Figure 4: Inter-individual variation.
From: Large-scale human skin lipidomics by quantitative, high-throughput shotgun mass spectrometry
(a,b) Amounts of lipids measured of (a) males and (b) females of different age (here the two most extreme pmol values are not shown to improve clarity). Lines represent 0-order polynomial smoothing function with 8 neighbours averaged. (c,d) Profiles of lipid groups of (c) males and (d) females of different age. Lines represent 0-order polynomial smoothing function with 8 neighbours averaged. (e) Principal component analysis of males and females of different age lipidomes with lipid sub-species as input. (f) Sex prediction accuracy (test error) of random forest classification with all lipid data and data excluding sebum lipids. Points show individual results of a 5 times repeated 10 times cross validation with horizontal lines representing their mean. The dashed line is the 0.65 Null Error Rate, the prediction error was based on always choosing the majority class.
