Random sampling or 'random' model in skin flux measurements? [Commentary on 'Investigation of the mechanism of flux across human skin in vitro by quantitative structure-permeability relationships']

Transdermal therapy receives increasing attention as an attractive alternative to traditional drug delivery. Unfortunately the exact algorithm of transdermal permeation that could guide medicinal chemists towards delivery optimization at an early stage of the drug design process still remains to be decoded. This paper discusses some major hurdles on the way to full understanding of Quantitative Structure-Activity Relationships (QSAR) of skin permeation. From the statistical perspective, a recently published combined data set is found to be inappropriate with respect to the distribution of major molecular descriptors, and therefore should be approached cautiously as a source for QSAR model training and in modelling of occupational and environmental skin exposures.