Human Whole-Genome Sequencing — Long-Read Structural Variant Discovery & Complete Genome Analysis by PacBio & Nanopore

Comprehensive human genome sequencing on PacBio HiFi (Revio / Sequel II) and Oxford Nanopore PromethION — delivering genome-wide structural variant detection, telomere-to-telomere genome analysis, haplotype phasing, and integrated epigenome profiling for rare disease research, cancer genomics, and population-scale studies

CD Genomics provides end-to-end human whole-genome sequencing services on PacBio HiFi (Revio/Sequel II) and Oxford Nanopore PromethION platforms — from HMW DNA extraction and library preparation through multi-caller structural variant detection, tandem repeat analysis, mobile element insertion discovery, de novo assembly, haplotype phasing, and pharmacogenomic annotation — delivering genome-wide SV detection at nucleotide resolution from a single sequencing experiment.

Human whole-genome sequencing (WGS) by long-read technology represents a fundamental advance in our ability to detect the full spectrum of human genetic variation. While short-read sequencing has been the dominant approach for the past two decades, its inherent read-length limitation (150–300 bp) means that approximately 5–10% of the human genome — including most structurally complex regions — remains inadequately resolved. Long-read sequencing platforms from Pacific Biosciences (HiFi on Revio / Sequel II) and Oxford Nanopore (PromethION) produce reads of 15–100+ kilobases, providing the single-molecule resolution needed to characterize structural variants (SVs), tandem repeats, mobile element insertions, and complex rearrangements that are systematically missed or mischaracterized by short-read approaches.

The impact of this technological shift is substantial. Large-scale studies such as the All of Us research program have demonstrated that long-read WGS at moderate coverage (8–10× HiFi) detects over twice as many disease-associated SVs as short-read WGS at 30× coverage, revealing previously hidden genetic contributions to rare and common diseases. Long-read WGS also provides haplotype-resolved genome data, direct detection of DNA base modifications without chemical pretreatment, and the ability to produce telomere-to-telomere genome assemblies — delivering a comprehensive genomic picture from a single sequencing experiment.

At CD Genomics, we provide human whole-genome sequencing services across both major long-read platforms — PacBio HiFi (Revio, Sequel II) and Oxford Nanopore PromethION — with coverage tailored to the specific requirements of each project. Whether your research involves rare disease variant discovery, somatic SV detection in cancer, population-scale genome characterization, or pharmacogenomic profiling, our team will design and execute a sequencing and analysis strategy optimised for your research objectives.

Why Choose Our Human WGS Service

• Comprehensive Structural Variant Detection — Long reads (15–100+ kb) capture all SV classes: inversions, translocations, mobile element insertions, tandem repeat expansions, and complex rearrangements that short-read WGS systematically fails to detect or resolve at nucleotide resolution. Refer to our structural variant detection service for detailed methodology and benchmark results
• Dual-Platform Flexibility — PacBio HiFi (Q30+ consensus accuracy, 15–25 kb reads) for precision variant calling and diploid phasing; ONT ultra-long reads (50–100+ kb) for spanning complex structural rearrangements; or integrated multi-platform strategies for the most comprehensive genome analysis
• Clinical Research-Grade Coverage — Coverage targets from 10–30× depending on application (germline variant discovery, somatic mutation detection, population screening), with recommendations derived from published benchmark studies and platform-specific performance characteristics
• Integrated Epigenome from a Single Run — Both PacBio HiFi and ONT sequencing natively detect DNA base modifications (5mC, 6mA, 4mC) from the same sequencing data without bisulfite conversion or antibody enrichment — simultaneous genome and epigenome characterization without additional experimental costs
• Scalable from Single Genomes to Population Cohorts — Multiplexed sample barcoding, flexible coverage depths, and tiered bioinformatics analysis packages enable cost-effective projects of any scale, from individual genome characterisation to cohort studies spanning hundreds of samples
• End-to-End Project Management — From sample QC and sequencing strategy design through primary and secondary analysis, variant interpretation, and publication-ready reporting — a dedicated project scientist manages every stage with regular progress updates

What Is Human Whole-Genome Sequencing by Long-Read Technology — and Why It Matters

Human whole-genome sequencing aims to determine the complete DNA sequence of a human genome, covering all 22 autosomes, the X and Y chromosomes, and the mitochondrial genome. The goal is not simply to produce a sequence, but to generate a comprehensive and accurate catalog of all genetic variation present in the sample — from single-nucleotide variants (SNVs) and small insertions/deletions (InDels) to large structural variants, repeat expansions, and mobile element insertions — and to place this variation in its haplotype context.

The limitations of short-read WGS for this task are now well documented. The 150–300 bp reads produced by Illumina sequencers are shorter than most human repetitive elements and structural variant breakpoints, making it impossible to uniquely map reads across these regions. Studies comparing long-read and short-read WGS in the same individuals have consistently shown that short-read data misses 50–70% of SVs, particularly insertions (which are the most common SV type in the human genome), tandem repeat expansions, and complex rearrangements involving segmental duplications. For clinical applications, this means that patients with SV-mediated genetic diseases may remain undiagnosed after short-read WGS, and potentially actionable pathogenic variants remain hidden in uncharacterized genomic regions.

Long-read WGS overcomes this limitation fundamentally. A single PacBio HiFi read of 15–25 kb or an ONT read of 50–100+ kb can span an entire SV breakpoint junction, resolve a tandem repeat array, or bridge a segmental duplication to place structural variation in its correct genomic context. With coverage as low as 10–15× HiFi, long-read WGS achieves comprehensive SV detection across the genome. At higher coverage (20–30×), it enables de novo assembly of individual genomes, haplotype-level phasing, and complete characterization of the 5–10% of the genome that is inaccessible to short reads.

At CD Genomics, our human WGS service is designed to deliver the full value of long-read sequencing — whether that means comprehensive SV discovery at moderate coverage for cohort studies, high-coverage de novo assembly for reference-quality genomes, or integrated genome and epigenome analysis for multi-omics research projects.

Long-Read Human WGS Applications Span Rare Disease, Cancer Genomics, Population Studies, and Pharmacogenomics

Beyond rare disease diagnosis, cancer genomics, and population studies, long-read WGS is increasingly applied in resolving somatic mosaicism and age-related clonal hematopoiesis, where detection of structural variants present at low variant allele fractions (5–10%) in a subset of cells requires the single-molecule resolution that only long-read sequencing provides. Mosaic SVs — including mobile element insertions, complex rearrangements, and large deletions — arise in repetitive and duplicated genomic regions that are systematically inaccessible to short-read SV detection, making long-read WGS the only current approach capable of characterizing the full spectrum of somatic structural variation in aging and diseased tissues.

Rare Genetic Disease Research

• Comprehensive SV detection in short-read-negative rare disease cohorts — long-read WGS reanalysis increases diagnostic yield by 5–15% by detecting SVs, repeat expansions, and complex rearrangements missed by short-read clinical pipelines
• Nucleotide-resolution breakpoint mapping of disease-causing structural variants, enabling precise genotype-phenotype correlations and family segregation analysis that is impossible with imprecise short-read SV calls
• Repeat expansion disorder characterization — long reads spanning full repeat arrays enable accurate sizing of pathogenic repeat expansions in genes associated with neurodevelopmental and neurodegenerative disorders
• Haplotype-resolved compound heterozygote resolution — phasing of two pathogenic variants in the same gene to determine whether they are in cis or trans, directly informing recessive disease diagnosis

Cancer Genomics

• Somatic SV detection in tumor genomes — long-read WGS detects gene fusions, intragenic rearrangements, and complex genomic rearrangements (chromothripsis, chromoplexy, templated insertion chains) that are hallmark signatures of many cancer types but are poorly resolved by short-read sequencing
• Tumor-normal paired analysis for somatic-germline discrimination, with phased haplotypes enabling identification of loss of heterozygosity and copy-number-neutral events in cancer genomes
• Integrated genome and epigenome characterization — simultaneous detection of somatic mutations and cancer-specific methylation changes from a single sequencing experiment

Population & Evolutionary Genomics

• Genome-wide SV discovery and frequency estimation in diverse populations — long-read WGS reveals the full spectrum of structural variation that shapes human genetic diversity, including population-specific SVs and complex variants that define haplotype structure
• Pangenome reference construction — haplotype-resolved assemblies from diverse individuals enable the construction of pangenome graphs that represent the full scope of human genetic variation, replacing the single linear reference genome paradigm
• Admixture and evolutionary analysis using phased haplotype data for understanding population history, selection signatures, and the evolutionary origins of structural variant alleles

Pharmacogenomics

• Complete characterization of complex pharmacogene loci (CYP2D6, CYP2B6, UGT1A, HLA region) — long-read WGS resolves gene duplications, deletions, hybrid genes, and structural rearrangements that determine drug metabolism phenotypes but are systematically misassigned by short-read genotyping
• Haplotype-resolved star allele determination — phased long reads enable accurate diplotyping of pharmacogenes, directly informing predicted drug metabolism phenotypes for clinical research applications

Our Bioinformatics Delivers SV Detection, De Novo Assembly, Methylation Calling, and Pharmacogenomic Annotation from Human WGS Data

Our bioinformatics pipeline for human WGS data supports comprehensive genome analysis from raw read processing through multi-platform variant detection, genome assembly, and advanced interpretation — organized across two service tiers to accommodate different research objectives and budgets.

Sample Input Requirements for Human WGS — DNA Quality, Coverage Targets, and Multiplexing Options

Case Study: Evaluation of PacBio and Oxford Nanopore Single-Molecule Sequencing for Structural Variant Detection in Two Swedish Human Genomes

Fatima N, Petri A, Gyllensten U, Feuk L, Ameur A. Evaluation of Single-Molecule Sequencing Technologies for Structural Variant Detection in Two Swedish Human Genomes. Genes. 2020;11(12):1444.

1. Background

Structural variants are a major source of genetic diversity in the human genome and are increasingly recognised as contributors to both rare and common diseases. Despite their importance, SV detection has lagged behind SNV/InDel detection — largely because short-read sequencing technologies, which produce reads shorter than most SV breakpoint-spanning distances, detect only a fraction of SVs present in any given genome. Long-read single-molecule sequencing technologies from Pacific Biosciences and Oxford Nanopore Technologies promised to overcome this limitation, but at the time of this study (2020), direct head-to-head comparisons of the two platforms for human WGS SV detection in the same individuals were limited.

In this study, Fatima et al. sequenced two Swedish human genomes (Swe1, male; Swe2, female) on both ONT PromethION (~32× coverage) and PacBio Sequel II (HiFi and CLR modes, ~66× CLR coverage, downsampled to ~33× for comparison) and systematically evaluated the SV detection performance of each platform.

2. Methods

DNA from two Swedish individuals (Swe1 and Swe2, selected from the SweGen project) was sequenced on ONT PromethION (six flow cells per individual, ~32× average coverage) and PacBio Sequel II (eight SMRT Cells per individual, ~66× CLR coverage). PacBio data was also processed in HiFi CCS mode. SV calling was performed using minimap2 for alignment and Sniffles for SV detection, with additional validation using PBSV (PacBio) and NanoSV (ONT). Overlap between platforms was assessed at multiple coverage levels, including downsampling of PacBio data to match ONT coverage (~33×) for fair comparison.

3. Results

Figure 2. Comparison of structural variant detection by PacBio and ONT in two Swedish human genomes. ONT detected ~17,000 SVs per genome, while PacBio detected ~23,000 SVs per genome at full coverage. Downsampling analysis showed that the majority of platform-specific SVs were detected at lower coverage by the other platform, and that combining both technologies provides the most comprehensive SV discovery. Adapted from Fatima et al. (2020), Genes, CC BY 4.0.

Key Findings

• Platform-specific SV detection: ONT detected an average of ~17,000 SVs per genome (primarily deletions and insertions), while PacBio detected ~23,000 SVs per genome — with both platforms detecting the majority of large SVs (>1 kb) but differing substantially in their sensitivity for smaller SVs and SVs in repetitive regions.
• Coverage normalisation: After downsampling PacBio data to match ONT coverage (~33×), the number of SVs detected by PacBio decreased to ~17,000 — closely matching ONT yield — indicating that coverage depth rather than platform chemistry was the primary driver of differences in total SV count.
• Platform overlap and complementarity: At matched coverage, 60–70% of detected SVs were shared between platforms. The remaining platform-specific calls were enriched in different genomic contexts: PacBio-specific SVs were enriched in GC-rich regions, while ONT-specific SVs were enriched in long repetitive elements and segmental duplications — demonstrating the complementary nature of the two technologies for comprehensive human genome analysis.
• Insertion detection: Both platforms detected a substantial fraction of insertions relative to the reference genome, with ONT detecting more insertions anchored in repetitive elements (LINE1, SINE/Alu) due to its longer read lengths, while PacBio provided more precise breakpoint resolution for the shared insertion calls.

4. Conclusions

This study demonstrated that ONT and PacBio long-read sequencing have similar overall performance for SV detection in human WGS data, with platform-specific differences driven primarily by read length (ONT advantage for spanning long repeats) and per-base accuracy (PacBio advantage for precise breakpoint resolution). The findings established that coverage depth is the single most important determinant of SV detection sensitivity for both platforms and that combining data from both technologies provides the most comprehensive view of human structural variation. These results continue to inform platform selection and coverage design for human WGS projects today.

Interpretation Boundaries for Long-Read Human Whole-Genome Sequencing

• SV detection sensitivity varies by variant class, size, and genomic context. While long-read WGS dramatically outperforms short-read for SV detection, sensitivity remains variant class-dependent: large insertions and deletions (>1 kb) are detected with high sensitivity at 15× coverage, while smaller SVs (<50 bp), SVs in centromeric satellite arrays, and SVs in recently duplicated segmental duplications may still be incompletely resolved. Detection rates for each variant class are documented in our project reports
• Coverage depth determines detection sensitivity, and optimal coverage is application-specific. Lower coverage (8–15×) is sufficient for population-scale SV discovery but may miss smaller SVs and lower-confidence breakpoints. Higher coverage (25–30×) is required for de novo assembly and the highest confidence variant calls. Coverage recommendations reflect published benchmark data and should be interpreted as project design guidance, not absolute guarantees
• Haplotype phasing accuracy depends on heterozygosity and read length. Phasing performance is highest in regions of average heterozygosity (~0.1% in human genomes) and decreases in homozygous or low-complexity regions. Phasing completeness is reported for each project with per-chromosome and per-genome statistics
• DNA methylation detection from native long-read signal is probabilistic. Both PacBio kinetic and ONT current-based modification calling produce per-site and per-read modification probability scores. These calls are statistically derived and may be affected by basecalling model version, coverage depth, and sequence context. Methylation frequencies at CpG sites with <10× coverage should be interpreted with caution
• Reference genome choice influences alignment-based results. Alignment to GRCh38 vs. T2T-CHM13 produces different variant call sets, particularly in previously unresolved regions. We provide alignment to the reference genome specified by the client and document any reference-specific effects on variant detection in the project report
• All human WGS data are generated for research use only. Our sequencing and analysis services are intended for research applications and are not validated for clinical diagnosis, treatment decisions, or clinical variant interpretation in individual patients or specimens. Data privacy and security are managed according to agreed protocols as documented in each project's data handling plan

What You Get — Human WGS Deliverables & Report Examples

Deliverable Examples for Human Whole-Genome Sequencing Projects

1. Raw sequencing data files (FASTQ, POD5 or BAM) with basecalling quality metrics, coverage statistics, and alignment reports (BAM/CRAM files aligned to GRCh38 or T2T-CHM13 reference).

2. Variant call files (VCF/BCF) for all detected variant classes: SVs (with breakpoint coordinates, SV type, and allele frequency estimates), SNVs/InDels, tandem repeat genotypes, mobile element insertions, and copy number variants — with variant quality scores and annotation.

3. Bioinformatics analysis report including: comprehensive variant summary with classification by type and size, SV overlap analysis between detection methods, genome-wide variant density plots, methylation frequency tracks (if requested), and comparison with known population frequency databases.

4. Publication-ready figures and NCBI submission support — including IGV validation screenshots for key variants, genome-wide SV distribution plots, and BioProject submission files for data release.

Figure 3. Representative deliverable formats for human WGS projects. Left: genome-wide SV distribution and classification. Center: multi-caller SV comparison and validation. Right: comprehensive variant annotation and analysis report. AI-generated representative data.