Peak alignment using reduced set mapping (PARS) is extended with a new baseline approximation and a new dendrogram alignment scheme, which is designed to avoid the issue of selecting a target chromatogram for the alignment. Two data sets with LC/UV data are studied and it is shown that peak alignment with PARS increases the class separation substantially in the principal component score space. The results indicate that it is possible to use PARS for calibration transfer of multivariate models of chromatographic data.