Using Pre-training and Interaction Modeling for ancestry-specific disease prediction in UK Biobank

TL;DR

This study explores how multiomic data, interaction modeling, and pre-training can improve disease prediction across diverse ancestries in the UK Biobank, addressing under-representation issues in genetic research.

Contribution

It demonstrates that interaction terms and pre-training can modestly enhance disease prediction accuracy for certain diseases in diverse populations.

Findings

16 models showed significant improvement in ROC-AUC scores

Interaction and pretraining models mainly relied on PRS scores

Improvements were limited to a subset of diseases

Abstract

Recent genome-wide association studies (GWAS) have uncovered the genetic basis of complex traits, but show an under-representation of non-European descent individuals, underscoring a critical gap in genetic research. Here, we assess whether we can improve disease prediction across diverse ancestries using multiomic data. We evaluate the performance of Group-LASSO INTERaction-NET (glinternet) and pretrained lasso in disease prediction focusing on diverse ancestries in the UK Biobank. Models were trained on data from White British and other ancestries and validated across a cohort of over 96,000 individuals for 8 diseases. Out of 96 models trained, we report 16 with statistically significant incremental predictive performance in terms of ROC-AUC scores (p-value < 0.05), found for diabetes, arthritis, gall stones, cystitis, asthma and osteoarthritis. For the interaction and pretrained models that outperformed the baseline, the PRS score was the primary driver behind prediction. Our findings indicate that both interaction terms and pre-training can enhance prediction accuracy but for a limited set of diseases and moderate improvements in accuracy