What Whole Genome Sequencing Costs in 2026

Oak
wgs genomics sequencing-cost bioinformatics data-ownership
Three glass plinths in a dark studio hold the same winged feathered creature, progressing from rough pale cast to fully iridescent living animal.

A 30x whole genome delivered as FASTQ, BAM, and VCF files costs roughly $200 to $1,000 in 2026. That assumes you are buying reagents and instrument time from a research core or a consumer reseller. Those three formats are the standard handoffs: FASTQ holds the raw sequencing reads as they come off the instrument, BAM holds those reads after they have been aligned to a reference genome, and VCF lists the positions where your genome differs from that reference. A clinically reported genome ordered through a diagnostic laboratory, with a board-certified geneticist signing an interpretation, costs $1,500 to $5,000 and is often billed to insurance.

The spread between those numbers is not explained by DNA chemistry. Sequencing the molecules is now the cheap part of the exercise. What you pay for above $300 is depth and coverage guarantees. It also covers read length, sample handling under CLIA and CAP regulation, and human interpretation. The sections below walk through each price tier and what the technical parameters buy you. They also cover what happens to the data after it leaves the sequencer and what we would purchase in each situation.

The three price tiers and what separates them

Prices for whole-genome sequencing cluster into three bands, and the differences between them are mostly differences in what surrounds the sequencing rather than the sequencing itself. Understanding which band you are shopping in tells you what you will receive and what you will not.

The lowest tier, $200 to $250, is what an academic sequencing core charges a lab that walks in with extracted DNA. The University of Minnesota Genomics Center’s advertised $221 human genome is a real number, and it is real because you are buying a slice of a NovaSeq X flow cell at cost plus overhead. You supply DNA at the required concentration and volume, you accept the core’s library prep, and you receive FASTQ files. There is no consent counseling and no sample collection kit. There is no clinical validation of the pipeline, and no one to call if a variant frightens you.

The middle tier, $400 to $1,000, is the consumer reseller market: Nebula, Dante Labs, Sequencing.com, and similar companies. They ship a saliva or cheek swab kit and run extraction and library prep. They sequence somewhere, often in Europe or Asia to arbitrage lab costs, then hand back a BAM and VCF plus a web report. The promotional prices advertised at $399 are usually for 30x short-read sequencing, with a caveat worth reading closely. The figure “30x” means each position in the genome is read an average of thirty times, but some vendors count that as raw coverage before duplicate removal and quality filtering, which can land you at 22–25x of usable depth. Ask specifically for mean mapped, deduplicated coverage and the fraction of the genome at ≥20x. A good short-read genome gets 95% or more of the callable genome to 20x.

The top tier is clinical. GeneDx, Invitae, Baylor Genetics, and hospital labs charge $1,500 to $5,000 for a proband genome, meaning the affected individual, and more for trio sequencing of a child plus both parents. The list price reflects CLIA-certified sample chain of custody and Sanger or other orthogonal confirmation of reported variants. It also covers a curated variant classification workflow against ACMG criteria and a signed report. Systematic reviews of sequencing economics put clinical whole genome costs in a wide band across health systems, with per-case costs depending heavily on whether analysis and reporting labor is counted alongside reagents 1. A more recent review of the same literature finds the pattern persisting: reported costs vary by more than an order of magnitude, largely because studies draw the cost boundary in different places 2.

That accounting problem is why published “cost per genome” figures are so hard to compare. Sequencing consumables are often less than half of total cost once you include DNA extraction, library prep labor, instrument depreciation, compute, storage, clinical scientist time, and genetic counseling 3. If you are budgeting for yourself, budget for the pipeline and the interpretation rather than the flow cell.

What you are buying at each depth and read length

Two technical parameters, coverage depth and read length, determine most of what your genome can and cannot tell you. They are also the two most commonly misunderstood items on a price sheet, so it is worth being precise about what each one contributes.

Coverage depth is the parameter that most affects price and the one most often misrepresented. For germline single-nucleotide variant and small insertion or deletion calling, 30x is the standard operating point, and the marginal value of going to 40x or 50x is small for single-nucleotide variants. Depth matters more for structural variants, for mosaicism, and for calling in segmental duplications. If your interest is pharmacogenomics, polygenic scores, and rare coding variants, 30x short reads are sufficient.

Read length is the more interesting lever. Illumina 150 bp paired-end reads cannot map uniquely into roughly 5–10% of the genome, a fraction that includes medically relevant genes with high-identity paralogs: SMN1/SMN2, PMS2, CYP2D6, HBA1/HBA2, and the HLA region. Oxford Nanopore or PacBio HiFi long reads resolve those regions and call structural variants and repeat expansions that short reads miss entirely. Long-read whole genomes currently run $1,000 to $2,500 at research cores. If you want one genome you will never re-sequence, and you care about the hard regions, long reads at 20–25x are a better buy than short reads at 60x. The primer literature on population-scale sequencing lays out the same tradeoffs between platform, depth, and the variant classes each can and cannot detect 4.

Exome sequencing, which reads only the protein-coding portion of the genome, still exists at $300 to $600. We would skip it. It covers about 1–2% of the genome, misses non-coding regulatory variants and most structural variation, and has uneven capture efficiency that leaves gaps in GC-rich first exons. The price gap to a full genome no longer justifies the loss.

Where the money goes after the sequencer

Once the sequencing is finished, the remaining work splits into two parts. Turning reads into a variant list is computationally cheap and largely automated. Deciding what those variants mean is neither. This section covers both.

Assume you have your FASTQs. The compute is inexpensive and the work is tractable. A 30x human genome is roughly 100–120 GB as gzipped FASTQ and 50–80 GB as a BAM. It is 15–25 GB as a CRAM, a more compact alignment format that stores reads by reference to GRCh38 rather than in full. Alignment with bwa-mem2 mem -t 32 -R '@RG\tID:1\tSM:sample\tPL:ILLUMINA' runs in a few hours on a 32-core machine. Variant calling with GATK HaplotypeCaller in GVCF mode is slower, so most people now use DeepVariant (--model_type=WGS --num_shards=32), which is faster and more accurate on indels. Total cloud spend for a single genome, alignment through VCF, is $15 to $40 on spot instances.

Annotation is where the real interpretive work begins. ANNOVAR and VEP both take a VCF and attach gene consequence and population allele frequency from gnomAD. They also add ClinVar assertions and in silico pathogenicity predictions 5. Your 4–5 million variants will collapse to a few hundred that are rare and protein-altering, and to a handful with any published clinical assertion. Most of what remains is variants of uncertain significance, which means exactly that: nobody knows.

Two calibrations are worth carrying into that process. First, the human genome is full of variation under selection and full of variation under none, and distinguishing the two from sequence alone remains hard 6. Second, diagnostic yield depends enormously on prior probability. Whole genome sequencing in critically ill children with suspected genetic disease found a molecular diagnosis in a substantial fraction of cases, because those children were selected for high prior likelihood 7. A healthy adult sequencing themselves out of curiosity should expect a much lower yield of clinically actionable findings, and that asymmetry is what the cost-effectiveness literature models when it evaluates testing strategies 8.

One caution follows directly from all of this. If your VCF shows a variant in ClinVar classified as pathogenic in a gene like BRCA1, MLH1, TTN, or MYBPC3, do not act on it from a research pipeline. Consumer and core-lab pipelines are not clinically validated, false positives in repetitive and low-coverage regions are common, and confirmation requires an orthogonal assay in a CLIA laboratory ordered through a clinician. Take the result to a genetic counselor or medical geneticist.

What we would buy

The right purchase depends on why you want the data, and the two common reasons point in different directions.

If the goal is to have your own genome as data you can work with, buy a 30x short-read genome with a contractual guarantee of raw FASTQ delivery, and verify the delivered files before the vendor’s retention window closes. Run samtools stats and mosdepth on the BAM and check three numbers: mean deduplicated depth, percent of the genome at ≥20x, and the duplicate rate. A duplicate rate above 15% suggests over-amplified library prep from a low-input saliva sample. Expect $400 to $700 for this in 2026.

If the goal is a clinical answer about a specific symptom, family history, or affected child, do not buy a consumer genome. Ask a clinician to order diagnostic sequencing so that the result arrives as a validated, reportable finding with follow-up attached.

Questions people also ask

How long does a genome test take? A research core turns around FASTQs in two to six weeks from sample receipt, with most of that time spent waiting for a flow cell to fill. Consumer vendors quote four to eight weeks and often run longer. Clinical diagnostic genomes take three to eight weeks routine, and rapid inpatient protocols can return provisional results in two to seven days when the situation requires it.

Why is clinical WGS ten times the price of a research genome? Because most of the cost is labor and regulation: CLIA-certified handling, orthogonal confirmation, ACMG-based variant classification by trained scientists, medical director sign-off, and counseling. Reagent and instrument cost is a minority of the total in clinical settings 3.

Does insurance cover it? Sometimes, for a documented diagnostic indication with prior authorization, and usually not for elective sequencing in a healthy adult. Health-economic evaluations of sequencing are built around this distinction between diagnostic and screening use 8.

Is the “$100 genome” real? As a reagent and instrument-time figure on a fully loaded high-output flow cell, roughly yes. As a price you can pay for your own sequenced, aligned, annotated genome, no. It excludes extraction and library prep labor. It also excludes compute, storage, and everything downstream 2.

Should I pay extra for 60x or 100x coverage? Not for germline single-nucleotide variant and indel calling, where returns above 30x are minimal. Spend the same money on long reads instead if the hard regions of the genome matter to you.

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. K Schwarze, J Buchanan, JC Taylor, et al. Are whole Exome and whole Genome Sequencing Approaches Cost-Effective? A Systematic Review of the Literature. Value in Health, 2018. https://doi.org/10.1016/j.jval.2018.04.677 ↩

  2. Frederick McElwee, Sally L. Sansom, James Buchanan, et al. The cost and cost-effectiveness of whole-exome and whole-genome sequencing: a systematic literature review. European Journal of Human Genetics, 2026. https://doi.org/10.1038/s41431-026-02146-2 ↩ ↩2

  3. Kurt Christensen, Dmitry Dukhovny, Uwe Siebert, et al. Assessing the Costs and Cost-Effectiveness of Genomic Sequencing. Journal of Personalized Medicine, 2015. https://doi.org/10.3390/jpm5040470 ↩ ↩2

  4. Rachel L Goldfeder, Dennis P Wall, Muin J Khoury, et al. Human Genome Sequencing at the Population Scale: A Primer on High-Throughput DNA Sequencing and Analysis. American Journal of Epidemiology, 2017. https://doi.org/10.1093/aje/kww224 ↩

  5. Hui Yang, Kai Wang. Genomic variant annotation and prioritization with ANNOVAR and wANNOVAR. Nature Protocols, 2015. https://doi.org/10.1038/nprot.2015.105 ↩

  6. Wenqing Fu, Joshua M. Akey. Selection and Adaptation in the Human Genome. Annual Review of Genomics and Human Genetics, 2013. https://doi.org/10.1146/annurev-genom-091212-153509 ↩

  7. NIHR BioResource—Rare Disease, Courtney E. French, Next Generation Children Project, et al. Whole genome sequencing reveals that genetic conditions are frequent in intensively ill children. Intensive Care Medicine, 2019. https://doi.org/10.1007/s00134-019-05552-x ↩

  8. Katherine Payne, Sean P. Gavan, Stuart J. Wright, et al. Cost-effectiveness analyses of genetic and genomic diagnostic tests. Nature Reviews Genetics, 2018. https://doi.org/10.1038/nrg.2017.108 ↩ ↩2