Integrating Schizophrenia Genetic Risk Variants Enhances Alzheimer's Disease Prediction
A recent preprint study published on medrxiv.org investigated the potential of integrating schizophrenia-associated genetic risk variants to enhance Alzheimer's disease (AD) prediction. The study utilized large-scale genome-wide association study (GWAS) data from 1,126,563 individuals with AD and 320,404 individuals with schizophrenia. By incorporating top 50 schizophrenia-associated single nucleotide polymorphisms (SNPs), the study found that AD prediction accuracy improved significantly, from 0.6032 to 0.6400, with a DeLong's test p-value of 0.0005. This breakthrough demonstrates meaningful genetic overlap between AD and schizophrenia and highlights the potential of this approach for dissecting the complex genetic architecture of brain disorders.
Key Takeaways:
- The study integrated statistically-selected schizophrenia-associated SNPs to enhance AD prediction beyond models using only AD-associated SNPs.
- The use of top 50 schizophrenia-associated SNPs improved AD prediction accuracy in the best-performing machine learning (ML) model, increasing the area under the receiver operating characteristic curve (AUC) from 0.6032 to 0.6400.
- The contributing schizophrenia SNPs implicated shared biological pathways relevant to AD, including immune regulation/neuroinflammation, tau protein biology, and synaptic vesicle trafficking.
- The study revealed novel predictive variants warranting further investigation, providing new genetic targets for cross-disorder research.
Statistics:
- 1,126,563 individuals with AD were used in the large-scale GWAS data.
- 320,404 individuals with schizophrenia were used in the large-scale GWAS data.
- The DeLong's test p-value for the improved AD prediction accuracy was 0.0005.
- The area under the receiver operating characteristic curve (AUC) for the baseline model was 0.6032.
- The improved AUC for the ML model using top 50 schizophrenia-associated SNPs was 0.6400.
Sources:
- medrxiv.org/content/10.1101/2025.04.24.25326362v1