Golden Helix · Clinical Genomics Guide

WES vs WGS
Choosing the Right Sequencing Strategy

A clinical comparison of whole exome and whole genome sequencing. Diagnostic yield, cost, variant detection, and the criteria that determine which is the right choice for your lab or your patient.

Diagnostic YieldCost Per DiagnosisWhen to Choose EachOperational Implications

Introduction

There is no universal right answer.
There is the right answer for this case.

Whole exome sequencing and whole genome sequencing are the two most powerful tools in clinical genomics. Both use next generation sequencing. Both can identify the genetic cause of rare disease. For years labs and clinicians have debated which one to order.

The honest answer is that it depends on the patient population, the clinical indication, the variants you need to detect, the infrastructure your lab has in place, and increasingly, the cost-per-diagnosis rather than the cost-per-test. This guide gives lab directors, bioinformaticians, and ordering clinicians the framework to make that decision, grounded in current evidence rather than vendor positioning.

~1–2%
Genome captured by WES (exome only)
~99%
Genome covered by WGS
85%
Known disease-causing variants found in coding regions
29%
WGS diagnostic yield in WES-negative rare disease families (NEJM 2023)
80–100x
Typical WES mean coverage
30–40x
Typical WGS mean coverage

Definitions

What Each Test Actually Is

Whole exome sequencing (WES)

WES sequences the protein-coding regions of all ~20,000 genes in the human genome, collectively the exome. About 60 million base pairs, roughly 1 to 2% of the total genome, but containing an estimated 85% of known disease-causing variants. A capture step uses synthetic probes that hybridize to exonic regions, pulling them out of the DNA library before the run. This enrichment focuses read depth on the clinically relevant regions, typically achieving 80 to 100x mean coverage on a standard run.

Whole genome sequencing (WGS)

WGS sequences all approximately 3.2 billion base pairs of the human genome: coding and non-coding, exonic and intronic, regulatory and repetitive. No capture step. The entire genome is sequenced directly. WGS typically achieves 30 to 40x mean coverage on a standard clinical run, with higher depths (100x+) used for somatic tumor analysis or mosaic variant detection.

Side by Side

Key Differences at a Glance

The headline differences, before we get into where each one belongs clinically.

DimensionWhole Exome SequencingWhole Genome Sequencing
Genomic coverage~1–2% (exons only)~99% (whole genome)
Base pairs sequenced~60 million~3.2 billion
Typical mean coverage80–100x30–40x
Raw data per sample~10–20 GB~100–200 GB
Typical clinical cost$1,000–$3,000$2,000–$5,000
Turnaround timeFaster (less data)Slower (more data)
SNV / Indel detectionStrong in exonsStrong genome-wide
CNV detectionLimitedStrong
Structural variantsLimitedStrong
Non-coding variantsNoYes
Repeat expansion detectionLimitedBetter (best with long-read)
Mitochondrial DNAPartialFull
Capture / GC biasYes (capture step introduces bias)None (no capture step)
Established clinical useExtensiveGrowing rapidly

The Evidence

Diagnostic Yield: What Studies Show

The central clinical question is not which technology is more comprehensive in theory, WGS obviously covers more of the genome. The question is whether that additional coverage translates into more diagnoses, and for which patients. The evidence has shifted decisively in recent years.

A landmark 2023 study in the New England Journal of Medicine sequenced the genomes of 822 families with suspected rare monogenic disease who had not received a diagnosis from prior testing. WGS achieved a molecular diagnosis in 29.3% of previously undiagnosed families: a substantial yield from a population already filtered through prior testing, including (in many cases) WES.

A 2015 study in PNAS (Belkadi et al.) comparing WES and WGS head-to-head on the same individuals found that WGS detected ~650 high-quality coding SNVs per sample that WES missed: roughly 3% of coding variants. These are variants in regions where WES probes do not capture efficiently, particularly low-complexity sequences, high-GC regions, and exon-intron boundaries.

A 2023 systematic review and meta-analysis in PLOS ONE found that WGS had a consistently higher diagnostic yield than WES for pediatric patients with suspected genetic disorders, diagnosing 15 to 35% of WES-negative families across multiple studies.

The practical implication: if a patient has had WES with no diagnosis, WGS should be considered the next step, not another panel and not a WES repeat. The incremental diagnostic yield is real and clinically meaningful.

Where WGS gains its diagnostic advantage

  • Non-coding regulatory variants. Promoters, enhancers, and splice regulatory regions that affect expression without altering the protein-coding sequence. WES misses these entirely.
  • Deep intronic variants causing aberrant splicing. Variants hundreds of base pairs from an exon that create cryptic splice sites. Invisible to WES; need WGS or RNA sequencing.
  • Structural variants and large CNVs. WES capture disrupts the read-pair architecture that SV callers depend on. WGS provides significantly better detection of inversions, translocations, and large insertions.
  • Incomplete exome capture. Every WES kit has regions of the exome it captures poorly: high-GC exons, first and last exons, regions with homology to other loci. Gaps vary by capture kit.
  • Repeat expansions. Huntington disease, fragile X, spinocerebellar ataxias. Poorly detected by short-read WES. WGS improves detection. Long-read WGS is the gold standard.

Decision Logic

When to Choose WES

Despite WGS's diagnostic advantages, WES remains the appropriate choice in many clinical scenarios. The decision is not always "WGS if you can afford it."

  • 01

    Clinical indication maps cleanly to coding variants

    For well-characterized Mendelian conditions where the mutational spectrum is dominated by coding SNVs and indels (BRCA1/2 hereditary cancer, CFTR-related cystic fibrosis), WES provides sufficient coverage at lower cost and faster turnaround.

  • 02

    Payer coverage supports WES but not WGS

    WGS reimbursement remains inconsistent. In many systems WES is the covered option for rare disease diagnosis. Ordering WGS when payer coverage is absent shifts cost to the patient, a clinical and ethical consideration.

  • 03

    Lab infrastructure is not validated for WGS-scale data

    WGS generates 10x more raw data than WES. If storage, compute, and pipeline are not validated for WGS scale, WES is the pragmatic choice until infrastructure is ready. Running WGS through a WES-validated pipeline is a validation failure waiting to happen.

  • 04

    Turnaround time is critical and rapid WGS is not available

    Rapid WGS for critically ill neonates is increasingly available but not at every institution. Where WGS TAT cannot meet clinical needs, WES may be the right interim approach.

  • 05

    Trio analysis on a constrained budget

    Sequencing proband plus both parents dramatically improves diagnostic yield over proband-only WES by enabling de novo variant detection. A trio WES at $6,000–$9,000 total may outperform proband-only WGS at $3,000–$5,000 for many indications.

  • 01

    Capture bias

    The hybridization capture step introduces GC-content bias. High-GC exons, including many disease-relevant promoter-proximal exons, are systematically underrepresented. This creates consistent coverage gaps that vary by kit vendor.

  • 02

    Missing non-coding disease

    An estimated 15 to 25% of Mendelian disease diagnoses in WES-negative patients are eventually explained by non-coding variants, most of which WGS would detect.

  • 03

    SV blindspot

    WES has poor sensitivity for copy number variants below ~50 kb and essentially no sensitivity for balanced structural variants. Conditions with a significant SV mutational mechanism are at risk of false negatives.

  • 04

    Inter-kit variability

    Different WES capture kits cover different exonic regions. A negative WES result on one platform may not mean the variant is undetectable. It may mean the variant falls in a gap specific to that kit.

Decision Logic

When to Choose WGS

WGS is no longer a research-only technology. It is entering clinical production at labs globally and is increasingly supported by professional society guidelines as a first- or second-line test for specific indications.

  • 01

    Patient is WES-negative with strong clinical suspicion

    The highest-yield application of WGS. A 2023 NEJM study demonstrated 29% diagnostic yield in this population. If a patient has had WES, gene panels, and chromosomal microarray without a diagnosis, WGS is the rational next step.

  • 02

    Critically ill neonate requiring rapid diagnosis

    Rapid WGS programs returning results in 24 to 72 hours have demonstrated clinical utility in NICUs and PICUs, where a molecular diagnosis can directly change management within a time-sensitive window. RCTs show rapid WGS changes management in 20 to 40% of cases.

  • 03

    Structural variant etiology is suspected

    For conditions where large deletions, duplications, or rearrangements are a primary mechanism (many neurodevelopmental disorders), WGS's superior SV detection makes it preferred over WES plus chromosomal microarray.

  • 04

    A comprehensive first-line approach is feasible

    When WGS cost and infrastructure are not limiting, WGS-first eliminates stepwise testing and the delays it introduces. The cost-per-diagnosis argument increasingly supports this approach for undiagnosed rare disease programs.

  • 05

    Mitochondrial disease is on the differential

    WGS provides full mitochondrial genome coverage at high depth (mtDNA is present in hundreds of copies per cell). WES mtDNA coverage is partial and inconsistent.

  • 06

    Non-coding or regulatory variants are clinically suspected

    For genes where non-coding pathogenic variants are documented (LDLR promoter variants in familial hypercholesterolemia, HBB regulatory variants in thalassemia, deep intronic variants in CFTR), WGS provides the only NGS-based route to detection.

Economics

The Cost-Per-Diagnosis Argument

The common framing of WES vs WGS as a cost comparison ("WES is cheaper, therefore WES first") is increasingly outdated. The relevant metric is not cost-per-test. It is cost-per-diagnosis.

A 2022 health economic analysis in Genetics in Medicine (Lavelle et al.) modeled cost-effectiveness of WES vs WGS for children with rare undiagnosed conditions. While WGS has a higher upfront cost, its higher diagnostic yield, particularly in WES-negative patients, results in a lower cost-per-diagnosis when accounting for the avoided cost of subsequent testing, repeat consultations, and empirical treatment of undiagnosed patients.

The stepwise approach (panel first, then WES, then WGS) was the most expensive path per diagnosis in the model, despite having the lowest cost at each individual step.

Operations

Operational Implications for the Lab

The WES vs WGS decision is not just a clinical genetics question. It is an infrastructure and operations question that directly affects the bioinformatics pipeline, storage architecture, and interpretation workflow.

Bioinformatics pipeline

WGS requires a pipeline validated for whole-genome scale. Not simply a matter of pointing a WES pipeline at a larger input file. Specific differences:

  • Variant callers must be configured for genome-wide operation, not just exon-targeted regions. A caller tuned for high-depth WES may perform differently at 30x WGS depth.
  • SV and CNV callers must be included. A WGS pipeline that only calls SNVs and indels is discarding a significant fraction of WGS's diagnostic value.
  • Non-coding annotation needs additional sources and filtering strategies. Without appropriate non-coding databases and effect predictors, WGS generates enormous noise in non-coding regions.
  • Compute and storage requirements are 5 to 10x higher. WGS needs significantly more RAM, CPU, and storage at every stage.

Interpretation workflow

A standard WES tertiary analysis might produce 200 to 500 candidate variants after filtering. WGS on the same patient with non-coding regions included can produce several thousand candidates without appropriate filtering strategies. Labs transitioning to WGS must invest in non-coding variant filtering frameworks, phenotype-driven prioritization, and expanded ACMG/AMP evidence gathering for non-coding variation. VarSeq handles both WES and WGS tertiary analysis within a single unified platform.

Storage at scale

MetricWESWGS
Raw FASTQ per sample~10 GB~100–200 GB
BAM file per sample~8–12 GB~80–120 GB
VCF per sample~100 MB~500 MB – 1 GB
Storage per 100 samples/week~2–3 TB/year~20–30 TB/year

A lab running 100 WGS cases per week needs 1 to 1.5 petabytes of raw data per year if retaining FASTQs. Most institutions adopt tiered retention: VCFs and clinical reports indefinitely, BAMs for 2 to 5 years, FASTQs compressed or discarded after a defined validation period.

Common Questions

Frequently Asked Questions

Is WGS better than WES?
For most clinical indications where a definitive genetic diagnosis is the goal, WGS has a higher diagnostic yield, particularly in patients who are WES-negative. A 2023 NEJM study demonstrated 29% diagnostic yield in WES-negative rare disease families using WGS. However, "better" depends on the clinical question, infrastructure, and cost context. WES remains appropriate where payer coverage supports it, where the indication maps to coding variants, and where lab infrastructure is not yet validated for WGS-scale analysis. The trend in clinical genomics is toward WGS as a first-line test, but WES remains the practical choice for many programs today.
Is WES the same as WGS?
No. Both use next generation sequencing technology but differ fundamentally in scope. WES sequences only the protein-coding regions of the genome, approximately 1 to 2% of the total, using a capture step to enrich those regions before sequencing. WGS sequences the entire genome without a capture step. WGS detects all the variants WES detects in coding regions, plus non-coding variants, structural variants, repeat expansions, and variants in poorly captured exonic regions that WES misses.
Can WES be rejected by insurance?
Yes. Insurance coverage for WES and WGS varies significantly by payer, indication, and jurisdiction. In the United States, most major commercial payers and Medicare cover WES for specific indications, typically including patients with undiagnosed rare disease, developmental delay, intellectual disability, or unexplained epilepsy when prior genetic testing has been non-diagnostic. WGS coverage is less consistent and more frequently requires prior authorization with documented medical necessity. Coverage criteria change regularly. Ordering clinicians and laboratory staff should verify current payer policies before ordering. Many labs have genetic counselors or reimbursement specialists who assist with prior authorization.
What is the diagnostic yield of WES vs WGS?
For unselected rare disease populations: WES yields 25 to 50%, WGS yields 35 to 55% in large clinical cohort studies. The yield difference grows in WES-negative populations: WGS achieves ~29% additional diagnoses where WES failed (NEJM 2023). The additional diagnoses cluster in non-coding regulatory variants, deep intronic splice variants, structural variants, repeat expansions, and variants in poorly captured exonic regions. For specific indications with well-characterized coding mutational spectra (BRCA, CFTR), the yield gap narrows.
Should I use hg19 or hg38?
For new clinical pipelines, use GRCh38/hg38. It is the current standard, corrects errors in hg19, and is the primary build for ClinVar, gnomAD, and ClinGen. Labs still on hg19 should plan a migration. Never mix builds in a pipeline: variants called on hg19 must be annotated with hg19 coordinates. See tertiary analysis for more on annotation consistency.
How much more storage does WGS need than WES?
About 10x more, end to end. WES produces ~10 GB FASTQ per sample; WGS produces 100 to 200 GB. BAM files are ~10 GB for WES and 80 to 120 GB for WGS. A lab running 100 cases per week needs ~2 to 3 TB per year for WES vs ~20 to 30 TB per year for WGS. Most institutions adopt a tiered approach: retain VCFs and clinical reports indefinitely, retain BAMs for 2 to 5 years, archive or discard FASTQs after a defined validation period.

Keep Reading

Related Resources

Go deeper on either test, or on how a VCF becomes a clinical report regardless of which test produced it.

Whole Exome Sequencing (WES): A Clinical GuideWhat WES is, how it works, when to order it.Whole Genome Sequencing (WGS): A Clinical GuideThe most comprehensive genetic test, in clinical context.Next Generation Sequencing: A Complete GuideThe full pipeline view.Tertiary Analysis in NGSHow either VCF becomes a clinical report.Rare Disease GenomicsFrom diagnostic odyssey to molecular diagnosis.Rare Disease Solutions at Golden HelixHow VarSeq supports both WES and WGS interpretation.

Handle Both WES and WGS in One Platform

VarSeq supports tertiary analysis for targeted panels, whole exome, and whole genome sequencing within a unified workflow, with variant filtering, ACMG-guided interpretation, and clinical reporting that scales from gene panels to whole genomes without requiring separate pipelines.

Explore VarSeq Request an Evaluation