Identifying Individuals at High Risk for Alzheimers Disease Among Hispanics Using Single and Multi-Ancestry Polygenic Risk Scores

IntroductionPolygenic risk score (PRS) is effective in predicting AD risk among Europeans but remains understudied in Hispanics. Diverse genome-wide association studies (GWAS) data across multiple ancestries may improve PRS predictions. We evaluated PRS performance to predict AD disease risk using novel methods in the largest available African, European, and Hispanic GWAS for AD. MethodsPrediction performance of APOE, single-ancestry PRS, and multi-ancestry PRS derived from GWAS-focused and method-focused approaches to clinical AD, incident AD, and cognition were evaluated in 2,961 Hispanics from two large studies. The GWAS-focused approach constructs PRS based on multi-ancestry GWAS, while the method-focused approach uses novel multi-ancestry PRS methods, integrating GWAS summary statistics across ancestries. Ten repetitions of 5-fold cross-validation were used. In a subset, plasma biomarker data were used in a tuning-validation split to examine PRS performance in predicting single and combined biomarkers. FindingsThe multi-ancestry PRS excluding APOE, constructed using the method-focused approach, outperformed both single-ancestry and multi-ancestry PRSs from the GWAS-focused approach. The best method-focused PRS, incorporating summary statistics from GWASs of African, European, and Hispanic populations, explained up to 1.6%, 3.9%, and 1.7% of the variance in clinical AD, incident AD, and cognition, respectively - comparable to or even higher than the variance explained by the APOE. Similar findings were observed in biomarker analyses. APOE accounted for more variation in plasma P-tau levels and PRS explained more variation in A{beta} levels. InterpretationIntegrating novel multi-ancestry PRS methods with GWAS across ancestries enhances prediction accuracy for AD risk among Hispanics. APOE and PRS may point to different biological aspects of AD. FundingNational Institutes of Health R01 AG072474, RF1 AG066107, 5R37AG015473, RF1AG015473, R56AG051876, R01 AG067501, and UL1TR001873. Research in contextO_ST_ABSEvidence before this studyC_ST_ABSWe searched PubMed for research related to PRS prediction of AD in Europeans, Africans, and Hispanics, published from database inception to August 1, 2024, without any language restrictions. The search terms used were "PRS," "PGS," "Hispanics," "Latinos," "AD," and "plasma biomarkers." We considered only peer-reviewed reports in English. Previously, few studies have examined the performance of PRS in predicting clinical AD, incident AD, and mild cognitive impairment (MCI) among Hispanics. However, none of these studies utilized advanced methods for constructing multi-ancestry PRS, validated PRS performance among Hispanics, or examined plasma biomarkers. Added value of this studyThe present study demonstrated that integrating novel multi-ancestry PRS methods with GWAS from African, European, and Hispanic populations enhanced prediction accuracy for AD risk among Hispanics. Among Hispanics, PRS explains a similar or higher amount of variance compared to APOE. Plasma biomarker analyses suggests that APOE may also be strongly related to variation in P-tau levels, while PRS may explain variance in A{beta} levels. Implications of all the available evidenceAmong Hispanics, PRS complements the effects of APOE and has the potential to identify at-risk populations for clinical trial eligibility and early biomarker screening. Although AD genetic studies are still limited among Hispanics, a dynamic combination of advanced methods with GWAS across populations could substantially improve prediction performance in this population, which in turn may reduce health disparities.