SA105 - IDENTIFICATION OF PHENOTYPIC PREDICTORS OF PATHOGENIC COPY NUMBER VARIANTS IN A PSYCHOSIS POPULATION

Abstract

Background Copy Number Variants (CNVs) significantly impact one's risk of developing a Neurodevelopmental Disorder (NDD). Eleven CNVs are robustly implicated in Schizophrenia (SCZ), increasing risk by 2 to 60-fold. Carriers are also at substantially higher risk of developing other NDDs (penetrance for autism/developmental delay/congenital malformations= 8–88%). Genetic testing is a standard of care for people with Autism Spectrum Disorder (ASD); 10–20% carry CNVs that may have implications for clinical management and genetic counselling. As The aim of this study was 1) to investigate clinically identifiable phenotypic features for association with putatively pathogenic CNVs, in a psychosis population, and 2) to investigate whether associated phenotypic features were predictive of CNV carrier status in patients with psychosis. Putatively predictive phenotypic characteristics may aid the identification of patients who would benefit from clinical genetic testing. Methods The primary analysis was carried out on a data set derived from an Irish SCZ research cohort with extensive clinical phenotype information and Genome-Wide Association Study (GWAS) data. Literature review identified neurodevelopmentally related phenotype variables for correlation analyses in a group of carriers of CNVs at 15 SCZ-associated loci, compared to a group without the CNVs. Phenotype variables identified included: early onset of symptoms, history of learning difficulties, specific learning disorder, remedial school support, low educational attainment, history of developmental delay, comorbid neurodevelopmental diagnosis (ASD, intellectual disability, epilepsy, ADHD), family history of NDD. The probability of pathogenic CNV carrier-status was analysed using logistic regression. A subsequent analysis was undertaken in a second psychosis data set (Cardiff) with similar neurodevelopmental phenotype and genetic data. Results In the Irish data set, nineteen individuals (1.6%) carried risk CNVs from a total sample of 1,215 cases. ‘Specific learning disorder’ (OR=9.0 (95% CI 1.38–39.98, p=0.03)); ‘History of learning difficulties’ (OR=4.0 (95% CI 1.55- 10.30, p=0.005)); ‘History of developmental delay’ (OR=5.8 (95% CI 1.91 - 16.44, p=0.005)) and ‘co-morbid NDD’ (OR=4.9 (95% CI 1.18 - 16.73, p=0.034)) were significantly associated with carrier status. ‘History of learning difficulties’ was significantly correlated with the other three variables. Logistic regression odds ratios for probability of a pathogenic CNV were 9.2, 5.3, 6.4 respectively in the presence of positive history of specific LD, history of developmental delay or co-morbid NDD. Analyses in the Cardiff data set (n=479) were supportive of a significant correlation between the CNV carrier group and presence of co-morbid NDD (p=0.0012). Correlation with history of developmental delay did not reach significance (p=0.13). Broad confidence intervals were observed in analyses in both data sets, indicating need for replication in larger datasets. Discussion These preliminary analyses indicate that symptoms of early developmental compromise, similar to symptoms observed in ASD and developmental delay may have clinical utility in screening for SCZ patients at increased risk of carrying pathogenic CNVs. The finding requires replication in larger datasets which is currently being pursued.