Cross-Domain Text Mining of Pathophysiological Processes Associated with Diabetic Kidney Disease

Krutika Patidar,Jennifer H. Deng,Cassie S. Mitchell,Ashlee N. Ford Versypt

DOI: https://doi.org/10.1101/2024.01.10.575096

2024-01-12

Abstract:Diabetic kidney disease (DKD) remains a significant burden on the healthcare system and is the leading cause of end-stage renal disease worldwide. The pathophysiology of DKD is multifactorial and characterized by various early signs of metabolic impairment, inflammatory biomarkers, and complex pathways that lead to progressive kidney damage. New treatment prospects rely on a comprehensive understanding of disease pathology. The study aimed to identify signaling drivers and pathways that modulate glomerular endothelial dysfunction in DKD via cross-domain text mining with SemNet 2.0. The open-source literature-based discovery approach, SemNet 2.0, leverages the power of text mining 33+ million PubMed articles to provide integrative insight into multiscalar and multifactorial pathophysiology. A set of identified relevant genes and proteins that regulate different pathological events associated with DKD were analyzed and ranked using normalized mean HeteSim scores. High-ranking genes and proteins intersecting three domains—DKD, immune response, and glomerular endothelial cells—were analyzed. The top 10% of ranked concepts mapped to the following biological functions: angiotensin, apoptosis, cell-cell function, cell adhesion, chemotaxis, growth factor signaling, vascular permeability, nitric oxide response, oxidative stress, cytokine response, macrophage signaling, NFκB factor activity, TLR signaling, glucose metabolism, inflammatory response, ERK/MAPK signaling, JAK/STAT signaling, T-cell mediated response, WNT signaling, renin angiotensin system, and NADPH response. High-ranking genes and proteins were used to generate a protein-protein interaction network. This comprehensive analysis identified testable hypotheses for interactions or molecules involved with dysregulated signaling in DKD, which can be further studied through biochemical network models.

Bioinformatics

What problem does this paper attempt to address?

This paper aims to identify the pathological and physiological processes related to diabetic kidney disease (DKD) through cross-domain text mining. DKD is a major cause of end-stage renal disease worldwide, with a complex pathogenesis involving metabolic disorders, inflammatory markers, and complex pathways leading to kidney damage. The study used the open-source literature-driven discovery tool SemNet 2.0, which utilizes over 33 million PubMed articles for text mining to provide a comprehensive understanding of multi-scale and multifactorial pathophysiology.

SemNet 2.0 analyzed a range of genes and proteins related to DKD, immune response, and renal glomerular endothelial cells, ranking them based on normalized average HeteSim scores. Highly ranked genes and proteins were concentrated in three areas: DKD, immune response, and renal glomerular endothelial cells. These highly ranked concepts are associated with biological functions such as angiotensin, apoptosis, intercellular function, cell adhesion, chemotaxis, growth factor signaling, vascular permeability, nitric oxide response, oxidative stress, cytokine response, macrophage signaling, nuclear factor κB activity, TLR signaling, glucose metabolism, inflammatory response, ERK/MAPK signaling, JAK/STAT signaling, T-cell-mediated response, WNT signaling, renin-angiotensin system, and NADPH response, among others.

By analyzing the highly ranked genes and proteins, a protein-protein interaction network was generated, and testable hypotheses involving the interactions or molecules regulating signal imbalance in DKD were proposed for further investigation through biochemical network modeling. Through this approach, the study provides a comprehensive understanding of dysregulated pathways and molecules related to DKD, offering potential targets for new therapeutic strategies.