Interpreting artificial neural networks to detect genome-wide association signals for complex traits

2024-07-26

arXiv.org

AI Summary

Conventional genome-wide association studies (GWAS) are constrained by linear assumptions and struggle to detect non-linear genetic effects and epistatic interactions. Method: We propose an interpretable AI-driven GWAS paradigm that employs deep neural networks to model genotype–phenotype relationships, integrates gradient-weighted class activation mapping (Grad-CAM) with SHAP for locus-level feature importance attribution, and introduces a novel framework unifying multiple post-hoc interpretability methods with p-value calibration for statistically robust significance assessment. Results: Applied to the Estonian Biobank schizophrenia cohort, our method successfully replicates established risk loci, yields significant enrichment in brain morphology- and function-related pathways, and achieves superior detection accuracy compared to standard GWAS and state-of-the-art machine learning approaches—providing a new tool for complex trait genetics that balances interpretability with statistical rigor.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models
• Philosophy and Ethics of AI: Accountability, Interpretability & Explainability

Application Category

• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Investigating the genetic architecture of complex diseases is challenging due to the multifactorial and interactive landscape of genomic and environmental influences. Although genome-wide association studies (GWAS) have identified thousands of variants for multiple complex traits, conventional statistical approaches can be limited by simplified assumptions such as linearity and lack of epistasis in models. In this work, we trained artificial neural networks to predict complex traits using both simulated and real genotype-phenotype datasets. We extracted feature importance scores via different post hoc interpretability methods to identify potentially associated loci (PAL) for the target phenotype and devised an approach for obtaining p-values for the detected PAL. Simulations with various parameters demonstrated that associated loci can be detected with good precision using strict selection criteria. By applying our approach to the schizophrenia cohort in the Estonian Biobank, we detected multiple loci associated with this highly polygenic and heritable disorder. There was significant concordance between PAL and loci previously associated with schizophrenia and bipolar disorder, with enrichment analyses of genes within the identified PAL predominantly highlighting terms related to brain morphology and function. With advancements in model optimization and uncertainty quantification, artificial neural networks have the potential to enhance the identification of genomic loci associated with complex diseases, offering a more comprehensive approach for GWAS and serving as initial screening tools for subsequent functional studies.

Problem

• Detect genome-wide association signals
• Interpret artificial neural networks
• Identify loci for complex traits

Innovation

• Artificial neural networks predict complex traits
• Feature importance scores identify associated loci
• P-values obtained for detected loci