Abstract: Objective To Screen serum protein markers related to CRF and establish the diagnosis model, exploring and discussing its significance in serodiagnosis by comparing differences of serum protein spectrum expression between patients with chronic renal failure（CRF） and control group.Methods Collecting 62 patients of CRF and control group with 28 normal ones.Serum samples were tested by surface enhanced laser desorption/ionization-time of flight-mass spectrometry (SELDI-TOF-MS)．The data were analyzed to screen serum proteomic biomarkers．By bioinformatics analysis ,decision classification tree models were to be established and tested.Results A total of 19 effective protein peaks were significantly different between CRF and normal control (P<0.001) at m/z range of 1500 to 30000，among which 18 showed low expression and 1 showed high expression in CRF. CRF and normal control was obviously different from the nature of the clustering; and samples of each group near each other, inter-group samples from each other. By the bioinformatics analysis , establishing a "CRF-normal controls " of the Diagnostic decision tree model, which was 87.8% in prediction accuracy rate with a sensitivity of 87.1% and a specificity of 89.3%.Conclusions Diagnostic decision tree model made more accurate judgments prediction and provided the experimental evidence for early clinical detection

Key words: Chronic renal failure, CM10 protein chip, serum