A method of formulating a skin-lightening composition by identifying actives with a data architecture and incorporating the actives into a skin care composition. The data architecture can be used to identify connections between perturbagens and genes associated with skin tone. A gene expression profile for a control human cell is compared to the gene expression profile a cell exposed to perturbagen to identify the genes which are differentially expressed. The differentially expressed genes are analyzed and arranged into an ordered list of identifiers, which along with other instances create a data architecture.