What a Whole Genome Sequencing Test Costs, and What the Price Buys

Oak
wgs sequencing pricing genomics insurance
Three versions of the same feathered reptile specimen on black plinths, from bare wire frame to fully plumed, under hard studio light.

A whole genome sequenced at 30x and sold direct to consumers costs between about $300 and $600 in 2026. The figure 30x refers to mean depth of coverage, meaning each position in the genome is read about thirty times on average. The same genome ordered through a physician looks quite different on paper. Run in a CLIA-certified laboratory and returned as a clinical report, it lists between roughly $1,000 and $3,000, and what you personally pay depends entirely on whether a payer covers it.

There is also a third product that borrows the name. Low-pass sequencing at 0.4x to 1x, imputed up to common variant calls, sells for $40 to $100. Across all of these, the sequencing chemistry itself accounts for a shrinking minority of the price. The differences you see mostly tell you about depth and library chemistry. They also reflect the deliverables and how much human curation comes attached.

The cost stack behind a $400 genome

It helps to see where the money in a consumer genome goes, because that explains why the price has a floor. At 30x mean coverage of a 3.1 Gb human genome you need roughly 100 Gb of aligned bases. On an Illumina NovaSeq X 25B flow cell at 2x150 bp, that is a small fraction of a lane. Amortized reagent and flow cell cost per genome sits in the low hundreds of dollars at scale. PCR-free library preparation adds perhaps $30 to $60 in enzymes and beads.

Computation is nearly a rounding error by comparison. Alignment and variant calling on a cloud GPU instance with DRAGEN or DeepVariant costs a few dollars of compute per sample. Storage is the quiet recurring line. A 30x CRAM is about 15-20 GB, an equivalent BAM 60-90 GB, and raw gzipped FASTQ 90-120 GB. Any vendor promising indefinite raw data hosting has therefore built a subscription into the model somewhere.

What scales badly is human attention. A clinical genome report involves a bioinformatician reviewing quality metrics, a variant scientist curating candidate variants against ACMG criteria and ClinVar, and a board-certified geneticist signing the report. Literature on pediatric diagnostic sequencing puts the fully loaded cost of genome sequencing with interpretation well above the cost of the sequencing itself. That gap is the single best explanation for why two tests using the same instrument can differ by a factor of five in price.1 Infectious disease genomics shows the same structure from a different angle. Implementation studies there find that instrument and reagent costs are tractable while personnel, validation, and analysis capacity dominate the economics.23

Specifications that change the price, and should change your decision

Before you compare two prices, make the two products comparable, because the word “genome” covers a wide range of data. The specifications below matter in roughly descending order of how much they affect what you can eventually do with the file.

  • Depth, and how it is defined. “30x” is usually mean coverage across the callable genome. Ask instead for the fraction of the genome at ≥20x, which is what determines confident heterozygous calls. A good PCR-free 30x run puts well over 90% of GRCh38 at ≥20x.
  • Library chemistry. PCR-free libraries reduce duplicate rates and GC bias, which matters for copy number and for GC-rich promoters. PCR-based libraries are cheaper, and they are what some sub-$300 offers use.
  • Read length and platform. Short reads at 2x150 bp resolve small variants well and structural variants poorly. Long-read sequencing (PacBio HiFi or Oxford Nanopore) at 15-30x still costs roughly $1,500 to $3,000. It is the only practical way to phase haplotypes across long distances, resolve repeat expansions, and call CYP2D6 hybrids reliably.
  • Reference build. Insist on GRCh38 with ALT-aware alignment. A hg19 deliverable in 2026 means you will spend a weekend on liftover and lose regions to it.
  • Deliverables. Look for FASTQ or CRAM, a gVCF, an annotated small-variant VCF (the standard text format listing the positions where your genome differs from the reference), and ideally separate structural variant, mitochondrial, and pharmacogenomic star-allele outputs. A vendor that ships only a web portal and a PDF has sold you an interpretation rather than a genome.

Verifying what you received

Whatever you paid, it is worth spending an hour confirming that the data matches the specification you were sold. The checks below are quick and they use standard open tools. Start with coverage, using mosdepth to summarize depth across the file:

mosdepth --fast-mode --no-per-base --by 1000 --threads 4 sample sample.cram
awk '$1=="total" && $2>=20 {s+=$3} END {print s}' sample.mosdepth.global.dist.txt

Next, run variant-level sanity checks with bcftools stats -F GRCh38.fa sample.vcf.gz. For a single European-ancestry genome against GRCh38, expect on the order of 3.5-4.5 million SNVs and several hundred thousand indels. The transition/transversion ratio should sit near 2.0. A number far below that range usually means shallow coverage or aggressive filtering, and a number far above it usually means contamination or a broken filter. You can confirm contamination directly with verifybamid2 --SVDPrefix ... --Reference GRCh38.fa --BamFile sample.cram. Anything above about 2% freemix warrants a conversation with the lab.

If you want a real accuracy figure rather than a vendor claim, ask about controls. Specifically, ask whether the lab runs a Genome in a Bottle control (HG002 or HG001) alongside production samples and will share the hap.py or rtg vcfeval output against the GIAB truth set inside the high-confidence regions. Labs that do this routinely will send you the numbers without much fuss. Labs that cannot produce them are telling you something.

Insurance, and why coverage is the volatile part of the price

For a healthy adult buying a genome out of curiosity or for longitudinal baselining, insurance does not apply and the sticker price is the price. For a diagnostic question, coverage is where the cost lands, and it is the least predictable part of the transaction. Payer policies for sequencing are fragmented across plans. They are applied through prior authorization, medical necessity criteria, and CPT coding that has historically lagged the tests being run.4 Studies of hereditary cancer panel ordering found that clinicians’ decisions were shaped directly by coverage rules and by what patients would owe out of pocket. Those pressures fell hardest on safety-net settings.5 For rare disease in particular, coverage for genome sequencing has broadened but remains inconsistent enough that access still depends on payer and geography. That inconsistency is why the field has been arguing publicly for standardized coverage.6

Two practical consequences follow. The first is procedural. Ask the ordering lab for a pre-authorization determination and a written patient-responsibility estimate before the sample ships, because the difference between “covered” and “denied” on the same test can be two thousand dollars.

The second concerns what happens after the results exist. Think through the downstream disclosure of results before you generate them. Genetic information is durable, familial, and hard to un-share. The concerns about discrimination and insurance that shaped US policy decades ago remain the reason those protections exist.7 The broader question of what individuals and families do with expanded genomic information, including incidental findings nobody asked for, is a live and unresolved one.8

What to expect on price in 2026

Looking ahead, we expect the consumer 30x floor to settle near $200 to $300 and then stop falling meaningfully. At that point shipping, extraction, storage, and support dominate the cost. Clinical genomes will not track that curve down, since their cost is curation labor rather than chemistry.

Long-read genomes are the interesting line to watch. If HiFi or nanopore 30x reaches the $800 range, it becomes the default recommendation for anyone who wants one genome for life. A phased long-read assembly answers questions that short reads cannot revisit later.

For a technically fluent buyer, our recommendation is to pay for a PCR-free 30x short-read genome with raw CRAM or FASTQ delivered on GRCh38. Verify it yourself with the checks above, and treat the vendor’s interpretation layer as disposable. Interpretation improves every year, while a well-generated raw file does not expire.

One caveat sits on top of all of this. If a variant you find has clinical implications, do not act on it from a research-grade file. Take it to a clinical geneticist or genetic counselor for orthogonal confirmation in a CLIA laboratory. Consumer pipelines are not validated for that purpose, and false positives in single genes are common enough to matter.

Oak builds longitudinal molecular profiles of individuals: whole-genome sequencing, RNA sequencing, proteomics, blood biomarkers, and continuous glucose data, integrated into one model of you. Build your profile.

Footnotes

  1. Michael P. Douglas, Patricia A. Deverka, Bruce Gelb, et al. Report – Cost and Clinical Utility of WES and WGS in pediatric patients with suspected genetic disease. 2022. https://doi.org/10.1101/2022.10.10.22280925 ↩

  2. Monica Vogel, Christian Utpatel, Caroline Corbett, et al. Implementation of whole genome sequencing for tuberculosis diagnostics in a low-middle income, high MDR-TB burden country. Scientific Reports, 2021. https://doi.org/10.1038/s41598-021-94297-z ↩

  3. My Tran, Kayla S Smurthwaite, Son Nghiem, et al. Economic evaluations of whole-genome sequencing for pathogen identification in public health surveillance and health-care-associated infections: a systematic review. The Lancet Microbe, 2023. https://doi.org/10.1016/s2666-5247(23)00180-5 ↩

  4. Kris Rickhoff, Andrew Drury, John Pfeifer. Billing and Reimbursement. Clinical Genomics, 2015. https://doi.org/10.1016/b978-0-12-404748-8.00026-5 ↩

  5. Grace A. Lin, Julia R. Trosman, Michael P. Douglas, et al. Influence of payer coverage and out‐of‐pocket costs on ordering of NGS panel tests for hereditary cancer in diverse settings. Journal of Genetic Counseling, 2021. https://doi.org/10.1002/jgc4.1459 ↩

  6. Vaidehi Jobanputra, Brock Schroeder, Heidi L. Rehm, et al. Advancing access to genome sequencing for rare genetic disorders: recent progress and call to action. npj Genomic Medicine, 2024. https://doi.org/10.1038/s41525-024-00410-2 ↩

  7. Kathy L. Hudson, Karen H. Rothenberg, Lori B. Andrews, et al. Genetic Discrimination and Health Insurance: An Urgent Need for Reform. Science, 1995. https://doi.org/10.1126/science.270.5235.391 ↩

  8. Pascal Borry, Heidi Beate Bentzen, Isabelle Budin-Ljøsne, et al. The challenges of the expanded availability of genomic information: an agenda-setting paper. Journal of Community Genetics, 2017. https://doi.org/10.1007/s12687-017-0331-7 ↩