#3119 DEEP LEARNING IDENTIFIES SUBPHENOTYPES OF DIABETIC KIDNEY DISEASE DRIVEN BY GENETIC VARIATIONS IN THE RHO PATHWAY

Abstract Background and Aims Diabetic kidney disease (DKD) is a leading cause of end-stage (ESKD), however therapies targeting causal pathways have been limited by heterogeneity. Integrating electronic health record (EHR) data genomics may uncover hidden subphenotypes in DKD. In this study, we use deep learning to identify novel genetic variant ARHGEF18 associated with significantly higher risk DKD ESKD (Figure 1A). We further employed quantitative microscopy techniques biochemical assays elucidate the mechanistic role its podocytes. Method patients from Mount Sinai BioMe Biobank were used study. Unsupervised clustering accounting for population structure framework identified two clusters: Cluster M (mild) S (severe). then performed genome wide association study (GWAS) within each cluster compared healthy controls. For studies variant, cytoplasmic, focal adhesion, cytoskeletal morphometrics as well live-cell motility, Rho GTPase activity, protein degradation experiments using confocal total internal reflection fluorescence (TIRF) cell-free immortalized human podocytes expressing wild-type (WT) mutant transcripts. Results autoencoders unsupervised EHR on 1,372 establish clusters differential prevalence ESKD. There was greater proteinuria 1B). Further exome sequencing these ARHGEF18, guanine exchange factor highly enriched Nephroseq database showed an increased expression chronic (CKD) biopsy samples controls 1C). Overexpression transcripts led impairments cell adhesion architecture, motility. Live TIRF preferential subcellular localization GEF18 periphery migrating whereas WT localized at perinuclear/cytoplasmic region 2A, B). cells also displayed RhoA activation 2C). Upon inhibition synthesis cycloheximide (CHX), observed slower over 12h period indicating stability 2D). resistance ubiquitin mediated pathologically levels. Conclusions report gain function that drives podocyte dysfunction through impaired degradation. Targeting pathway could help regulate rearrangements preventing effacement