platforms We Use
PacBio Revio (SPRQ-Nx Chemistry)
Our primary platform for long-read WGRS. Revio HiFi reads carry a median consensus accuracy of Q30 or better at read lengths of 15–20 kb, giving high-confidence SNV, indel, and SV calls in a single pass, with on-instrument 5mC methylation calling included.
ONT PromethION (R10.4.1, Kit 14)
An alternative or complementary platform offering ultra-long read options and high per-run throughput, well suited to large cohort studies, population-scale resequencing, or projects that also require native RNA or ultra-long DNA context.
Both platforms are supported through PacBio SMRT sequencing and Oxford Nanopore sequencing. Our team can help you choose based on read-length needs, cohort size, and budget.
Key Advantages
Scientific Advantages
- Structural variant sensitivity
Long reads span the repetitive and GC-extreme regions where short-read WGRS loses sensitivity, improving detection of deletions, insertions, inversions, duplications, and translocations.
- Precise breakpoints
Reads that fully span an SV give tighter, more reproducible breakpoint calls than reads reconstructed from short fragments.
- No amplification bias
Both platforms support PCR-free library preparation, avoiding coverage skew introduced by amplification.
- One dataset, multiple variant classes
SNVs, indels, SVs, and—on Revio—CpG methylation are all recoverable from the same sequencing run.
Business & Project Advantages
- Organism-agnostic service
Human, animal, plant, and microbial resequencing projects are all supported on the same platform pair.
- Flexible downstream analysis
Combine WGRS with variant calling, GWAS, or population evolution analysis as a single project.
- Scalable coverage
Coverage can be tuned to project goals, from cost-efficient population screening to deep, comprehensive variant profiling.
- Clear, interpretable reports
Deliverables include annotated variant call files and a plain-language summary suitable for methods sections.
Our WGRS Services
Variant Calling
Comprehensive SNV, indel, SV, and CNV detection from long-read WGRS data, providing the foundation for molecular marker development and functional gene discovery.
Genome-Wide Association Study (GWAS)
Association analysis between genome-wide variants and phenotypic traits, supported by the more complete variant catalog long reads provide.
Population Evolution
Population-scale variant profiling to study genetic diversity, population structure, gene flow, and speciation.
Human Whole-Genome Resequencing
SNV, indel, CNV, and SV profiling for disease gene screening, pathogenesis, and genetic mechanism studies.
Animal/Plant Whole-Genome Resequencing
Reference-aligned resequencing across chromosomal, mitochondrial, and (for plants) chloroplast DNA, supporting evolution, GWAS, and breeding programs.
Microbial Whole-Genome Resequencing
Strain-level comparison against reference genomes for bacteria, fungi, and other microbes, detecting key and low-frequency mutations.
Workflow – How It Works
1. Sample Preparation and Library Construction
High molecular weight DNA is extracted and converted into a PCR-free SMRTbell (PacBio) or ligation/rapid (ONT) library.
2. Long-Read Sequencing
Libraries are sequenced on PacBio Revio or ONT PromethION to the coverage depth appropriate for your variant classes of interest.
3. Alignment and Variant Calling
Reads are aligned to the appropriate reference genome, then processed through SNV, indel, and SV calling modules, with joint phasing where relevant.
4. Annotation and Reporting
- Variants are annotated against relevant databases
- SV calls are filtered and merged using the most robust tool combination for your coverage and organism
- A summary report and annotated VCF files are delivered for downstream analysis
Workflow of long-read whole-genome resequencing, from sample preparation through PacBio Revio or ONT PromethION sequencing to annotated variant calls.
Variant Calling Toolchain
| Variant Class | Typical Tools | Notes |
| SNVs and small indels | pbmm2/minimap2 alignment, DeepVariant | High precision and recall from HiFi or nanopore reads |
| Structural variants (alignment-based) | pbsv, cuteSV, Sniffles2 | Strong genotyping accuracy at low-to-moderate coverage (5–10×); best for complex SVs (translocations, inversions, duplications) |
| Structural variants (assembly-based) | hifiasm/verkko assembly with PAV, SVIM-asm, Dipcall | More sensitive to large insertions and robust to coverage changes; more computationally demanding |
| Phasing | HiPhase | Jointly phases small variants, SVs, and tandem repeats from HiFi read-backed information |
| SV benchmarking/merging | Truvari | Used to compare and merge SV calls across tools and samples |
Choosing Coverage and Platform
Tool and coverage choice both affect the sensitivity, precision, and cost of long-read WGRS. The guidance below summarizes trade-offs confirmed by independent benchmarking of PacBio and ONT SV calling methods.
| Goal | Recommended Coverage | Recommended Approach | Trade-off |
| Standard SNV/indel/SV profiling | 15–20× | Alignment-based calling (pbsv, cuteSV, Sniffles2) | Fast and cost-efficient; genotyping accuracy at low coverage varies by tool |
| Comprehensive SV detection, large insertions | ≥ 20–30× | Assembly-based calling (hifiasm + PAV/SVIM-asm/Dipcall) | More robust to coverage changes and evaluation stringency, but substantially more compute time |
| Population-scale screening | 10–15× | ONT PromethION, alignment-based calling | Cost-efficient at scale; precision remains high even as coverage drops, though recall for some variant classes declines |
| Complex SVs (translocations, inversions, duplications) | ≥ 20× | pbsv or cuteSV for translocations; tool choice varies by SV type | No single tool performs best across all complex SV types |
How to interpret this comparison
- For most WGRS projects, alignment-based calling at 15–20× coverage on PacBio Revio gives a strong balance of sensitivity, precision, and turnaround time.
- When maximizing sensitivity for large insertions or building a new reference is also a goal, pairing WGRS with de novo assembly-based SV calling adds robustness at the cost of compute time.
- For large cohorts, ONT PromethION's throughput often makes population-scale coverage targets more cost-effective, with alignment-based calling as the default.
Sample Requirements
| Category | Requirement | Notes |
| Sample type | High molecular weight genomic DNA | From blood, tissue, cultured cells, leaf, or microbial isolate depending on organism |
| Minimum input – PacBio Revio | ≥ 3–5 µg HMW gDNA | Lower inputs may be discussed for precious samples |
| Minimum input – ONT PromethION | ≥ 1–3 µg HMW gDNA | Depends on ligation vs. rapid library kit |
| DNA integrity | High molecular weight, minimal fragmentation | Avoid repeated freeze–thaw cycles and vigorous vortexing |
| Purity criteria | A260/280 = 1.8–2.0 A260/230 ≥ 2.0 |
Avoid phenol, ethanol, or salt carryover |
| Reference genome | Provided by customer, or selected jointly from public databases | Required for alignment-based WGRS analysis |
| Shipping conditions | Dry ice (preferred) | Ship in DNase/RNase-free tubes with clear labeling |
Why Choose CD Genomics
Dual-Platform Expertise
Long-read WGRS on both PacBio Revio and ONT PromethION, matched to your organism, cohort size, and budget.
Complete Variant Catalogs
SNVs, indels, and structural variants—including complex SVs short reads miss—from a single sequencing run.
Benchmark-Informed Tool Selection
We select variant callers based on published, independent benchmarking rather than a single default pipeline, matching tool choice to your coverage and SV types of interest.
Organism-Agnostic Experience
Human, animal, plant, and microbial WGRS projects are all supported, with dedicated downstream analysis through our human genome structural variation detection and PacBio sequencing data analysis services.
Transparent Reporting
Every project is delivered with annotated variant call files and a clear summary suitable for internal review or a methods section.
Case Study: Benchmarking Long-Read Structural Variant Calling Methods
Liu, Y.H., Luo, C., Golding, S.G. et al. Tradeoffs in alignment and assembly-based methods for structural variant detection with long-read sequencing data. Nature Communications 15, 2447 (2024).
1. Background
Long-read sequencing has driven the development of many structural variant (SV) calling tools, but until this study there was no comprehensive, independent benchmark comparing alignment-based and assembly-based methods across PacBio HiFi, PacBio CLR, and ONT datasets, or across a wide range of sequencing coverages.
2. Methods
The authors systematically compared:
- 12 alignment-based SV callers (including deep learning-based and hybrid methods, such as pbsv, cuteSV, Sniffles2, and DeBreak)
- 4 assembly-based SV callers (Dipcall, Smartie-sv, SVIM-asm, and PAV), built on assemblies from 6 different assemblers including hifiasm
- 4 upstream aligners across 11 real PacBio HiFi, CLR, and ONT datasets, 9 simulated datasets, and 2 paired tumor-normal datasets
Performance was evaluated against the GIAB SV gold-standard callset using Truvari across 31 combinations of evaluation stringency, sequencing coverage (5–56×), and SV type, with orthogonal validation against the T2T-CHM13 reference and a trio-based Verkko assembly.
3. Results
The benchmarking workflow comparing 12 alignment-based and 4 assembly-based SV calling methods across PacBio HiFi, CLR, and ONT datasets, aligners, and assemblers.
Key Findings
- Assembly-based tools excel at large insertions:
- Assembly-based callers detected more large insertions (>1 kb) than most alignment-based tools and were markedly more robust to changes in SV size and evaluation stringency.
- Alignment-based tools win at low coverage:
- pbsv and MAMnet achieved F1 scores of 93.4% and 94.0% respectively at just 5× coverage, while several other alignment-based tools' recall dropped sharply below 20×.
- No universally superior tool:
- Across 31 criteria combinations, the best-performing tool varied by SV type, size, and coverage—the authors provide a user recommendation table rather than a single default.
- Complex SV detection remains hard:
- cuteSV and pbsv performed best for translocations (F1 > 95%), but no tool achieved strong performance for inversions, and duplication detection outside PacBio HiFi data remained weak across all tools.
- Assembly quality matters:
- hifiasm produced the best downstream SV-calling performance among six assemblers tested on HiFi data.
- GIAB benchmark is incomplete:
- Over 20% of insertions flagged as false positives by assembly-based tools were independently supported by the T2T-CHM13 reference and a Verkko trio assembly, suggesting the standard benchmark under-counts true positives.
4. Conclusions
This study demonstrates that tool selection for long-read SV calling should be matched to coverage, SV type, and evaluation goals rather than defaulting to a single pipeline. Importantly:
- Alignment-based callers such as pbsv remain the most robust choice at low-to-moderate coverage (5–10×)
- Assembly-based callers add sensitivity for large insertions and robustness to parameter choice, at higher computational cost
- hifiasm is the strongest assembler for downstream SV calling from PacBio HiFi data
- Complex SV types (inversions, duplications) remain the hardest to call accurately regardless of method
- Orthogonal validation against T2T-CHM13 and trio assemblies can recover true positives missed by standard benchmark sets
FAQs
Q: How is whole-genome resequencing different from de novo sequencing?
Resequencing aligns your reads to an existing reference genome to detect variants. De novo sequencing builds a new genome sequence from scratch, without a reference, and is the right choice when no adequate reference exists for your organism.
Q: Which platform should I choose, PacBio Revio or ONT PromethION?
PacBio Revio is our default recommendation for standard WGRS projects, given its high per-read accuracy. ONT PromethION is a strong option for large cohorts or population-scale studies where throughput and cost per sample are the priority. Our team can help you decide based on your project goals.
Q: How much coverage do I need?
Standard SNV/indel/SV profiling typically works well at 15–20× coverage. Comprehensive SV detection, especially for large insertions, benefits from 20–30× coverage combined with assembly-based calling. Population-scale screening can often use 10–15× coverage, particularly on ONT PromethION.
Q: Can long-read WGRS detect structural variants that short-read sequencing misses?
Yes. Because long reads span repetitive and GC-extreme regions where short reads cannot, they detect and precisely localize structural variants, particularly large insertions, that short-read WGRS frequently misses or mislocalizes.
Q: Do you support microbial and non-model organism resequencing?
Yes. Our microbial whole-genome resequencing service supports bacteria, fungi, and other microbes, and our animal/plant WGRS service covers non-model organisms with a suitable reference genome, as outlined above.
Q: Can I combine WGRS with GWAS or population evolution analysis?
Yes. Our GWAS and population evolution services, described above, build directly on WGRS variant catalogs and can be scoped together as a single project.
References
- Liu, Y.H., Luo, C., Golding, S.G. et al. Tradeoffs in alignment and assembly-based methods for structural variant detection with long-read sequencing data. Nat Commun. 15, 2447 (2024).
- Logsdon, G.A., Vollger, M.R., Eichler, E.E. Long-read human genome sequencing and its applications. Nat Rev Genet. 21, 597–614 (2020).
- Amarasinghe, S.L., Su, S., Dong, X. et al. Opportunities and challenges in long-read sequencing data analysis. Genome Biol. 21, 30 (2020).
For Research Use Only. Not for use in diagnostic procedures.