We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
A review of the applications and challenges of applying partial least squares (PLS) to exposomics research.
Summary
This review evaluated partial least squares (PLS) regression as a statistical tool for exposomics research, where the goal is to link complex environmental exposure mixtures to health outcomes. The authors identified key analytical advantages of PLS for high-dimensional data but noted that its full potential in exposome science remains unrealized due to methodological gaps.
BACKGROUND: Partial Least Squares (PLS) regression is a valuable tool in Environmental Health for elucidating the intricate relationships between external environmental exposures and internal biological responses linked to health outcomes in both populations and individuals. However, its full analytical potential within this domain has yet to be realized. Analysing complex exposomics data, which frequently exhibits significant multicollinearity and poses difficulties in establishing exposure-disease causality, presents a considerable analytical challenge. RESULTS: This review underscores the potential of PLS for a wide array of applications within this expanding field. It highlights the development of robust PLS-based models for in-depth exposure studies focusing on the internal exposome, single- and multi-pollutant exposures across environmental compartments, persistent and "forever" contaminants, heavy metals, emerging concerns such as microplastics, and various influential environmental and lifestyle factors, including green space, noise pollution, radiation, and dietary patterns. In contrast to conventional regression techniques often limited by multicollinearity, PLS excels at performing effective dimensionality reduction and identifying key latent variables that capture the covariance between comprehensive exposure profiles and diverse health outcomes. This inherent capability renders PLS highly applicable to both broad population-level investigations of environmental health risks and detailed personalized exposomic studies characterizing individual exposure trajectories. SIGNIFICANCE: This review critically evaluates the suitability of PLS models for these applications and offers a forward-looking vision on the future evolution and enhanced integration of PLS within exposomics to advance our understanding of disease aetiology, thereby informing the development of more targeted and effective public health interventions.