慢性肾病（CKD）的早期诊断与多组学研究

Chronic kidney disease (CKD) is a global public health challenge, with a prevalence of approximately 14.3%, affecting the health of nearly 40 million patients. The incidence of CKD is increasing due to the rise in diabetes, hypertension, and acute kidney injury. When CKD progresses to its end stage, it becomes end-stage renal disease (ESRD, or uremia), at which point patients can only be managed through renal replacement therapies such as dialysis or by awaiting kidney transplantation. The key to CKD treatment lies in early diagnosis and intervention; however, for diagnosed patients, the risk of progression depends on reliable predictive models, which can better guide personalized medical care and allocate healthcare resources. Commonly used progression risk prediction models include the Z6 model based on serum markers (34298143) and the PROGRES-CKD system that integrates renal function indicators and medical history information (34886378). However, these models lack critical molecular-level risk information. Multi-omics studies based on genomics (38909968), proteomics (25589610), transcriptomics (40305204), and metabolomics (23687356) can help elucidate disease mechanisms and facilitate the translation from laboratory to clinical settings, leading to the development of diagnostic and prognostic tools for CKD. With the development and application of machine learning algorithms, analyzing multi-omics data using these algorithms to construct diagnostic and prognostic models has become a new research direction (36857859). However, there is a lack of diagnostic and prognostic models based on CKD transcriptomics, as well as further experimental functional studies. This study aims to use machine learning algorithms to build diagnostic and prognostic models for CKD and to explore the feasibility of in-depth research on candidate biomolecules in mouse experimental models, identifying key risk molecules involved in the onset and progression of CKD.