摘要：In this study, we split 2156 individuals from the Chinese Longitudinal Healthy Longevity Survey (CLHLS) data into two groups, establishing a phenotype of exceptional longevity & normal cognition versus cognitive impairment. We conducted a genome-wide association study (GWAS) to identify significant genetic variants and biological pathways that are associated with cognitive impairment and used these results to construct polygenic risk scores. We elucidated the important and robust factors, both genetic and non-genetic, in predicting the phenotype, using several machine learning models. The GWAS identified 28 significant SNPs at p-value $$< 3 imes 10^{-5}$$ significance level and we pinpointed four genes, ESR1, PHB, RYR3, GRIK2, that are associated with the phenotype though immunological systems, brain function, metabolic pathways, inflammation and diet in the CLHLS cohort. Using both genetic and non-genetic factors, four machine learning models have close prediction results for the phenotype measured in Area Under the Curve: random forest (0.782), XGBoost (0.781), support vector machine with linear kernel (0.780), and $$ ll _2$$ penalized logistic regression (0.780). The top four important and congruent features in predicting the phenotype identified by these four models are: polygenic risk score, sex, age, and education.
其他摘要：Abstract In this study, we split 2156 individuals from the Chinese Longitudinal Healthy Longevity Survey (CLHLS) data into two groups, establishing a phenotype of exceptional longevity & normal cognition versus cognitive impairment. We conducted a genome-wide association study (GWAS) to identify significant genetic variants and biological pathways that are associated with cognitive impairment and used these results to construct polygenic risk scores. We elucidated the important and robust factors, both genetic and non-genetic, in predicting the phenotype, using several machine learning models. The GWAS identified 28 significant SNPs at p -value $$< 3 \times 10^{-5}$$ 3 × 10 - 5 significance level and we pinpointed four genes, ESR1 , PHB , RYR3 , GRIK2 , that are associated with the phenotype though immunological systems, brain function, metabolic pathways, inflammation and diet in the CLHLS cohort. Using both genetic and non-genetic factors, four machine learning models have close prediction results for the phenotype measured in Area Under the Curve: random forest (0.782), XGBoost (0.781), support vector machine with linear kernel (0.780), and $$\ell _2$$ ℓ 2 penalized logistic regression (0.780). The top four important and congruent features in predicting the phenotype identified by these four models are: polygenic risk score, sex, age, and education.
