Abstract
Renal fibrosis is a common pathological pathway of progressive chronic kidney disease (CKD). However, kidney function parameters are suboptimal for detecting early fibrosis, and therefore, novel biomarkers are urgently needed. We designed a 2-stage study and constructed a targeted microarray to detect urinary mRNAs of CKD patients with renal biopsy and healthy participants. We analysed the microarray data by an iterative random forest method to select candidate biomarkers and produce a more accurate classifier of renal fibrosis. Seventy-six and 49 participants were enrolled into stage I and stage II studies, respectively. By the iterative random forest method, we identified a four-mRNA signature in urinary sediment, including TGFβ1, MMP9, TIMP2, and vimentin, as important features of tubulointerstitial fibrosis (TIF). All four mRNAs significantly correlated with TIF scores and discriminated TIF with high sensitivity, which was further validated in the stage-II study. The combined classifiers showed excellent sensitivity and outperformed serum creatinine and estimated glomerular filtration rate measurements in diagnosing TIF. Another four mRNAs significantly correlated with glomerulosclerosis. These findings showed that urinary mRNAs can serve as sensitive biomarkers of renal fibrosis, and the random forest classifier containing urinary mRNAs showed favourable performance in diagnosing early renal fibrosis.
Highlights
Using targeted microarrays, our previous study showed that urinary vimentin mRNA was significantly upregulated in moderate-to-severe fibrosis[9]
We demonstrated that urinary podocalyxin, CD2-AP, α-actin[4], and podocin mRNAs correlated with SCr in patients with diabetic nephropathy (DN)[8]
By iterative random forest analysis of a targeted microarray, we aimed to discover a panel of mRNAs and develop a more powerful classifier for improved diagnosis of renal fibrosis
Summary
Our previous study showed that urinary vimentin mRNA was significantly upregulated in moderate-to-severe fibrosis[9]. Whether urinary mRNAs can efficiently identify patients with renal fibrosis in the context of CKD has not been investigated yet. Another problem is that datasets generated by microarrays are often noisy, multicollinear, and high dimensional, which make it difficult to process. The primary objective of this two-stage study was to test and validate the hypothesis that mRNAs from urinary sediment could provide useful information of early renal fibrosis. To our knowledge, this is the first machine-learning analysis of the diagnostic performance of urinary mRNAs in renal fibrosis. By iterative random forest analysis of a targeted microarray, we aimed to discover a panel of mRNAs and develop a more powerful classifier for improved diagnosis of renal fibrosis
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
