11 June 2026
As genomic testing becomes more widely adopted in clinical care, many cases remain unsolved—not because the sequencing failed, but because the variants identified are exceptionally rare, difficult to classify, or hidden in challenging regions of the genome. Clinicians and laboratory scientists are increasingly facing a flood of genomic information without enough evidence to make sense of all the data.
In many difficult cases, sequencing produces technically sound results. The challenge is determining whether a detected variant is truly disease-causing—or merely incidental background noise.
Often testing will be inconclusive because of variants of uncertain significance, or VUSs. These are genetic findings that lack sufficient evidence to be classified as either benign or pathogenic.
“It might not be significant at all,” says Livia Loureiro, a medical science liaison at Illumina. “We don’t have evidence yet to tell whether a variant is associated with any phenotype.”
Sometimes a variant appears rare and biologically plausible, perhaps affecting a protein-coding region or altering protein function. But without published studies, additional patient cases, or functional evidence, laboratories often cannot confidently determine whether it contributes to disease. Cases with only VUS results are often reported back as unresolved—and from there, the path forward remains unclear.
The rise of reanalysis
One emerging strategy for solving previously negative cases is reanalysis. As genomic knowledge evolves, new gene-disease associations and updated databases can transform prior results into diagnoses.
Loureiro describes a case in which reanalysis changed everything. The case was presented during one of Illumina’s virtual Grand Rounds sessions. A boy with global motor delay, developmental regression, seizures, bronchospasms, and hypotonia received negative results when his physician ordered exome sequencing. Two years later, the family requested a reanalysis of the exome. This time, the report identified a pathogenic variant in the GLUL gene.
“The diagnosis was only possible because, just a few months before the reanalysis, a new publication had described the association between de novo variants in GLUL and developmental and epileptic encephalopathy through stabilization of glutamine synthetase. This case highlights the critical importance of periodic genomic reanalysis and the rapidly evolving nature of genomic knowledge. New scientific evidence enabled the healthcare team to recognize and report the disease-causing variant, ultimately providing the family with a long-awaited, definitive diagnosis for their child.”
The ability to continuously integrate new evidence is becoming increasingly important as laboratories process growing test volumes. Many labs maintain home-built analysis pipelines, but keeping those systems updated with the latest publications, databases, and classification guidelines can be resource-intensive.
Difficult variants in oncology
Interpretation challenges extend beyond rare disease diagnostics.
In oncology, identifying actionable biomarkers increasingly depends on detecting complex genomic signatures such as tumor mutational burden (TMB), microsatellite instability (MSI), homologous recombination deficiency (HRD), and gene fusions. Comprehensive genomic profiling (CGP) can detect all these signatures and increase the number of actionable mutations, potentially leading to improved outcomes. However, testing can also produce a vast amount of data, which can be challenging for oncologists to interpret.
“If you don’t know how to do the interpretation, you might miss the criteria to qualify that specific patient for a targeted therapy or drug,” says Loureiro.
Tertiary analysis interprets variant data and reports the clinical relevance of each variant. “By characterizing the somatic landscape, we can refine diagnoses, expand target identification, identify individuals likely to benefit from target therapies and/or hereditary testing, detect novel therapeutic targets, provide prognostic insights, and potentially improve patient outcomes,” Loureiro explains.
Some tumor types, such as sarcomas, require especially sophisticated analysis. Detecting a fusion event may not be enough; clinicians often need to identify the precise fusion partner to determine oncogenicity and treatment relevance.
Integrating other information and expertise
Other difficult cases demand broader contextual information. Phenotype-driven interpretation—matching genomic findings to detailed clinical symptoms—can dramatically influence diagnostic accuracy.
“If you have a rare disease and you input ‘developmental delay,’ there are a lot of genes that are associated with developmental delays,” Loureiro explains. “But if you can include specific information…it might help you to do this phenotype-genotype match.”
That means genomics increasingly depends on collaboration between clinicians, genetic counselors, geneticists, pediatricians, pathologists, neurologists, oncologists, cardiologists, and bioinformaticians.
“For rare diseases, we may not always have the same structured multidisciplinary team commonly seen in molecular tumor boards in oncology. However, the involvement of multiple experts is essential to fully understand the patient’s clinical presentation, interpret the reported variants, and guide appropriate clinical management,” Loureiro says.
AI enters the interpretation workflow
As genomic datasets grow larger and more complex, artificial intelligence (AI) is becoming an important enabler of efficient interpretation workflows. Many laboratories are exploring AI-assisted analysis tools to help prioritize variants and accelerate review.
“AI is not going to replace expert judgment,” Loureiro says. “But it can prioritize variants early in the review process and significantly reduce time to insight.” AI is already embedded in many genomic workflows, from variant calling to predictive algorithms used to assess splice effects and pathogenicity. The challenge now is helping laboratories understand how to responsibly implement these tools. Illumina is supporting this shift by enabling scalable, continuously updated analysis environments that integrate evolving evidence with advanced prioritization tools, such as Emedgene and Illumina Connected Insights for genetic disease research and oncology research.
Building a global learning community
To help address growing interpretation challenges, Illumina has launched an educational Grand Rounds in Genomic Medicine webinar series featuring clinical geneticists, molecular pathologists, and genomic scientists discussing real-world genomic cases across oncology, rare disease, cardiology, and reproductive health.
These sessions highlight how experts interpret challenging variants, integrate clinical context, and navigate uncertainty, while fostering a global learning community. The audience can ask questions during the live presentations, creating highly interactive discussions. “People ask about the tools being used, the types of evidence considered for variant interpretation, and what clinical information is necessary to evaluate a specific case,” Loureiro says. The webinars have drawn attendees from around the world, reflecting the growing universal demand for genomic interpretation expertise.
As genomics continues moving deeper into mainstream medicine, solving the most complex cases will require more than sequencing power alone. It will depend on intelligent workflows, collaborative expertise, and continuous evolving knowledge.
To watch a recorded session, or to register for the next live session, follow this link.
Find variant interpretation cases and other educational resources at End the Odyssey.
Learn more about AI-enabled analysis tools for rare disease research and oncology research.
