PROJECT SUMMARY The American Academy of Dermatology reports that 1 in 4 Americans (~84.5 million) are impacted by skin disease. Skin disease is the fourth leading cause of disability worldwide, significantly impacts quality of life, and costs ~$75 billion annually to treat. Skin conditions like atopic dermatitis (AD) are commonly diagnosed by practitioners using clinical history and physical exam features; however, because of limited understanding of the diverse pathophysiological mechanisms that underlie complex skin lesions, disease management still follows a ‘one-size-fits-all’ paradigm. This lack of evidence-based personalization or precision medicine leads to poor treatment outcomes and patient frustration. The central objective of this proposal is the development of a molecular-level skin assessment platform that will allow evidence-based diagnosis of skin conditions as well as the delivery of supplementary information on the pathophysiological mechanisms of the disease state to aid practitioners in choosing treatments and monitoring treatment progress. The final product skin assessment platform includes: 1) a standardized sample collection kit which allows for easy, non-invasive collection of material from a patient’s stratum corneum via tape-stripping, and 2) a pipeline to elucidate biomarker data consisting of liquid chromatography-mass spectrometry (LC-MS/MS) analysis and big data artificial intelligence approaches (i.e., deep neural networks, etc.). The test can be shipped through the mail and completed at home, allowing for the technology to be used for remote dermatological care and expanding access to groups historically underserved. Successful completion of Phase I will provide proof-of-principle of using skin biomarkers for prediction of atopic dermatitis in samples collected at-home. In Aim 1, we will validate our sample collection process to verity the robustness of at-home sample collection. In a study of 25 individuals, we will assess the quality of data obtained from untrained (at-home) sample collection versus trained (in-office) sample collection through assessing the protein content and similarity of compounds detected between these samples. In Aim 2, we will identify predictive biomarkers of AD in a study of 75 healthy (control) and 75 individuals (patients) diagnosed with AD. Feature selection and machine learning prediction analysis will be used to determine small molecule biomarkers associated with AD, and success will be measured as 90% predictive ability (area under curve (AUC) ≥ 0.90) of the biomarker set on an isolated cross validation dataset. These studies will demonstrate proof of concept and prove product feasibility through the identification of diagnostic, monitoring and predictive skin biomarkers associated with AD and AD therapy, provide critical analytical validation of the at-home sample collection kit by users, and increase the success of a future Phase II program focused on the clinical validation for t...