Gene panels vs exome sequencing for pediatric immunodeficiency

The Technical Divide: Coverage Depth versus Genomic Breadth

The choice between a targeted gene panel and whole exome sequencing determines which variants are easiest to detect, how quickly a laboratory can report them, and how much uncertainty remains after the first analysis.

In primary immunodeficiency and the broader group of inborn errors of immunity, the interval between symptom onset and molecular diagnosis can extend from two to five years across tertiary referral settings. That delay has many causes: incomplete phenotyping, overlapping clinical presentations, changing disease classifications, and limited access to specialized testing. Sequencing strategy is only one part of the problem, but it is an important one.

Targeted panels and whole exome sequencing are not simply two versions of the same assay. They prioritize different kinds of information.

Targeted panels use PCR-based or hybridization-capture methods to enrich a curated set of disease-associated genes. Immunodeficiency panels commonly include several hundred loci, often in the range of 200 to 400 genes, although the exact content varies between laboratories and changes over time. Because the assay is restricted to a defined target, sequencing reads can be concentrated at much greater depth. In the draft comparison considered here, panel coverage can reach approximately 5,000x.

That depth is clinically useful when the suspected variant is present in only a small proportion of cells. It can improve the detection of low-level mosaicism and help compensate for difficult regions that are unevenly represented by standard capture methods. It does not, however, make every region technically perfect. Some genes still require special laboratory methods, and a panel cannot detect a pathogenic change in a gene that is not included in its design.

Whole exome sequencing captures the protein-coding exons of approximately 20,000 genes. These regions account for only about 1% to 2% of the human genome, but they contain a large proportion of known disease-causing variants. Standard WES generally operates at a lower average depth, around 100x in the comparison used here. Its strength is not maximal read depth; it is the breadth of the search.

That breadth matters when a child’s phenotype does not map neatly onto a classic immunodeficiency syndrome. WES can examine genes not yet associated with immunodeficiency, genes linked to more than one clinical phenotype, and genes that have been added to disease classifications since the laboratory’s panel was designed. It can also preserve the possibility of reanalysis when the medical literature and gene–disease knowledge change.

The trade-off is clear: panels exchange breadth for depth, while WES exchanges some depth for a much wider search space. Neither platform is universally superior. The right first-line test depends on the clinical question, the quality of the phenotype, the laboratory’s validation data, and what will happen if the first assay is negative.

Diagnostic Yields in Clinical Practice: Analyzing Large-Scale Cohorts

Diagnostic yield is often presented as if it were a stable property of a sequencing platform. In practice, it is a property of the entire diagnostic system: the patient population, referral criteria, prior testing, family structure, phenotyping, variant interpretation, and the assay itself.

A longitudinal study of 878 patients with primary immunodeficiency evaluated a 264-gene targeted panel as first-line testing. The reported yield was 56% among the 780 patients with complete data, corresponding to 433 diagnoses. When panel-negative cases underwent reflex WES, the additional diagnoses increased the overall yield to 58%. In that cohort, WES added a relatively small number of diagnoses after the panel had already been applied.

A 10-year tertiary-care study from Turkey evaluated 1,407 patients with suspected inborn errors of immunity. The reported diagnostic yields were 66.8% for targeted panel testing and 68.4% for WES. The difference was 1.6 percentage points in that particular clinical population. It is useful evidence about the performance of the two workflows in that setting, but it should not be treated as a universal rule that the platforms are interchangeable.

The distinction matters because a near-identical yield in one cohort does not prove that the same children would have received the same diagnoses on each platform. Two tests can produce similar aggregate percentages while identifying partly different patients. A panel may be especially effective for deep coverage of a well-established group of genes, whereas WES may find a diagnosis in a child whose phenotype points beyond the panel’s boundaries. The total number of diagnoses alone does not show how much overlap exists between the results.

The Mexican pediatric cohort illustrates a second limitation. Among 157 children with suspected immunodeficiency, the overall yield from next-generation sequencing was 32.48%. That figure reflects the characteristics of the cohort, including patient selection, the phenotypic spectrum, and the genetic structure of the population. It cannot be used as a direct head-to-head comparison of targeted panels and WES unless the study design actually compares those platforms under comparable conditions.

The same caution applies to the reported 10.42% reclassification rate in that cohort. It indicates that a proportion of diagnosed children were assigned to a different IUIS category after molecular findings were considered than had been suspected from the clinical phenotype alone. It does not show that targeted panels and WES had equivalent diagnostic capture, and it does not establish that WES alone caused the reclassification.

Diagnostic yield is shaped by the cohort as much as by the platform. Similar percentages do not mean that the same variants, genes, or patients were detected by the same route.

Across the available cohort results in the draft comparison, the practical message is more restrained than a simple platform ranking. In well-characterized groups with classical presentations, targeted panels can produce a high yield and may approach WES in the number of diagnoses returned. WES becomes more valuable when the phenotype is broad, atypical, syndromic, or difficult to reconcile with the genes already on a panel.

The decisive difference may therefore be found in the cases that sit outside the center of the phenotype distribution. Aggregate yield tells the laboratory how many answers it found. It does not fully explain why some cases were solved by one method and not the other, which variant classes remained invisible, or how many negative cases could become solvable after reanalysis.

When High-Depth Panels Outperform Whole Exome Sequencing

WES is broader, but broader does not automatically mean more sensitive for every variant or every clinical question. Several situations favor a high-depth panel, particularly when the laboratory has a well-designed assay and the suspected disease mechanism is already reasonably focused.

Low-level mosaicism

Average WES depth around 100x may be insufficient for reliable detection of variants present at a low allelic fraction. In practical terms, a mosaic variant can be diluted by a large number of reads carrying the reference allele, especially when the variant is present only in a subset of blood cells. Targeted panels can concentrate thousands of reads over the relevant locus. In the comparison used here, depths of up to 5,000x create a substantially better opportunity to detect variants present at approximately 1% to 3% allele frequency, although the actual limit of detection depends on sample type, base quality, the variant caller, validation, and confirmation testing.

This distinction is relevant to disorders in which mosaicism or somatic reversion is biologically plausible, including some presentations involving ADA-SCID or STAT1 gain-of-function. It is also a reminder that the blood sample may not represent every tissue in which a variant is present. High sequencing depth improves analytical sensitivity; it does not remove the need to choose the right specimen or interpret the result in clinical context.

Difficult-to-sequence regions

Some immunologically relevant genes contain repetitive elements, pseudogene-related complexity, high GC content, or other sequence features that interfere with uniform capture and read alignment. The IKBKG region, for example, can present technical challenges. HLA-adjacent regions and other structurally complex loci may also be unevenly covered by standard exome kits.

A targeted assay can be designed around known problem regions. The laboratory may adjust probe density, optimize capture conditions, or supplement sequencing with another method. That does not mean every targeted panel covers every difficult locus better than every exome assay. The meaningful question is whether the specific laboratory has validated performance for the gene and variant class relevant to the child.

A negative result is therefore only as strong as the coverage and variant types the test was designed to detect. A report that lists a gene as technically included is not the same as proof that all exons, splice boundaries, copy-number changes, and challenging regions were adequately assessed.

A strongly suspected single-gene disorder

When the phenotype points toward a relatively narrow diagnostic question, a focused panel may be faster and more efficient than exome-wide analysis. X-linked agammaglobulinemia, for example, raises a focused question about BTK, while chronic granulomatous disease may lead the laboratory to examine CYBB, CYBA, NCF1, NCF2, and NCF4 alongside the functional oxidative burst results.

In these cases, the value of a panel is not merely that it tests fewer genes. It can provide deeper coverage across the genes that matter, simplify interpretation, and reduce the number of incidental or uncertain findings. If the clinical suspicion is already concentrated, screening the entire exome may add analytical work without adding an equivalent amount of useful information.

That conclusion depends on the phenotype being genuinely informative. A child with recurrent bacterial infections and neutropenia may initially appear to have a narrow problem, but syndromic features, unusual organisms, severe viral disease, autoimmunity, growth abnormalities, or a family history of early death can broaden the differential substantially. The panel should follow the clinical question, not replace it.

Time-sensitive diagnostic work

Targeted panels often have a shorter reported turnaround than WES. In the comparison used here, panels return results in approximately two to four weeks, while WES takes approximately four to eight weeks. The difference may matter in a rapidly evolving pediatric case, particularly when the result could change the next diagnostic or therapeutic branch.

SCID requires especially careful wording. A molecular result can clarify the genetic cause, support family counseling, refine transplant planning, and identify additional testing needs, but urgent protective management should not be postponed until sequencing confirms the diagnosis. When newborn screening, lymphocyte findings, infection history, or other clinical evidence raises concern for SCID, clinicians may begin protective isolation, infection-prevention measures, antimicrobial prophylaxis, immunology consultation, and transplant evaluation while molecular testing is still in progress, according to the clinical situation and local protocols.

A faster panel can still be valuable because it may shorten the time to etiologic clarification. It should not be described as the point at which urgent SCID management begins. The practical benefit is faster diagnostic resolution alongside immediate action based on screening and clinical suspicion.

High depth can answer a narrow question with unusual sensitivity. It cannot compensate for a gene that was never included or for a clinical suspicion that was framed too narrowly.

The distinction is analytical sensitivity versus diagnostic scope. Panels are optimized for sensitivity within a defined target. WES is optimized for a broader search across coding genes. In a narrowly framed case, depth may be the dominant advantage. In an unresolved case with an incomplete or atypical phenotype, breadth may be more important than maximum coverage of a familiar gene list.

Economic and Workflow Implications of First-Line Sequencing Strategies

The first-line decision also changes the laboratory workflow. A panel-first strategy followed by WES for negative cases is not simply two tests arranged in sequence; it is a diagnostic pathway with its own costs, delays, reporting rules, and opportunities for information loss.

Cost modeling from a multicenter study of 878 patients with immunodeficiency reported commercial savings of approximately $300 to $950 per patient for a WES-first workflow compared with a tiered panel-then-WES strategy. The reason is straightforward. In a tiered model, patients whose panel does not provide a diagnosis may undergo a second sequencing assay. If the reflex-testing group is large enough, running WES once for everyone can cost less than running a panel for everyone and then adding WES for a substantial subset.

That result does not mean that WES-first is automatically cheaper for every laboratory. Economic comparisons depend on local conditions, and a published model cannot substitute for an institution’s own numbers.

Several variables shape the balance:

1. Reagent and platform costs. The price difference between a targeted panel and an exome capture kit depends on the vendor, instrument, batch size, service contracts, purchasing agreements, and regional distribution. The same nominal test can have a different cost structure in different laboratories.

2. Bioinformatic infrastructure. WES produces a larger and more complex dataset. Storage, compute capacity, pipeline maintenance, quality review, variant curation, and specialist interpretation all contribute to the real cost. A laboratory that already performs exome analysis may absorb these costs more efficiently than one building the workflow from scratch.

3. Panel yield and reflex rate. If a validated panel solves a large proportion of appropriately selected cases, fewer children proceed to WES. The economic advantage of a WES-first pathway then becomes less obvious, especially if the panel is substantially cheaper and has a meaningful turnaround advantage.

4. Clinical urgency. A shorter panel turnaround can have value even when the per-test price is higher. That value is not limited to a formal cost calculation. Earlier etiologic information may help clinicians organize family testing, choose confirmatory assays, interpret functional studies, and plan specialist care.

5. Data ownership and future analysis. WES creates a stored dataset that can be revisited, but reanalysis is not free. It requires a reason to reanalyze, suitable pipeline support, updated knowledge resources, and staff time. A theoretical future benefit becomes a practical benefit only if the laboratory has a process for using it.

A tiered pathway can also be clinically coherent. A child with a classic phenotype and a suspected gene group may benefit from a rapid, high-depth panel. A negative result can then trigger WES, genome sequencing, copy-number analysis, repeat testing in another tissue, or a focused laboratory method depending on the suspected mechanism. This is different from treating a panel as a definitive exclusion test.

Conversely, WES-first may be sensible when the phenotype is heterogeneous, the family structure supports trio analysis, or the probability of needing reflex testing is high. It may also reduce the risk that a panel becomes outdated while the child moves through several rounds of testing. Yet WES does not remove the need for targeted confirmation, deletion and duplication analysis, repeat-expansion testing, functional assays, or other methods that are outside routine exome capabilities.

No single economic model applies across all diagnostic laboratories. The meaningful comparison is local: diagnostic yield in the actual referral population, proportion of cases needing reflex testing, time to a clinically useful result, interpretation capacity, reimbursement, and the consequences of a delayed or incomplete answer.

Navigating Reclassification: The Impact of Genomic Data on Immunological Diagnosis

A molecular diagnosis can do more than attach a gene name to an existing clinical label. It may change the immunological category assigned to the child, alter the expected disease course, and redirect surveillance or family counseling.

The Mexican pediatric cohort provides an important example. Among children with suspected immunodeficiency who received a diagnosis through the study’s overall next-generation sequencing approach, 10.42% were reclassified into a different IUIS category than the category initially suspected from the clinical phenotype alone. This finding shows that molecular testing can expose a mismatch between the initial clinical framework and the child’s underlying biology.

It does not, by itself, establish that WES produces this reclassification more often than a targeted panel. Nor does it demonstrate that panels and WES capture the same diagnoses. The reported percentage belongs to the cohort’s overall next-generation sequencing results and should be interpreted within that study design.

The clinical consequences of reclassification can nevertheless be substantial. A child initially considered to have an antibody-predominant disorder may be found to have a genetic condition associated with combined immunodeficiency, immune dysregulation, autoinflammation, or a broader syndromic phenotype. That information may affect infection surveillance, assessment for organ involvement, the choice and timing of immunomodulatory treatment, transplant discussions, and testing of relatives.

Atypical genotype–phenotype relationships

The relationship between gene and phenotype in inborn errors of immunity is not fixed. The same gene can be associated with different clinical presentations, and similar presentations can result from variants in different genes. Some children do not show the full textbook pattern at the time of testing, particularly early in life.

WES can be useful in this setting because it does not require the laboratory to commit in advance to a short list of gene–phenotype pairings. It may reveal a variant in a gene that was not considered clinically, a gene whose recognized phenotype has expanded, or a candidate finding that requires functional and family-based follow-up.

A panel can also accommodate phenotypic complexity if it is sufficiently broad and regularly updated. The limitation is structural: the search remains bounded by the genes and variant classes included in the assay. A panel designed around classical presentations may be less useful when the child’s features fall between established categories.

Reanalysis as a longitudinal resource

One of WES’s durable advantages is the possibility of revisiting the same sequence data as knowledge changes. A previously negative exome may become informative when a new immunodeficiency gene is described, a variant is reclassified, or the child develops additional features that sharpen the phenotype.

That potential should not be described as unlimited. Reanalysis depends on the quality of the original data, whether the relevant gene and variant type were adequately covered, the retention of raw and processed files, the laboratory’s policies, and the availability of updated interpretation resources. Some diagnoses will require new sequencing because the original assay did not capture the relevant region or structural variant.

Panel data can also be reinterpreted within the boundaries of the original assay. If the panel includes a newly implicated gene or if a previously uncertain variant becomes pathogenic, stored panel data may yield a new answer without repeat sequencing. If the relevant gene was absent from the panel, however, the laboratory must redesign the assay or move to a broader test.

The longitudinal value of WES is therefore real, but it is not a substitute for careful first-pass analysis. A broad dataset that is poorly interpreted, inadequately phenotyped, or not retained for future review does not deliver its full promise.

Making the First-Line Choice Without Pretending the Platforms Are Equivalent

The phrase targeted gene panel vs whole exome sequencing pediatric describes a genuine clinical decision, not a contest with a single winner. A panel is usually strongest when the question is focused, the relevant genes are well established, and rapid, high-depth analysis is important. WES is usually strongest when the phenotype is heterogeneous, the suspected mechanism is uncertain, or the probability of needing to look beyond a conventional immunodeficiency list is high.

The choice can be organized around the diagnostic question:

• A recognizable phenotype with a short differential may justify a high-depth panel, especially when the laboratory has demonstrated strong coverage of the relevant genes and variant classes.
• A severe or unusual phenotype without a clear syndrome match may favor WES, particularly when the analysis can include parental samples and be revisited later.
• A concern about mosaicism makes depth, specimen choice, and validated low-level variant detection central to the decision.
• A negative first-line result should prompt review of coverage, variant type, sample adequacy, phenotype quality, and the need for functional or orthogonal testing rather than an automatic assumption that the child has no genetic diagnosis.
• An urgent presentation requires clinical management to proceed according to screening results and clinical suspicion. Sequencing should support that process, not serve as permission to begin it.

Reported cohort yields of 56%, 58%, 66.8%, 68.4%, and 32.48% describe different populations and different testing contexts. They should not be collapsed into a universal diagnostic-yield ranking. The Turkish comparison suggests that panel and WES yields can be close in a tertiary-care cohort. The panel-first cohort shows that reflex WES may add diagnoses after a substantial number of children have already been solved by panel testing. The Mexican cohort demonstrates both the variability of yield between populations and the possibility that molecular findings can change disease classification.

The 10.42% reclassification figure is clinically meaningful because it shows that phenotype-based labels can be revised by genomic evidence. It is not evidence of platform equivalence, and it is not a WES-specific reclassification rate. That distinction is more than statistical housekeeping. It prevents a reasonable observation about the value of genomic diagnosis from becoming an unsupported claim about which sequencing platform caused it.

For laboratories designing a pediatric immunodeficiency pathway, the most defensible strategy is often a deliberately conditional one. Use the assay that matches the current clinical question, define what a negative result does and does not exclude, and build a clear route to broader or complementary testing. Panels and WES should be treated as parts of a diagnostic system rather than mutually exclusive ideologies.

The practical question is not whether breadth or depth is inherently better. It is whether the chosen test has enough breadth to include the plausible diagnosis, enough depth to detect the relevant variant, enough technical validation to make a negative result meaningful, and enough workflow support to deliver an answer while it can still change care.