A non-invasive urinary diagnostic signature for diabetic kidney disease revealed by machine learning and single-cell analysis

Sep 28, 2026

Uploaded by Mabel

Diabetic kidney disease (DKD) remains a severe complication of diabetes, yet current diagnostic indicators lack sufficient sensitivity. This study integrates urinary and renal single-cell sequencing with machine learning to construct a non-invasive biomarker signature. A computational pipeline distinguishes kidney-originating urinary sediment cells, identifies injury-associated gene signatures, and refines them into a validated diagnostic model. Single-cell and transcriptomic/histological analyses support biological relevance across DKD and multiple nephropathies.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages19

Reading Time48 min

AccessFree to download

Abstract

Background

Methods

Results

Conclusions

Introduction

OPEN ACCESS Citation: Chen Y, Luo J, Zheng Y, Jiang X, Yang Z, Liu X (2026) A non-invasive urinary diagnostic signature for diabetic kidney disease revealed by machine learning and single-cell analysis. PLoS One 21(1): e0340096. [https://](https://)[ ](https://)[doi.org/10.1371/journal.pone.0340096](doi.org/10.1371/journal.pone.0340096) Editor: Yusuf Oloruntoyin Ayipo, Kwara State University, NIGERIA Received: October 20, 2025 Accepted: December 15, 2025 Published: January 2, 2026 Copyright: © 2026 Chen et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data availability statement: All relevant data are within the manuscript and its Supporting Information files. Funding: The author(s) received no specific funding for this work. RESEARCH ARTICLE A non-invasive urinary diagnostic signature for diabetic kidney disease revealed by machine learning and single-cell analysis Yonggang Chen1☯, Jintai Luo2,3☯, Yingying Zheng4☯, Xiaomei Jiang5, Zixiang Yang1, Xiaobing Liu1* 1 Department of Urology, Loudi Central Hospital, Loudi, Hunan Province, China, 2 Department of Urology and Andrology, Minimally Invasive Surgery Center, Guangdong Provincial Key Laboratory of Urology, First Affiliated Hospital of Guangzhou Medical University, Guangzhou, Guangdong, China, 3 Loudi Hospital, First Affiliated Hospital of Guangzhou Medical University, Loudi, Hunan Province, China, 4 Fangchenggang Hospital of Traditional Chinese Medicine, Guangxi University of Chinese Medicine, Fangchenggang, Guangxi, China, 5 Department of Cardiovascular Medicine, Loudi Central Hospital, Loudi, Hunan Province, China ☯ These authors contributed equally to the work. * [ldlxb216399@163.com](ldlxb216399@163.com) Abstract Background Diabetic kidney disease (DKD) poses a significant health burden with inadequate diagnostic sensitivity. This study develops non-invasive biomarkers by integrating urinary and renal single-cell sequencing with machine learning. Methods This study analyzed DKD single-cell and bulk transcriptomic data from public repositories. We established a computational pipeline to distinguish kidney-originating cells in urinary sediments, enabling the identification of injury-associated gene signatures. These signatures were refined using machine learning to develop a diagnostic model, which was validated in independent cohorts. The biomarkers were further verified in DKD renal tissues at single-cell resolution and across multiple nephropathies. Functional and spatial analyses confirmed biological relevance using transcriptomic and histological validation. Results Single-cell analysis of 2,089 urine-derived cells identified eight renal cell types, including injured proximal tubule cells (Inj-PTC) showing upregulated injury markers (HAVCR1, VCAM1) and enriched apoptotic/TGF-β pathways. A machine learningselected biomarker panel (PDK4, RHCG, FBP1) demonstrated strong diagnostic value (area under the curve, AUC >0.9), with consistent downregulation across PLOS One | [https://doi.org/10.1371/journal.pone.0340096](https://doi.org/10.1371/journal.pone.0340096) January 2, 2026 1 / 19 Competing interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Abbreviations: DKD: diabetic kidney disease; CKD: chronic kidney disease; ESRD: end-stage renal disease; eGFR: estimated glomerular filtration rate; GSVA: gene set variation analysis; GEO: Gene Expression Omnibus; KEGG: Kyoto Encyclopedia of Genes and Genomes; LASSO: least absolute shrinkage and selection operator; DEGs: Differentially expressed genes; FDR: false discovery rate; ROC: receiver operating characteristic; LR: logistic regression; XGBoost: eXtre

FAQs

What diagnostic challenge in diabetic kidney disease does this study address?

It addresses inadequate diagnostic sensitivity of current clinical markers for DKD by developing non-invasive biomarker signatures with machine learning.

How were urinary-derived cells and injury gene signatures identified?

A computational pipeline distinguished kidney-originating cells in urinary sediments, enabling identification of injury-associated gene signatures from single-cell and bulk transcriptomic data.

Which biomarker panel showed diagnostic value in the results?

A three-gene panel (PDK4, RHCG, FBP1) showed strong diagnostic performance with AUC values above 0.9, alongside injury-related pathway enrichment.

Sep 28, 2026

Uploaded by Mabel