The German healthcare system has created a framework for identifying patients with ultra-rare diseases by analyzing next-generation sequencing (NGS) and phenotyping data. While a structured phenotypic assessment combined with an advanced sequencing test such as exome sequencing could reduce diagnostic delays, the addition of AI-driven modeling significantly improved variant prioritization. The study, published in Nature Genetics, demonstrates the synergy of using NGS and phenotyping to diagnose ultra-rare diseases in routine healthcare and discover novel etiologies by multidisciplinary teams.
Finding a national framework for diagnosing patients with ultra-rare disorders
Exome sequencing has been demonstrated to be more cost-effective than sequencing potentially multiple gene panels for complex and non-monogenic genetic diseases. However, using exome sequencing presents two significant challenges. First, the analysis becomes much more complex because this situation often identifies additional genetic variants that must be assessed. As a result, clear indications for exome sequencing and practical data analysis strategies are critical. Second, a healthcare system must have the infrastructure and resources to carry out such an endeavor.
The German healthcare system presents a viable testing ground for a nationwide approach as around 90% of the population has statutory health insurance, and the current reimbursement scheme allows physicians to request a variety of genomic tests, including chromosome analyses, molecular karyotyping, and sequencing of single genes or gene panels. The German healthcare system conducted a prospective study called TRANSLATE NAMSE from 2018 to 2020 to create a structured method for incorporating exome sequencing into the analysis of clinical phenotypes effectively and efficiently for diagnosing ultra-rare disorders and finding new monogenic disorders. Additionally, a companion study of a consenting subcohort from TRANSLATE NAMSE called PEDIA (prioritization of exome data by image analysis) assessed how the results from computer-assisted pattern recognition in facial dysmorphism contribute to variant interpretation.
Next-generation phenotyping and exome sequencing
The Nature Genetics article analyzes this national healthcare endeavor, looking at data from the TRANSLATE NAMSE and PEDIA studies. The TRANSLATE NAMSE consisted of phenotypic and molecular genetic data from 1,577 patients undergoing exome sequencing and was partially analyzed with next-generation phenotyping approaches. The study established genetic diagnoses in 499 patients (32%), totaling 370 distinct molecular genetic causes, most with a prevalence below 1:50,000. This diagnostic process identified 34 novel and 23 candidate genotype-phenotype associations, mainly in individuals with neurodevelopmental disorders.
To determine the likelihood that exome sequencing will lead to a molecular diagnosis in a given patient based on the respective clinical features only, the researchers developed a statistical framework called YieldPred. The resulting model can use clinical features to discriminate between solved and unsolved cases—meaning, whether a variant was found in an established disease gene—and also identify patient phenotypes, such as “dysfunction of higher cognitive abilities,” “hematological abnormalities,” and “ataxia,” as very influential predictors in terms of the establishment of a molecular diagnosis via exome sequencing.
A total of 224 of the 1,577 patients also provided written informed consent to participate in PEDIA, in which their facial images would be analyzed by the AI tool GestaltMatcher to aid in exome variant interpretation. In 94 of these PEDIA subcohort cases, a molecular diagnosis was established, and for 81 of these, the gestalt scores improved prioritization results (that is, the correct diagnosis was ranked higher). The analyses by PEDIA showed that computer-assisted analysis of facial images prioritized variants more efficiently compared with approaches based solely on clinical features and molecular scores.
The study’s results show that this new, structured diagnostic concept makes it easier to find ultra-rare disorders nationally. This national framework is necessary to give patients not yet diagnosed a chance to participate in research and create a place where data can be shared. The systematic, consortium-based collection of molecular and clinical data is the first critical step toward achieving personalized medicine.
