1 Spatial quantification and classification of skin response follow- ing perturbation using organotypic skin cultures

Motivation: For a mechanistic understanding of skin and its re-sponse to an induced perturbation, systems biology is gaining in-creasing attention. Unfortunately, quantitative and spatial expression data for skin, like for most other tissues, are almost not available. Results: Integrating organotypic skin cultures, whole-slide-scanning and subsequent image processing provides bioinformatics with a novel source of spatial expression data. We here used this approach to quantitatively describe the effect of treating organotypic skin cul-tures with SDS in a non-corrosive concentration. We firstly meas-ured the differentiation related spatial expression gradient of Heat-Shock-Protein 27 in a time series of up to 24h. Secondly, a multi-dimensional tissue classifier for predicting skin irritation was devel-oped based on abstract features of these profiles. We obtained a high specificity of 0.94 and a sensitivity of 0.92 compared to manual classification. Our results demonstrate that the integration of tissue cultures, whole-slide-scanning and image processing is well suited for both, the standardized data acquisition for systems biological tissue models and a highly robust classification of tissue responses. Contact