CLASSIFYING SUBJECTS BASED ON THEIR BIOLOGICAL RESPONSE TO UV IRRADIATION

摘要

The present invention relates to methods for classifying humans into Molecular Phototypes based on the biological response of their skin to UV irradiation. The method comprises: a) providing skin cells irradiated with UV light obtained from a plurality of human subjects; b) determining the methylation levels and/or expression levels of at least about 100 features in said skin cells, wherein the features are selected from CpG sites and RNAs; c) using the methylation data and/or expression data obtained in step b) to generate a subject similarity network; d) performing cluster analysis on the subject similarity network; and e) defining Molecular Phototypes based on the clusters in the subject similarity network. The invention also relates to methods for identifying biological pathways that are associated with the response to UV irradiation of skin in human subjects belonging to a Molecular Phototype determined by a method described herein. The method comprises: a) providing skin cells irradiated with UV light and control skin cells obtained from a plurality of human subjects belonging to a Molecular Phototype determined by a method described herein; b) determining the expression levels of at least one gene in said irradiated and control skin cells, wherein the at least one gene is annotated to a biological pathway; c) using a machine-learning model trained on the data obtained in step b) to identify whether said at least one gene can discriminate between irradiated and control skin cells; and d) determining the strength of the association between the biological pathway and the response to UV irradiation of skin in the human subjects belonging to the Molecular Phototype, wherein the strength of the association is based on the degree to which the machine-learning model can discriminate between irradiated and control skin cells. Also provided are computer programs relating to the methods.