Figure 3.

Nonlinear alignment by support vector regression. Support vector regression (SVR) was evaluated here on the retention time dimension between 2 illustrative samples described by a nonlinear, “S-shaped” relationship in retention time. To model this relationship, a radial basis function (RBF) kernel was selected. Measurements between samples varied negligibly in drift time and m/z, and thus alignment was only necessary in retention time. An example involving samples with a linear relationship in retention time is included in Figure S7.