应用SELDI-TOF-MS技术诊断2型糖尿病肾病

Application of protein fingerprinting model in diagnosing type 2 diabetic nephropathy

目的比较2型糖尿病（T2DM）肾病和对照人群血清蛋白质指纹图谱的差异，建立T2DM肾病诊断模型，探讨此技术对该病诊断的价值。方法采用表面增强激光解析电离飞行时间质谱（SELDI-TOF-MS）技术检测51例T2DM肾病患者和66例对照人群血清，获得蛋白质指纹图谱。结合人工神经网络软件建立诊断模型并进行验证。结果在相对分子量2000-30000范围内共检测到175个蛋白峰，其中有17个蛋白峰明显表达差异（P〈0．01）。筛选其中质荷比（m／z）分别为5420、5782、6472、6666、10277和11770的6个蛋白峰作为标志蛋白建立人工神经网络诊断模型。利用该模型对T2DM肾病进行盲法预测，结果表明其对该病的诊断敏感性和特异性分别为81．0％和96．2％。结论利用SELDI-TOF-MS和生物信息学技术建立了敏感性和特异性均较高的T2DM肾病诊断模型，为该病诊断提供了新途径。

Objective To compare the serum protein fingerprint difference between type 2 diabetic patients with nephropathy and healthy persons and to establish diagnostic model for diagnosing type 2 diabetic nephropathy. Methods Surface enhanced laser desorption/ionization time-of-flight mass spectrometry （SELDI-TOF-MS） technology was used to analyze sera from 51 type 2 diabetic patients with nephropathy and 66 healthy controls. Diagnostic model was developed and validated using artificial neural network software. Results A total of 175 protein peaks were detected at the molecular weight ranging from 2 000 to 30 000 Dalton,of which 17 were significantly different between tvoe 2 diabetic neDhrooathv and controls （ P 〈 0.01 ）. Among them ,6 proteins （ m/z at 5 420,5 782,6 472,6 666,10277, and 11 770） were chosen to develop artificial neural network diagnostic model. The model was blindly tested with the testing set for diagnosing type 2 diabetic nephropathy. The sensitivity and specificity was 81.0% and 96.2%, respectively. Conclusion Diagnostic model with high sensitivity and specificity based on SELDI-TOF-MS and bioinformatics technology is a new approach for diagnosing patients with type 2 diabetic nephropathy.