A generalized kernel machine approach to identify higher-order composite effects in multi-view datasets
In recent years, a comprehensive study of multi-view datasets (e.g., multi-omics and imaging scans) has been a focus and forefront in biomedical research. State-of-the-art biomedical technologies are enabling us to collect multi-view biomedical datasets for the study of complex diseases. While all the views of data tend to explore complementary information of a disease, multi-view data analysis with complex interactions is challenging for a deeper and holistic understanding of biological systems. In this paper, we propose a novel generalized kernel machine approach to identify higher-order composite effects in multi-view biomedical datasets. This generalized semi-parametric (a mixed-effect linear model) approach includes the marginal and joint Hadamard product of features from different views of data. The proposed kernel machine approach considers multi-view data as predictor variables to allow more thorough and comprehensive modeling of a complex trait. The proposed method can be applied to the study of any disease model, where multi-view datasets are available. We applied our approach to both synthesized datasets and real multi-view datasets from adolescence brain development and osteoporosis study, including an imaging scan dataset and five omics datasets. Our experiments demonstrate that the proposed method can effectively identify higher-order composite effects and suggest that corresponding features (genes, region of interests, and chemical taxonomies) function in a concerted effort. We show that the proposed method is more generalizable than existing ones.