STATISTICAL GENETICS MODELS FOR INTEGRATING EPIGENOMIC AND GENOMIC DISEASE RISK FACTORS

Abstract

Background: Complex human diseases are contributed by both genomic variations and epigenomic modifications via regulating gene expression, cellular pathways and disease susceptibility. Single-omics approaches are often not enough to fully explain the heritability of disease and biological complexity.

Objective: The study evaluated advanced statistical genetics models for genomic and epigenomic disease risk factor integration to improve the accuracy of disease prediction and precision medicine applications.
Methods:  We analysed genome-wide association study (GWAS) datasets, DNA methylation profiles and multi-omics datasets using Bayesian regression, mixed linear models, polygenic risk scoring and deep learning frameworks. The performance of predictive efficiency and biomarker identification was evaluated based on machine learning algorithms and cross-validation approaches.

Findings: Integrative multi-omics statistical models reached disease classification accuracies of approximately 88-94%, greatly outperforming traditional genomic-only approaches. Deep learning based models achieved the highest predictive accuracy and improved identification of disease associated genomic and epigenomic biomarkers. The heritability estimation and the personalized disease risk prediction were also improved by multi-omics integration.

Conclusion: Models in statistical genetics that integrate epigenomic and genomic information provide powerful platforms for precision disease prediction and biomarker discovery. Advanced AI-assisted multi-omics systems could greatly benefit future personalized medicine and early disease intervention strategies.