ASN's Mission

To create a world without kidney diseases, the ASN Alliance for Kidney Health elevates care by educating and informing, driving breakthroughs and innovation, and advocating for policies that create transformative changes in kidney medicine throughout the world.

learn more

Contact ASN

1401 H St, NW, Ste 900, Washington, DC 20005

email@asn-online.org

202-640-4660

The Latest on X

Kidney Week

Abstract: SA-PO0009

Molecular Taxonomy of Urinary Diseases Through Unsupervised Data-Driven Classification Based on 7000-Plex Urinary Proteome

Session Information

Category: Artificial Intelligence, Digital Health, and Data Science

  • 300 Artificial Intelligence, Digital Health, and Data Science

Authors

  • Maruvada, Vinaika, University of Houston, Houston, Texas, United States
  • Ma, Yewei, University of Houston, Houston, Texas, United States
  • Vanarsa, Kamala, University of Houston, Houston, Texas, United States
  • Saxena, Ramesh, The University of Texas Southwestern Medical Center, Dallas, Texas, United States
  • Mohan, Chandra, University of Houston, Houston, Texas, United States
Background

It is not clear if one can ascertain what renal disease a subject has based just on urine proteomics. Here, we attempt to define the molecular taxonomy of renal diseases using an unsupervised approach based on high-dimensional urinary proteomics.

Methods

Urine from 85 subjects—healthy controls (HC), lupus nephritis (LN), diabetic chronic kidney disease (DCKD), urology clinic controls (UC, presenting with hematuria), and bladder cancer (BC) were profiled using a 7000-plex aptamer-based proteomic platform. Following preprocessing, a protein–protein correlation network was constructed and partitioned into modules using Louvain community detection. Each patient’s expression was summarized by the mean expression of each module, creating a matrix of module activity scores. Hierarchical clustering using Ward’s method was applied to these module profiles to stratify patients.

Results

12 discrete protein modules were identified. Hierarchical clustering of patient module activity revealed 16 distinct patient groups (G1–G16). The resulting clusters (G4,G5 and G11) partially or fully aligned with clinical diagnoses. While some groups were diagnosis-specific, others included mixed categories. The heatmap patterns highlighted group-specific activation of modules like 2, 4, 8, and 9, underscoring potential biological stratification not evident in traditional classifications.

Conclusion

Unsupervised clustering revealed distinct molecular subtypes of renal disease that do not always match traditional diagnoses, but may offer an alternate or more accurate way to classify, monitor, and eventually personalize treatment for patients with renal diseases. Analysis is underway to understand the underlying biology of this classification and any confounding variables.

Protein Module Detection and Patient Stratification via Unsupervised Clustering of Urinary Proteomics Data

Digital Object Identifier (DOI)