Analyses from paper

Analysis code from our [gemmr] publication is collected here. At the same time, these analyses demonstrate use cases of gemmr and show properties of CCA and PLS:

• Samples per feature dependence
• Samples per feature is a key parameter
• Cross-validated estimation
• Compatible true vs observed associations
• Weight errors
• Empirical data: evaluate analyses
• How many PCs are really necessary?
• Analysis of prior literature
• Required sample sizes
• Sample size calculator
• Sample size dependence on powerlaw-decay constants
• Loadings and Cross-loadings
• Sparse PLS
• CCA vs PLS errors