Human Whole Genome Sequencing · ONT PromethION P2i & P24

One assay, the whole genome: variants, phasing, and methylation.

Short-read WGS averages away exactly the variants that matter most: structural variants, repeat expansions, phasing, and methylation. Oxford Nanopore reads native DNA molecules directly — tens to hundreds of kilobases long — so one genome delivers variants, haplotype, and methylation in a single workflow, with real-time basecalling on every run.

~125–160 Gb
per flow cell (~40–50× human)
99.8%
modal raw read accuracy
Native 5mC/5hmC
methylation in every run
P2i + P24
— benchtop to production scale

Human WGS on Oxford Nanopore at Nucleome Informatics

Nucleome Informatics offers human whole genome sequencing on the Oxford Nanopore PromethION platform family, operating both the PromethION 2 Integrated (P2i) — a compact benchtop device running up to two PromethION Flow Cells with onboard GPU basecalling — and the PromethION 24 (P24) — the production-scale sequencer running up to 24 independent flow cells. Both run current PromethION Flow Cells (FLO-PRO114M) with Q20+ chemistry and Dorado super-accuracy basecalling.

A single PromethION Flow Cell yields ~125–160 Gb in a 72-hour run — roughly 40–50× coverage of a human genome — so one flow cell delivers one high-quality human genome. The P2i covers pilots and small cohorts; the P24 scales to population-level programmes without batching.

  • Raw (FAST5/POD5) and basecalled (FASTQ/BAM) reads with full run QC (yield, read-length N50, modal accuracy)
  • Alignment to GRCh38 or T2T-CHM13 (minimap2) with coverage and QC metrics
  • Small variant calling: SNVs and indels (Clair3 / DeepVariant), with deep-annotation (VEP)
  • Structural variant calling (Sniffles2 / cuteSV) and copy-number analysis
  • Short tandem repeat (STR) expansion genotyping
  • Haplotype-phased VCF (WhatsHap) — variants assigned to parental haplotypes
  • Genome-wide methylation calls (5mC/5hmC) from the same reads — no bisulfite conversion
  • A written analysis report with methods, versions, and interpretation notes

Why nanopore for human WGS?

  • Structural variants are first-class citizens: long reads detect deletions, duplications, insertions, inversions, and complex rearrangements that short reads systematically miss.
  • Phasing comes free: reads span multiple variants, so the phased VCF is built from single molecules — not statistical inference.
  • Methylation in the same run: native base modifications (5mC, 5hmC, 6mA, 4mC) are detected directly from the reads, at up to 99.7% accuracy per Oxford Nanopore — no separate bisulfite or array assay.
  • Real-time everything: basecalling happens during the run; data streams as it is produced.

From first principles to expert-level analysis

New here? Start with Foundations. Planning a study or reviewing results? Jump to Practitioner or Expert. Each level builds on the one before it, so you can read this page as a short course or dip in where you need to.

Start here if you are new to this service — six ideas that explain what it is and why it matters.

LEVEL 1

What is whole genome sequencing?

Reading the entire human genome — all ~3 billion bases — rather than just the exome's coding regions. WGS captures coding and non-coding variation, structural variants, and copy number in one assay.

LEVEL 1

What is different about nanopore sequencing?

Nanopore reads single DNA molecules as they pass through protein pores — natively, without amplification. Reads are long (tens to hundreds of kilobases) and basecalled in real time as the run proceeds.

LEVEL 1

How much data does a human genome need?

About 100 Gb for 30× coverage. A single PromethION Flow Cell yields ~125–160 Gb in a 72-hour run — roughly 40–50× coverage — so one flow cell delivers one genome with margin.

LEVEL 1

What are P2i and P24?

Two members of Oxford Nanopore's PromethION family. The P2i is a compact benchtop device running up to two flow cells with onboard GPU basecalling; the P24 runs up to 24 flow cells for production-scale throughput. Both use the same flow cells and chemistry, so data quality is equivalent — only scale differs.

LEVEL 1

What does "native methylation calling" mean?

The sequencing reads preserve the DNA's chemical modifications, so methylation (5mC, 5hmC) is detected directly from the same reads — no bisulfite conversion, no separate assay.

LEVEL 1

What sample do I provide?

High-molecular-weight DNA (typically from blood), or whole blood for extraction at our laboratory. HMW integrity is the main quality determinant — we provide extraction protocols or handle extraction.

For scientists planning a project or reading a report: how each step works and how it can mislead.

LEVEL 2

Library preparation

HMW DNA is prepared with ligation-based kits (SQK-LSK114 family) with optional native barcoding (24- or 96-plex) for multiplexing. No PCR is required in standard workflows, preserving methylation and avoiding amplification bias.

LEVEL 2

Sequencing on P2i and P24

Flow cells run up to 72 hours; basecalling happens in real time on onboard GPUs (Dorado super-accuracy). The P2i runs two flow cells independently; the P24 runs up to 24 — flow cells can be started individually, so projects are never batch-locked.

LEVEL 2

Small variant calling

Reads are aligned with minimap2 and variants called with Clair3 or DeepVariant. The long-read human variation workflow benchmarks at 99.8% SNV F1 accuracy against truth sets.

LEVEL 2

Structural variants and repeats

Sniffles2/cuteSV detect deletions, duplications, insertions, and inversions from split-read evidence; STR expansion genotyping resolves tandem repeats — including loci short reads cannot size.

LEVEL 2

Phasing and methylation

WhatsHap builds the phased VCF from read-span evidence; methylation calls are extracted with modkit from the same basecalled reads. Both come from the primary data — no extra assays.

LEVEL 2

Quality reporting

Every genome ships with yield, read-length N50, modal accuracy, coverage uniformity, and per-region coverage metrics — plus honest flags where coverage falls short of targets.

For study design: the decisions that determine whether the data answers your question.

LEVEL 3

Coverage planning

30× is the standard for germline WGS; flow-cell yields of ~125–160 Gb give 40–50× with margin for failed pores and QC losses. For structural-variant-focused studies, additional depth improves sensitivity; we scope depth per design.

LEVEL 3

Reference choice: GRCh38 vs T2T-CHM13

T2T-CHM13 resolves acrocentric short arms and segmental duplications that GRCh38 leaves gappy — but downstream annotation ecosystems remain GRCh38-centric. We support both and report per-reference metrics.

LEVEL 3

Pharmacogenomics and complex loci

Long reads resolve the loci short reads cannot: CYP2D6 copy number and hybrid alleles, HLA haplotypes, GBA pseudogene fusions, and SMN1/2 copy number — discuss targeted follow-up with our PureTarget workflows where relevant.

LEVEL 3

Rare disease applications

Long-read WGS is the single most comprehensive first-line assay for suspected structural-variant or repeat-expansion disease; trio designs with phasing strengthen interpretation and de novo calling.

LEVEL 3

Methylation-aware WGS

Because methylation comes free with every run, cohort studies can layer epigenetic analysis onto germline variant discovery without additional sequencing — a genuinely unique capability of the platform.

LEVEL 3

Data management

A 40× human genome generates ~100 GB of reads plus analysis artefacts. We deliver through secure transfer with structured metadata, and support submission-ready formatting for controlled-access repositories.

P2i
PromethION 2 Integrated; benchtop sequencer running up to two PromethION Flow Cells with onboard GPU basecalling.
P24
PromethION 24; production-scale sequencer running up to 24 independent flow cells.
Modal raw read accuracy
the most common per-read accuracy value across a run (99.8% with current chemistry).
Dorado
Oxford Nanopore's GPU basecaller; super-accuracy models are the current standard.
Phasing
assigning variants to parental haplotypes.
modkit
Oxford Nanopore's tool for processing base-modification calls.

How it works, in two figures

Oxford Nanopore's human variation workflow benchmarks 99.8% SNV F1 and 95.8% SV CDS F1 accuracy with current chemistry; long-read WGS is established for structural-variant and repeat detection, native methylation calling, and haplotype-resolved genotyping. See references below.

Figure 1 Short reads vs long reads. schematic
Short-read WGS Nanopore long-read WGS Averages across repeats and SVs No phasing without inference Methylation needs a separate assay Single reads span SVs and repeats Phased VCF from read-span evidence Native 5mC/5hmC in every run
Figure 1 | Short reads vs long reads. (Illustrative schematic, not sample data.)
Figure 2 Workflow. schematic
HMW DNA Ligation library (± barcoding) P2i / P24 sequencing (72 h) Real-time Dorado basecalling SNV + SV + STR calling Phasing + methylation report
Figure 2 | Workflow. (Illustrative schematic, not sample data.)

Built for the questions you are asking

The platforms, pipelines and standards behind this service.

PromethION 2 Integrated (P2i)

Compact benchtop device running up to two PromethION Flow Cells with onboard NVIDIA Ampere GPU basecalling, 15 TB storage, and full real-time data handling — the flexible engine for pilots and small cohorts.

PromethION 24 (P24)

Production-scale sequencer running up to 24 independent flow cells with an A-Series compute tower (4 NVIDIA GPUs, 512 GB RAM, 60 TB SSD) — terabases per run for population-scale programmes, with no batching required.

Q20+ chemistry on FLO-PRO114M

Current PromethION Flow Cells with Q20+ reagents deliver modal raw read accuracy of 99.8% and flow-cell yields of ~125–160 Gb for human libraries.

Dorado super-accuracy basecalling

Real-time GPU basecalling with super-accuracy models, plus native base-modification calling (5mC, 5hmC, 6mA, 4mC) in every run.

Complete variant toolchain

Clair3/DeepVariant for SNVs and indels, Sniffles2/cuteSV for structural variants, STR genotyping, and WhatsHap phasing — the long-read human variation workflow with published benchmark accuracy (99.8% SNV F1; 95.8% SV CDS F1).

T2T-CHM13-ready alignment

Assemblies and references of the completeness era need long reads to exploit them: we align to GRCh38 or T2T-CHM13 and report per-region coverage, including the acrocentric short arms and segmental duplications short reads cannot map. ---

What this service does and does not cover

Scope note. Nanopore human WGS delivers comprehensive germline variant discovery — SNVs, indels, structural variants, STR expansions, copy number, phasing, and methylation — from one assay. Per-read accuracy (99.8% modal) is below short-read rates; small-variant calling compensates via depth and consensus, and benchmark accuracy against truth sets is high (99.8% SNV F1), but single-read-level error means homopolymer-associated indels require careful calling — we report caller-specific caveats honestly. Methylation calls are research-grade base-modification detections, not bisulfite-equivalent quantitative assays. Clinical reporting pathways are agreed at intake with the requesting clinician. Turnaround times and pricing are defined per project.---

Service highlights

What you can expect from our team in Hyderabad.

Both scales in-house

P2i for pilots and small cohorts; P24 for population-scale programmes — same chemistry, same data quality.

One assay, five data layers

Variants, structural variants, STRs, phasing, and methylation from a single sequencing run.

Real-time basecalling

Dorado super-accuracy models run on onboard GPUs — data streams as it is produced.

India's PromethION service provider

Nucleome Informatics operates PromethION platforms in Hyderabad with full analysis support.

T2T-ready

Alignment to GRCh38 or T2T-CHM13, with per-region coverage reporting.

Consultative design

Coverage, reference, and analysis plan agreed before sequencing — with honest feasibility scoping. ---

Frequently asked, in plain language

From the basics to the finer points of design and interpretation.

basicWhat is the difference between the P2i and P24?
Scale only. The P2i runs up to two PromethION Flow Cells with onboard GPU basecalling; the P24 runs up to 24. Both use identical flow cells and chemistry, so data quality is the same — the P24 adds throughput for large cohorts.
basicHow many genomes fit on one flow cell?
One high-quality human genome. Flow-cell yields of ~125–160 Gb give roughly 40–50× human coverage in a 72-hour run.
basicWhat do I get in the report?
Raw and basecalled reads, alignment, SNV/indel calls, structural variant calls, STR genotypes, a phased VCF, methylation calls, and full QC — with a written interpretation report.
intermediateHow accurate is nanopore WGS compared with short reads?
Modal raw read accuracy is 99.8%. Small-variant calling against truth sets benchmarks at 99.8% SNV F1 — comparable to short-read workflows — while structural variant detection is substantially better. Homopolymer-associated indels need careful calling; we report caller-specific caveats.
intermediateIs methylation really included?
Yes — native base modifications (5mC, 5hmC, 6mA, 4mC) are detected from the same reads in every run, at up to 99.7% accuracy per Oxford Nanopore. No bisulfite conversion or separate assay is needed.
intermediateCan you phase my genome?
Yes — long reads span multiple variants per molecule, so WhatsHap builds the phased VCF directly from read-span evidence rather than statistical inference.
advancedCan nanopore WGS resolve CYP2D6, HLA, or SMN1 copy number?
Yes — these are exactly the loci where long reads excel. For focused pharmacogenomic or carrier questions, we complement WGS with our PureTarget workflows for targeted, higher-depth resolution.
advancedGRCh38 or T2T-CHM13 — which reference?
We support both. T2T-CHM13 resolves regions GRCh38 leaves gappy; GRCh38 retains the richest annotation ecosystem. We report per-reference metrics and can deliver both alignments.
advancedHow does multiplexing work?
Native barcoding kits support 24- or 96-plex designs. On the P24, up to 24 flow cells run independently — cohorts sequence without batching delays.
advancedWhat are the data volumes?
A 40× genome generates ~100 GB of reads plus analysis artefacts. We deliver via secure transfer with structured metadata and support controlled-access repository submission.

Key literature & resources

Selected sources behind the methods and reference values on this page. Figures labelled "schematic" or "illustrative" are generated for this page, not client data.

  1. Oxford Nanopore Technologies. PromethION 2 Integrated (P2i) specification sheet and technical specification. ONT documentation, 2024–2025.
  2. Oxford Nanopore Technologies. PromethION 24 / PromethION A-Series technical specification. ONT documentation, 2024–2025.
  3. Oxford Nanopore Technologies. Human variation end-to-end workflow (accuracy benchmarks: 99.8% SNV F1; 95.8% SV CDS F1). nanoporetech.com/application-workflows, 2024–2025.
  4. Zheng Z, et al. Symphonizing pole talk and base talk for whole-genome sequencing (Clair3). Nature Biotechnology. 2022;40(4):596–602.
  5. DePristo MA, et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data (DeepVariant: Poplin R, et al. Nature Biotechnology. 2018;36(10):987–990).
  6. Sedlazeck FJ, et al. Accurate detection of complex structural variations using single-molecule sequencing (Sniffles). Nature Methods. 2018;15(6):461–468. (Sniffles2: Smolka M, et al. Nature Biotechnology. 2024.)
  7. Nurk S, et al. The complete sequence of a human genome (T2T-CHM13). Science. 2022;376(6588):eabj6987.
  8. De Coster W, et al. — see also: De Coster W, Rademakers R. NanoPack: porechop and Filtlong tools for nanopore data. (Read QC tooling context.)