Clinical geneticists confronting unresolved rare disorders should shift their diagnostic focus from isolated genomic coordinates toward comprehensive molecular configurations that capture structural context, phasing, and epigenetic marks across native DNA molecules. In a perspective published in Frontiers in Genetics, researchers led by Dr. Rajdeep Das and colleagues examine why standard diagnostic pipelines frequently leave patients without answers despite technically successful short-read exome and genome sequencing. While short-read sequencing has transformed clinical genetics and related medical specialties, it evaluates discrete variants individually using population frequencies, predicted consequences, segregation, and gene-disease validity. This established framework serves single-nucleotide variants and small insertions or deletions exceptionally well, yet it falters when a pathogenic mechanism depends on continuity across extended genomic intervals. By maintaining molecular continuity across kilobases of native genetic material, long-read sequencing assays can preserve vital relationships among sequence variants, structural breakpoints, repeat motifs, and allele-specific regulatory states, transforming how clinicians identify and classify complex disease-causing alleles.
The persistent limitation of standard diagnostic workflows is highlighted by the diagnostic odyssey experienced by families enduring repeated specialist consultations, serial testing procedures, provisional diagnoses, and protracted uncertainty. These inconclusive trajectories carry substantial personal and financial costs, depriving patients of timely disease-specific surveillance, targeted therapeutics, reproductive counseling, or clinical trial eligibility, while burdening health systems with low-yield testing regimens. Citing a 2025 cohort study by Glaubitz and colleagues, the authors note that patients suspected of having rare diseases demonstrate substantially greater pre-referral healthcare utilization and direct medical costs than matched controls, demonstrating that unresolved diagnostic ambiguity directly drains clinical resources. The failure of standard testing in these scenarios often stems not from an inability to detect that a variant exists, but from an inability to discern how separate genomic signals relate to one another in physical space. Deeper short-read sequencing may increase statistical confidence at an isolated coordinate, but it cannot recreate physical molecular linkages that were never captured in the original fragmented library preparation.
To illustrate how technical success can coexist with diagnostic failure, the authors describe a composite clinical presentation involving a child exhibiting unambiguous features of a recessive disorder. Short-read genome sequencing in this scenario identifies one convincing pathogenic coding variant alongside an adjacent, small copy-number gain categorized as a variant of uncertain significance, leaving the patient with two isolated findings, two unlinked classifications, and no definitive diagnosis. When long-read sequencing is applied to the same presentation, it reveals that the coding variant resides on the maternal haplotype while the copy-number gain on the paternal haplotype represents a tandem duplication spanning two exons in trans. Complementary long-read RNA sequencing from blood subsequently confirms the expression of an out-of-frame transcript derived specifically from the paternal allele. The complete molecular configuration demonstrates compound heterozygosity without requiring the discovery of an entirely new disease gene, resolving the diagnostic question by physically connecting, phasing, and characterizing existing signals that short-read assays could only observe as disconnected fragments.
Three primary molecular relationships frequently account for diagnostically complex cases that elude conventional variant-centered evaluation. The first critical relationship is phase, because two pathogenic variants located within a recessive gene yield a clinical diagnosis only when positioned in trans on opposing chromosomes; the identical alterations occurring in cis on the same chromosome may instead represent a benign co-occurrence or a complex allele with unique functional consequences. The second relationship involves structural context, where simple copy-number enumeration fails to explain the true biological impact of genomic gains and losses. Orientation, insertion sites, topologically associated domains, proximity to regulatory elements, and exact breakpoint architecture determine whether a duplication disrupts a transcription unit, generates a fusion gene, alters dosage, or exerts minimal functional effect. The third relationship involves regulatory and epigenetic state, where sequence variation, DNA methylation, and transcription are routinely captured through disconnected platforms, obscuring which parental allele is methylated, expressed, or silenced in imprinting disorders and repeat-associated conditions.
Long-read sequencing technologies address these diagnostic challenges through two primary, complementary platforms: PacBio HiFi sequencing and Oxford Nanopore sequencing. PacBio HiFi technology produces highly accurate consensus reads that excel at precise small-variant detection, breakpoint definition, repeat characterization, and regional haplotype phasing. Oxford Nanopore technology sequences native DNA or RNA molecules directly, producing substantially longer reads capable of spanning complex structural rearrangements, resolving ultra-long haplotypes, supporting adaptive sampling, and reading chemical base modifications without requiring bisulfite conversion. Because PCR amplification strips native epigenetic marks, assessing direct DNA methylation requires native-DNA sequencing workflows rather than amplified libraries. Furthermore, the authors emphasize that long-read RNA sequencing must be recognized as a distinct, complementary assay assessing expressed transcript consequences rather than an additional readout obtained from the same genomic DNA molecule. When appropriately applied, these sequencing modalities span repetitive elements and structural junctions, allowing deletions, insertions, inversions, translocations, and mobile-element insertions to be reconstructed as intact biological events.
The value of characterizing intact molecular structures becomes especially evident at homologous or pseudogene-rich genomic regions and repeat-expansion loci. Clinically critical loci such as CYP2D6, SMN1 and SMN2, GBA1, HLA, and KIR are notoriously intractable on short-read platforms because sequence homology obscures whether an observed variant belongs to a functional gene, an inactive pseudogene, or a specific structural haplotype. Similarly, for repeat-expansion disorders, basic length estimation provides an incomplete diagnostic descriptor. Internal repeat purity, interrupting motifs, somatic mosaicism, methylation patterns, and haplotypic background heavily dictate clinical penetrance and age of onset. In RFC1-related cerebellar ataxia, neuropathy, and vestibular areflexia syndrome, pathogenicity depends entirely on the specific expanded pentanucleotide motif rather than repeat length alone. In FGF14-related spinocerebellar ataxia 27B, disease penetrance tracks tightly with motif purity near narrow size boundaries, whereas in NOTCH2NLC-related disorders, internal GGC composition and methylation distinguish distinct clinical phenotypes at identical overall allele sizes.
Recent clinical cohorts demonstrate the diagnostic utility of long-read sequencing while illustrating important methodological nuances. In a targeted study of four families evaluated for HYDIN-related ciliopathy, nanopore sequencing uncovered pathogenic deletions in two probands and resolved apparent homozygosity in two others. A study by Sinha and colleagues achieved five genetic diagnoses among fifty-one exome-negative patients by identifying structural alterations, copy-number changes, and methylation abnormalities. Evaluating ninety-six short-read-negative individuals with neurodevelopmental phenotypes, Hiatt and colleagues found that seven cases required long reads to resolve complex structural or repeat alterations, whereas nine additional findings could be identified through reanalysis of existing short-read data. In another cohort of one hundred fourteen families, Steyaert and colleagues solved twelve cases involving small variants, structural changes, and repeats, using phasing to clarify ambiguous relationships. Similarly, Negi and colleagues identified diagnostic variants in eleven of forty-two affected individuals, though five had prior diagnoses, three were potentially resolvable via short-read reanalysis, and only three strictly depended on long-read phasing, methylation, or paralog resolution.
Because published cohorts consist of clinically preselected individuals who previously underwent extensive, non-diagnostic investigations, their reported yields cannot be extrapolated to unselected rare-disease populations. The authors propose three practical implementation models tailored to specific healthcare contexts: a reflex model, a targeted model, and a comprehensive whole-genome model. The reflex model deploys long-read sequencing specifically after standard exome or short-read genome testing proves inconclusive, directly addressing residual molecular questions. The targeted model applies long-read assays directly to loci known to defeat short reads, including repeat-expansion genes, complex pseudogene regions, pharmacogenes, and immune loci. The whole-genome model applies broad long-read sequencing under rigorous institutional policies covering validation, consent, secondary findings, and reporting. Regardless of the deployment strategy, laboratories must validate assays independently for each reportable variant class, establishing separate reference materials and analytical performance metrics for small variants, structural rearrangements, copy-number variants, repeat expansions, phasing, methylation, and transcript isoforms rather than treating long-read sequencing as a single, uniform test.
Beyond analytical validation, the transition to context-aware genomics faces structural bottlenecks in data infrastructure, standard reference materials, and health economics. Current clinical databases and electronic reporting systems remain overwhelmingly formatted around isolated variants mapped to single linear genomic coordinates, leaving them fundamentally incapable of cataloging multi-breakpoint structural rearrangements, phased haplotypes, or joint epigenetic states. Although resources like the telomere-to-telomere T2T-CHM13 assembly and the Human Pangenome Reference Consortium provide sophisticated pangenomic frameworks that capture human structural diversity, clinical adoption of these tools remains limited. This infrastructure gap also intersects with health equity, as reference resources and population allele frequencies remain disproportionately thin for historically underrepresented ancestries. Furthermore, biological limitations persist, as long-read molecules do not always span the full distance required for definitive phasing, base modification calls vary by chemistry and base-calling software, and long-read RNA testing depends strictly on obtaining accessible tissues that express the pathogenic transcript.
Evaluating the economic viability of long-read sequencing requires comprehensive Health Economics and Outcomes Research frameworks rather than crude, isolated comparisons of laboratory test costs. Although raw per-sample sequencing costs routinely favor short-read workflows, simple price comparisons fail to capture the broader diagnostic trajectory, where long-read reflex testing competes against lengthy series of low-yield specialist visits, imaging studies, and invasive procedures. Economic assessments must evaluate complete diagnostic pathways from prespecified health-system and societal perspectives across appropriate time horizons, tracking costs associated with library preparation, sequencing failures, data storage, bioinformatic compute, clinical interpretation, genetic counseling, and orthogonal confirmation. Documented clinical benefits must measure avoided diagnostic procedures, time to conclusive diagnosis, definitive management modifications, and patient or caregiver outcomes, balanced against potential harms stemming from ambiguous findings. Crucially, the authors note that a molecular diagnosis does not automatically reduce downstream healthcare expenditure, highlighting that prospective studies must rigorously evaluate the specific hypothesis that long-read datasets yield superior clinical and economic value during periodic genomic reanalysis.
Regulatory and variant classification frameworks must simultaneously adapt to interpret integrated molecular configurations. Current guidelines from the American College of Medical Genetics and Genomics, the Association for Molecular Pathology, and the Clinical Genome Resource evaluate variants strictly on a coordinate-by-coordinate basis. Within these conventional rubrics, criteria such as PM3 treat trans configurations merely as external supporting evidence for a single variant rather than recognizing the phased compound-heterozygous pair as an indivisible diagnostic entity, while structural variant guidelines focus heavily on genomic dosage rather than fine architecture. At present, reflex long-read testing is most robustly supported for phenotype-consistent recessive disorders with one unresolved allele or unknown phase, suspected repeat expansions with ambiguous motif structure, imprinting disorders with unresolved methylation, and complex rearrangements requiring structural reconstruction. Transforming long-read sequencing from an exploratory technology into routine care will require professional societies, diagnostic laboratories, and payers to establish coordinated classification standards, flexible database architectures, and prospective clinical trial evidence demonstrating that reading complete molecular configurations directly improves patient outcomes.
Subject of Research: Context-aware genome interpretation using long-read sequencing technologies for rare disease diagnosis
Article Title: Addressing the diagnostic odyssey beyond variant calling: a perspective on long-read sequencing and context-aware genome interpretation
Article References: Das, R., Pina, J., Deeb, K. K., Reddi, H. V., Mao, R., & Shetty, S. (2026). Addressing the diagnostic odyssey beyond variant calling: a perspective on long-read sequencing and context-aware genome interpretation. Frontiers in Genetics, 17, 1965842. https://doi.org/10.3389/fgene.2026.1965842
Image Credits: AI Generated
DOI: 10.3389/fgene.2026.1965842
Keywords: long-read sequencing, context-aware interpretation, diagnostic odyssey, rare disease genomics, haplotype phasing, structural variants, repeat expansions, allele-specific methylation, PacBio HiFi, Oxford Nanopore, long-read RNA sequencing, recessive disease, health economics, pangenome reference, variant classification, clinical genetics, paralog resolution
News Source: Juliet Wilcox. (November 13, 2026). Long-Read Sequencing Framework Evaluates Molecular Configurations to Resolve Complex Genetic Diagnoses. Scienmag.
