Choosing WGS, WES, and Panels Based on the Question, Not Coverage
Why It Matters
The intuition that “more reads always mean a better test” is frequently incorrect when selecting a genetic test. Targeted panels, whole-exome sequencing (WES), and whole-genome sequencing (WGS) differ not only in scope but also in specimen requirements, variant detection methodologies, interpretation burden, and potential for reanalysis. The optimal choice is not necessarily the broadest assay, but rather the one that most reliably detects the variants anticipated by the clinical or research question.
Test nomenclature does not guarantee performance. Even among WES assays, capture designs and coverage profiles vary; similarly, among WGS assays, library preparations and pipelines differ in their capacity to resolve structural variants and repeat expansions. Therefore, one must consult actual validation documentation rather than relying on product names or average depth metrics.
Core Concepts
Panels concentrate sequencing resources on defined genes or regions. They can be efficient when gene–disease relationships are well-established and high depth or rapid turnaround is required. However, if new genes are discovered, existing test results cannot be reanalyzed within those regions, and the included genes and transcripts vary by panel.
WES primarily captures protein-coding exons and adjacent splice regions. Because many causes of Mendelian disorders lie in coding regions, WES is useful for broad exploration; however, capture is uneven, and some exons, GC-rich regions, and homologous regions may be poorly covered. The term “whole exome” does not imply complete observation of all exons.
WGS provides more uniform coverage of coding and non-coding coordinates and expands opportunities for CNV and SV analysis. However, repetitive sequences and highly homologous regions remain challenging even with short-read WGS, and clinical interpretation of non-coding variants is limited. It is essential to distinguish between the volume of data generated and the number of interpretable results.
Workflow from Question to Assay
First, characterize the phenotype and possible modes of inheritance, and identify known causative genes and expected variant mechanisms. The required technologies vary depending on whether SNVs and small indels are predominant, or if exon-level deletions, balanced rearrangements, repeat expansions, mitochondrial variants, or mosaicism are significant.
Next, consider specimen quality, the required turnaround time, and the detection allele fraction. Finally, determine whether raw data can be reanalyzed in the event of a negative result and whether there are plans to proceed to alternative assays. Assay selection and post-negative-result planning constitute an integrated workflow.
Interpreting Coverage Correctly
Mean depth merely summarizes the total read count and does not indicate whether key loci were sufficiently covered. It is essential to assess per-exon depth for disease-relevant exons, the proportion of callable bases, mapping quality, and allele balance. Even with high mean depth, if specific exons are poorly covered, a negative result in those regions does not constitute strong evidence of exclusion.
While benchmarks such as GIAB provide a basis for comparing accuracy against defined truth sets, they do not represent all difficult genomic regions or all clinical specimens. Laboratory-specific validation and continuous quality control are required.
Small Selection Example
In diseases where the causative gene and hotspot are well-characterized, a validated high-depth panel may be advantageous for rapidly detecting low-level mosaic variants. If the phenotype is non-specific and there are many candidate genes, WES or WGS can broaden the exploratory scope. If structural variants or repeat sequences are central, separate CNV/SV assays, repeat-primed PCR, or long-read sequencing may be required.
This example illustrates question–technology fit rather than a fixed hierarchy.
Common Misconceptions
- The nominal target and the actual callable region are not identical.
- A higher mean depth does not necessarily increase sensitivity for all variant classes.
- A negative whole-genome sequencing (WGS) result does not rule out genetic disease, and a negative whole-exome sequencing (WES) result does not completely guarantee the absence of coding variants.
- Updates to the analysis pipeline and reanalysis can alter results without new specimen collection, but they cannot recover blind spots in the original data.
Limitations
Because costs and technologies evolve rapidly, presenting fixed numbers as universal rules quickly becomes outdated. In actual selection, it is essential to verify current test menus, validation scope, result return policies, and patient consent. Broad testing may also increase incidental and secondary findings, along with the burden of data governance.
Reading in Context
Low allele fraction and assay LOD are linked to the concepts of Germline and somatic VAF, the blind spots of homologous regions are linked to Pseudogene mapping, and structural variant annotation is linked to the concept of AnnotSV.