Rationale:
INTRODUCTION
Seizures are a common complication in many rare diseases and often contribute to diagnostic delays and management challenges. Epilepsy, characterized by recurring seizures due to abnormal neuronal activity, is frequently linked to underlying genetic mutations. Its presentation varies widely across rare conditions, often co-occurring with other neurodevelopmental symptoms. Identifying patient-level patterns of seizure-related comorbidities is critical to improving understanding and care strategies across these diverse populations.
RARE-X, the research initiative of Global Genes, collects patient-reported data based on symptom profiles rather than diagnosis alone. With over 9,000 participants representing 80+ rare conditions—including many neurodevelopmental disorders—RARE-X enables hypothesis-agnostic, cross-condition research. The platform’s focus on shared and unique symptom patterns supports discovery across traditionally siloed diseases.
Methods:
This proof-of-concept study applied cluster analysis to identify comorbidity patterns associated with seizures across rare diseases using RARE-X data. We examined patient-reported data from 10 rare conditions, categorizing them into High Seizure (HS: 50–85%) and Low Seizure (LS: 1–33%) cohorts, comprising 610 and 1,030 participants, respectively. A total of 246 patients with seizures and 299 without were included.
Previously, we conducted heatmap analyses using aggregate, condition-level data, which suggested the existence of more than one phenotypic cluster associated with seizure symptoms. Building on those findings, we now extend the analysis by incorporating patient-level data and applying unsupervised clustering. This approach uses combinations of key neurodevelopmental symptoms (e.g., developmental delay, motor impairment, behavioral challenges) to detect phenotypic patterns at the individual level. Heatmaps were also used to visualize symptom co-occurrence, and seizure frequencies were compared with the NIH’s GARD database to assess concordance and condition coverage.
Results: Initial review of disease-specific data highlights a range of co-occurring neurodevelopmental symptoms reported alongside seizures. Early clustering efforts are exploring symptom patterns across conditions with high and low seizure frequencies. These analyses aim to identify potential subgroups based on patient-reported comorbidities and to assess the value of PRO data in uncovering shared phenotypic profiles across rare diseases.
Conclusions: -Cluster analysis of patient-reported outcomes (PRO) data revealed distinct subgroups based on epilepsy and other related symptoms, underscoring the potential of PRO data to uncover meaningful phenotypic patterns that may not be apparent through diagnosis alone.
-The shift from condition-level to patient-level data analysis enhances resolution and allows for the identification of more nuanced patterns within and across rare diseases.
-RARE-X’s symptom-based structure enabled the identification of cross-condition similarities, even in rare disorders with limited sample sizes.
Funding: N/A