Combination of partial least squares regression and design of experiments to model the retention of pharmaceutical compounds in supercritical fluid chromatography.

•Accurate retention prediction of new compounds was performed with the PLS model built.•Pharmaceutical compounds served as retention’s probes in experiments defined by DoE.•Fit individual MLR models showed the effects of experimental parameters on retention.•PLSR, DoE & ionic LSER descriptors allowed establishment of a global retention models.