What an exome test costs, and what you get for the money

A research-grade whole exome with raw data returned runs roughly $250 to $600 per sample in 2026. A clinical, CLIA/CAP-certified exome with an interpreted report and variant classification is typically $1,000 to $2,500 self-pay for a singleton, and $2,500 to $5,000 for a trio (proband plus both parents). Genetic counseling, if bought separately, adds $200 to $500 on each side of the test. The published health-economics literature reports a much wider band than any price list, because different studies count different things: a systematic review of 36 studies found reported per-sample costs for exome sequencing ranging from about US$555 to US$5,169, with genome sequencing spanning roughly US$1,906 to US$24,810, and most of that variation came from what each study included in “cost” rather than from the assay itself.¹

The important split is whether you are buying molecules or buying judgment. The wet lab is cheap and getting cheaper. Interpretation is human labor and has not.

Line-by-line: where the money goes

For a 36 Mb capture target at 100x mean coverage, the consumables at volume look approximately like this:

• Capture probes (Twist Exome 2.0, IDT xGen Exome Research Panel v2, Agilent SureSelect V8): $40 to $120 per sample when pooling 8 to 16 libraries per hybridization reaction. At single-sample volumes, three to five times that.
• Library prep enzymes, adapters, unique dual indexes: $20 to $60.
• Sequencing. You need about 6 to 10 Gb of raw 2x150 data per exome, because 20% to 40% of reads land off-target and another 5% to 15% are optical or PCR duplicates. On a NovaSeq X 25B flow cell run full, per-Gb cost lands under $2, so the sequencing itself is $15 to $40. On an older NovaSeq 6000 S4, closer to $5 per Gb, so $40 to $70.
• DNA extraction and QC (Qubit, TapeStation, sometimes a fragment analyzer trace per library): $15 to $40.

That is $100 to $250 of reagents. Everything above that price is instrument depreciation, labor, accessioning, sample tracking, LIMS, storage, proficiency testing, accreditation overhead, and margin. For a clinical lab, the largest single non-reagent cost is a board-certified molecular geneticist reading the case: one to four hours of curation per singleton against ACMG/AMP criteria, longer if the phenotype is vague or the variant is novel. Trios cost more in sequencing (three libraries) but the marginal interpretation cost is often lower per answer, because de novo filtering collapses the candidate list from thousands to dozens.

The price bands, and who each one is for

$250 to $600, raw data only. Usually a research-use-only service: you send saliva or blood or extracted DNA, you get FASTQ and often a BAM/CRAM and a VCF. No report, no clinical validity, no counseling. This is the right purchase if you intend to do your own analysis and you understand that a result from a non-CLIA lab will not be accepted for clinical decisions without confirmatory testing.

$1,000 to $2,500, clinical singleton. CLIA/CAP lab, Sanger or orthogonal confirmation of reported variants, a written report with ACMG classifications, and usually an option for secondary findings from the ACMG recommended gene list. Ask explicitly whether raw data is included and in what format. Many clinical labs charge $100 to $500 extra for a BAM, and some will only release a VCF.

$2,500 to $5,000, trio or quad. The standard for a suspected rare Mendelian condition in a child. The parental samples are what make de novo and compound-heterozygous calls tractable.

When you compare quotes, normalize on four numbers: capture kit and target size, mean on-target depth, percent of target at ≥20x, and which files you receive. A lab quoting $400 at 50x mean is not selling the same thing as a lab quoting $900 at 100x with ≥98% of target at 20x.

Why the exome is cheaper, and what that discount costs you technically

The exome is about 1.5% of the genome, so you sequence far fewer bases for the same depth. The tradeoff is capture bias. Hybrid capture produces uneven coverage: GC-rich first exons drop out, and a persistent 1% to 3% of coding bases sit below 10x in any given sample, varying by kit. Genome sequencing gives more uniform coverage and detects structural and copy-number variation far better, which is the main technical argument for paying more.² Exome data also misses repeat expansions, most deep intronic and promoter variants, and mitochondrial variants unless the kit includes chrM probes or you scavenge off-target reads.

Pseudogene regions are a specific failure mode worth knowing: short reads in GBA/GBAP1, SMN1/SMN2, PMS2/PMS2CL, and CYP21A2 mismap in ways that silently produce wrong calls. If a gene like that matters to you, panel-based or long-read follow-up is the answer, not a deeper exome.

What lands in your bucket, and the first commands to run

Expect 5 to 8 GB of gzipped FASTQ per sample, a 4 to 8 GB BAM (or a 1.5 to 3 GB CRAM), and a VCF on the order of 20 to 60 MB containing roughly 20,000 to 25,000 variants inside the capture target.

A pipeline we would run on a returned FASTQ:

bwa-mem2 mem -t 32 -R '@RG\tID:L1\tSM:S1\tPL:ILLUMINA' GRCh38.fa R1.fq.gz R2.fq.gz \
  | samtools sort -@ 8 -o S1.bam -
gatk MarkDuplicates -I S1.bam -O S1.md.bam -M S1.dup.txt
mosdepth --by exome_targets.bed --thresholds 1,10,20,30 S1 S1.md.bam
run_deepvariant --model_type=WES --ref=GRCh38.fa --reads=S1.md.bam \
  --regions=exome_targets.bed --output_vcf=S1.vcf.gz --num_shards=32
bcftools norm -m -any -f GRCh38.fa S1.vcf.gz -Oz -o S1.norm.vcf.gz
vep -i S1.norm.vcf.gz --cache --assembly GRCh38 --everything --vcf -o S1.vep.vcf

Read the mosdepth thresholds file before anything else. The fraction of target at ≥20x is your single best QC number, and it tells you which genes you cannot say anything about. Then check bcftools stats for Ti/Tv (expect about 3.0 to 3.3 across a capture target; a value near 2.1 means you are picking up noise) and the het/hom ratio, which will shift with ancestry.

For filtering, gnomAD v4 allele frequency plus ClinVar is the workhorse combination. A reasonable first pass: keep variants with a coding or splice consequence, popmax AF below 0.001 for dominant hypotheses and below 0.01 for recessive, then sort by ClinVar status and constraint metrics. Copy-number calling from exome data needs a batch of samples processed identically (ExomeDepth, CNVkit, GATK gCNV); a single exome in isolation gives you very little CNV signal.

Where you need a clinician, plainly

A variant you find in your own VCF is a hypothesis about a data file, not a finding about your health. Classification depends on phenotype, family history, segregation, and functional evidence that pipelines do not have. Exome sequencing also routinely turns up secondary findings unrelated to the reason for testing, and the professional guidance has consistently been that these should be handled inside a relationship with a clinician who can arrange pre-test and post-test counseling.³ If anything in your data looks medically consequential, take it to a clinical geneticist or genetic counselor and expect the finding to be reconfirmed in a CLIA lab before anyone acts on it.

Questions people also ask

How long does exome sequencing take? Wet lab is 3 to 7 days: extraction, library prep, an overnight hybridization, and 24 to 44 hours of NovaSeq runtime depending on flow cell and read length. Research services usually quote 3 to 6 weeks total, mostly batching and queueing. Clinical labs quote 3 to 8 weeks for standard turnaround and 5 to 14 days for rapid or critical-care exomes, where the added cost is a dedicated flow cell and an analyst working the case immediately.

What is exome testing used for? Finding coding-region variants that explain a suspected single-gene condition, mostly in rare disease diagnosis, plus carrier status and pharmacogenomic star alleles as secondary uses. It is a broad first-line test when the phenotype does not point at one gene.

Is a genome worth the extra money over an exome? Cost-effectiveness comparisons between the two remain unsettled, largely because downstream diagnostic and management costs are rarely measured alongside the assay.¹ Technically, genomes give better coverage uniformity and structural variant detection.² If you plan to reanalyze your own data over years, the genome is the better substrate because reinterpretation improves without resequencing.

Why do quoted prices differ by 10x between labs? Accreditation, interpretation labor, confirmation testing, and whether the quote includes counseling and reanalysis. The reagents are a small and shrinking fraction of the total.

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. Katharina Schwarze, James Buchanan, Jenny C. Taylor, et al. Are whole-exome and whole-genome sequencing approaches cost-effective? A systematic review of the literature. Genetics in Medicine, 2018. https://doi.org/10.1038/gim.2017.247 ↩ ↩²

2. Britt‐Sabina Petersen, Broder Fredrich, Marc P. Hoeppner, et al. Opportunities and challenges of whole-genome and -exome sequencing. BMC Genetics, 2017. https://doi.org/10.1186/s12863-017-0479-5 ↩ ↩²

3. Wayne W. Grody, Barry H. Thompson, Louanne Hudgins. Whole-Exome/Genome Sequencing and Genomics. PEDIATRICS, 2013. https://doi.org/10.1542/peds.2013-1032e ↩