573-P: Construction and Validation of a Hierarchical-Driven Network for Decoding the Potential Molecular Mechanisms in Diabetic Microvascular Complications

Aim: Diabetic microvascular complications, including diabetic nephropathy (DN) and diabetic retinopathy (DR), lead to poor life quality and increased mortality. Accumulating evidence indicates that the complexity of pathophysiology is controlled by hierarchical molecular factors at the epigenetic, transcriptional and post-transcriptional levels. Yet most studies focus on one of the driven factors, it still lacks a systematic classification model that links hierarchical molecular factors to the precise diagnosis of diabetic microangiopathy. This study aimed to construct driven networks that reveal the potential molecular mechanisms for the precise diagnosis of diabetic microangiopathy.

Methods: By combining multi-regulators, transcriptional kinetics, gene expression, DNA-protein binding signals and lncRNA targets, we based on the hierarchical Bayesian Model, DrivenRN, to construct driven networks for detecting the signature of key function genes. Then we tested this model using public datasets, followed by real-time PCR validation of the identified genes’ expression in the peripheral blood mononuclear cells (PBMCs) from patients.

Results: Through DrivenRN analysis, we identified four genes that are specifically involved in driving DN, including TIMP3, CNDP1, MYH9 and ENPP1. We also identified three genes that are specifically involved in driving DR-, including EDN1, CD34 and CYR61. Furthermore, we validated CNDP1, MYH9 and ENPP1 in DN PBMCs, and CD34 in DR PBMCs.

Conclusion: This study has constructed a novel model, DrivenRN, which effectively maps the driven networks for the prediction of potential pathological factors in diabetic microvascular complications.