Whole genome sequencing in cerebral palsy: a UK paediatric pilot study
Thiloka E Ratnaike1,2,3, Heather H Pierce1, Alison J Coffey4
1Department of Paediatrics, University of Cambridge, Cambridge, UK.
Insights
Whole genome sequencing (WGS) is a clinically useful tool for diagnosing genetic causes of cerebral palsy (CP) in UK patients. This pilot study found diagnostic variants in 12.8% of individuals, informing prognosis and management.
Area of Science:
- Genetics
- Neurology
- Genomic Medicine
Background:
- Monogenic conditions affect 9-36% of individuals with cerebral palsy (CP).
- The diagnostic utility of whole genome sequencing (WGS) for UK National Health Service (NHS) CP patients was previously unevaluated.
Purpose of the Study:
- To evaluate the clinical utility of gene-agnostic trio WGS for diagnosing genetic conditions in UK CP patients.
- To assess the effectiveness of AI-driven variant prioritization and human phenotype ontology (HPO) terms in diagnostic yield.
Main Methods:
- Prospective pilot study of 86 individuals with CP undergoing trio WGS.
- AI-based variant prioritization followed by application of a CP-specific gene list.
- Multidisciplinary review of candidate pathogenic/likely pathogenic variants and HPO term analysis using machine learning.
Main Results:
- Diagnostic pathogenic/likely pathogenic variants were identified in 11/86 (12.8%) participants.
- Variants strongly suggestive of disease causation were found in 9.3% of cases.
- Findings significantly informed patient prognosis, management, and familial recurrence risk.
Conclusions:
- Whole genome sequencing (WGS) is a valuable diagnostic and management tool for genetic conditions associated with CP in the UK.
- Further validation in a larger cohort is recommended to confirm these findings.
Background:
Prior international studies indicate that 9-36% of people with cerebral palsy (CP) have a monogenic condition. However, the utility of whole genome sequencing (WGS) as a diagnostic tool for United Kingdom (UK) National Health Service (NHS) patients has not been evaluated.
Methods:
This prospective pilot study recruited 86 individuals with CP from specialist clinics in Bedford, Cambridge, Colchester, Newcastle, and Luton NHS Foundation Trusts. Gene-agnostic trio WGS was performed using AI-based variant prioritisation, with subsequent application of a CP gene list. Candidate diagnostic pathogenic (P) or likely pathogenic (LP) variants were reviewed at multidisciplinary meetings and confirmed in an NHS Genomic Laboratory Hub prior to issuing a clinical report. The use of human phenotype ontology (HPO) terms was evaluated to estimate probability of a diagnostic variant using a supervised linear discriminant analysis (PCA + LDA) model.
Findings:
86/157 (54.7%) individuals approached consented to the study. Variants meeting P/LP diagnostic criteria were identified in 11/86 cases (12.8%). 8/86 participants (9.3%) carried variants strongly suggestive of disease causation. Variants of uncertain significance were identified in 27/86 cases (31.4%). In all cases with P/LP variants, findings informed patient prognosis, specialist care, clinical management, and familial recurrence risk. Machine learning approaches were used to segregate the probability of diagnosis for participants based on HPO terms.
Interpretation:
WGS is clinically useful for diagnosis and management of genetic conditions associated with CP in the UK. Validation of these findings in a larger cohort is warranted.
Funding:
Rosetrees Charitable Trust, Isaac Newton Trust, NIHR Cambridge Biomedical Research Centre, and the Wellcome Trust.
