O28: GREGoR: Accelerating genomics for rare diseases

detailed information on growth dynamics, such as changes in shape and trajectory over time, limiting their clinical applicability.Although disease-specific growth curves exist for some conditions, such as Trisomy 21 and Turner syndrome, many genetic diseases lack published growth curves.To address these gaps, we utilized real-world electronic health record (EHR) data to systematically evaluate weight, height, and body mass index (BMI) differences in 376 genetic disorders.This effort produced a comprehensive catalog of growth-related abnormalities and disease-specific growth curves aimed at supporting personalized growth assessments for individuals with genetic diseases.Methods: We utilized EHR data from the Clinical Genetic Database (CGdb), a curated collection of genetic testing results from individuals seen at Vanderbilt University Medical Center between 2002 and 2023.Biometric data on height, weight, and BMI were cleaned and analyzed separately for adults (age 20 years) and pediatric individuals (age 2 and < 20 years).Growth-related terms from the Human Phenotype Ontology (HPO) were categorized and linked to OMIM disease IDs.Positive controls (eg, achondroplasia for low height) were established for each biometric measurement.Linear regression models were used to evaluate the association between disease status and the median z-scores of each biometric, adjusting for demographic factors.Pediatric individuals and adults were assessed separately and only diseases with at least five affected cases were included in the analysis.Associations were considered significant at a Bonferroni-corrected p-value threshold (p < 0.05).Significant findings were compared against OMIM annotations for replication.Disease-specific growth curves for the pediatric cohort were modeled using nonlinear mixed effects models, capturing clinically relevant growth parameters such as size, timing, and velocity.Results: Our cohort included 1,467,025 individuals with 18,216 having one or more molecularly confirmed genetic diagnoses in CGdb.We analyzed 376 diseases (327 pediatric and 186 adult).One hundred forty-one significant associations were identified across the three measurements with height accounting for 87 (62%) of these.Of the significant associations, 112 (79%) had growth abnormality annotations in OMIM, and 76 (54%) were exact replications.Many conditions showed strong and consistent replication across the adult and pediatric cohorts (eg, Down syndrome, DiGeorge syndrome, Prader-Willi, etc) while others were replicated in only the adult cohort (eg, Klinefelter syndrome, Trisomy X, etc), reflecting variation in how these conditions manifest at different ages.Our analysis highlighted numerous associations that align with existing literature but are currently not documented in OMIM, including an association with shorter height for cystic fibrosis and Multiple Endocrine Neoplasia Type 1, and lower BMI for Huntington disease.We generated sex-specific growth curves for 15 unique genetic diseases, including Down syndrome, cystic fibrosis, and Duchenne muscular dystrophy, that had sufficient cases in the pediatric cohort to ensure model reliability, providing a comprehensive catalog of disease trajectories.Conclusion: Our findings demonstrate the potential of EHR data to replicate and expand upon existing HPO annotations of growth abnormalities for genetic diseases.The disease-and sex-specific growth curves derived from this study offer a robust, data-driven foundation for enhancing the diagnosis, monitoring, and personalized management of genetic conditions.