Identifying hidden drivers of heterogeneous inflammatory diseases

Chronic inflammatory diseases of the cardiovascular system, brain, gut, joints, skin and lung are characterized by complex interactions between genetic predisposition and tissuespecific immune responses. This heterogeneity complicates diagnoses and the ability to exploit omics approaches to improve disease management, develop more effective therapeutics, and apply precision medicine. Using skin inflammation as a model, we developed a bio-computational approach that assigns deep clinical phenotyping information to transcriptome data of lesional and non-lesional skin (564 samples) to identify biologically-relevant gene signatures. This identified previously unknown key factors, including CCAAT Enhancer-Binding Protein Beta (CEBPB) in neutrophil invasion, and Pituitary Tumor-Transforming 2 (PTTG2) in the pathogenic epithelial response to inflammation. These were validated using genetically-modified human skin equivalents, migration assays, and in situ imaging. Thus, by combining deep clinical phenotyping and omics data with sophisticated bio-computational algorithms we present a methodological advance to identify hidden drivers of clinically-relevant biological processes within omics datasets. One Sentence SummaryDeciphering the pathogenesis of chronic inflammatory diseases by assigning transcriptome profiles to deep clinical phenotyping.