This case study walks through using sparse PLS (sPLS) to explore relationships between gene expression and clinical measurements in a liver toxicity study. sPLS helps uncover correlated patterns across datasets and identify key variables driving these links. Here, we use the multivariate (PLS2) approach to tune the model, interpret sample groupings, and explore networks between genes and clinical traits.
🔍 More on sPLS
🔍 PLS1 case study
📄 Download R script
Data used on this page:liver.toxicity
Key functions used on this page:pca()
spls()
perf()
tune()plotIndiv()
plotArrow()plotVar()
network()cim()
References:
1. Tenenhaus M. (1998) La régression PLS: théorie et pratique. Paris: Editions Technic.
2. Wold, S., Sjöström, M., and Eriksson, L. (2001). Pls-regression: a basic tool of chemometrics. Chemometrics and intelligent laboratory systems, 58(2), 109–130.
3. Bushel, P., Heinloth, A., Li, J., Huang, L., Chou, J., & Boorman, G. et al. (2007). Blood gene expression signatures predict exposure levels. Proceedings Of The National Academy Of Sciences, 104(46), 18211-18216. https://doi.org/10.1073/pnas.0706987104