摘要
目的建立蛋白质芯片技术检测血清蛋白质指纹图谱的方法,探讨基于人工神经网络(ANN)的血清蛋白质指纹图谱模型在先天性巨结肠(HD)患儿诊断中的应用价值。方法应用蛋白质指纹图谱分析仪测定HD患儿64例、巨结肠类缘病患儿25例和健康儿童23例血清标本的蛋白质指纹图谱,并结合ANN方法进行数据分析。112例标本随机分成训练组66例(HD40例,巨结肠类缘病14例,健康儿童12例)和盲法测试组46例(HD24例,巨结肠类缘病11例,健康儿童11例)。利用从训练组得出的基于ANN的血清蛋白质指纹图谱模型,对46例未知血清进行检测,并与X线影像学检查结果进行比较。结果筛选出质荷比(m/z)位于7211.6和2864.8的蛋白质标志物2个。在HD组表达强度分别为6.15±2.21和2.78±1.21,巨结肠类缘病组表达强度分别为12.82±7.56和4.86±0.91(Pa<0.01)。筛选出m/z位于6884.2和5639.2的蛋白质标志物2个。HD组表达强度分别为4.09±1.78和15.57±8.87,健康对照组表达强度分别为8.31±3.07和30.31±6.18(P<0.01)。应用该方法对HD患儿进行诊断的准确率、敏感度和特异度分别为89.13%(41/46例)、87.50%(21/24例)和90.91%(20/22例),明显高于X线影像学检查68.8%(77/112例)、82.8%(53/64例)和50.0%(24/48例)。结论在HD患儿的诊断中,利用从训练组得出的基于ANN的血清蛋白质指纹图谱模型较传统方法有更高的敏感性和特异性。
Objective To set up a method for the detection of serum protein finger print pattern by using the protein chip technology for exploration of serum protein finger print pattern models based on the artificial neural network （ANN）in diagnosis of Hirschsprung＇s disease （HD）. Methods One hundred and twelve serum samples from subjects with HD, allied HD, and healthy individuals were detected with protein biochip surface enhanced laser desorption / ionization time - of - flight mass spectrometry （ SELDI - TOF - MS） for protein finger print pattern,and analyzed with ANN. The 112 samples were randomly put into a training group（40 patients with HD, 14 patients with allied HD and 12 healthy individuals） and a blind test group（ 24 patients with HD, 11 patients with allied HD, and 11 healthy individuals）. The serum protein finger print patteru model obtained by ANN from training group was used to detect the 46 unknown serum samples, and the technology was compared with X - ray. Results Two protein markers whose nt/z positioned at 7 211.6 and 2 864.8 were detected. The expression inten- sity of HD was 6.15 ± 2.21 and 2.78 ± 1.21, the expression intensity of allied HD was 12.82 ± 7.56 and 4.86 ± 0.91 （ Pa 〈 0.01 ）. Two protein markers whose m/z positioned at 6 884.2 and 5 639.2 were detected. The expression intensity of HD was 4.09 ± 1.78 and 15.57 ± 8.87,the expression intensity of healthy individuals was 8.31± 3.07 and 30.31± 6.18 （ Pa 〈 0.01 ）. The accuracy, sensitivity and specificity of this method in detection of HD were 89.13% （41/46） ,87.50% （21/24） and 90.91% （20/22） respectively, significantly superior to X - ray 68.8% （77/112） ,82.8% （53/64） and 50.0% （24/48）. Conclusion In comparison with the traditional methods,this new method has higher sensitivity and specificity in diagnosis of HD.
出处
《实用儿科临床杂志》
CAS
CSCD
北大核心
2008年第17期1382-1384,共3页
Journal of Applied Clinical Pediatrics
关键词
先天性巨结肠
人工神经网络
蛋白质指纹图谱
表面增强激光解析-电离-飞行时间质谱仪
Hirschsprung＇s disease
artificial neural network
protein finger print pattern
surface enhanced laser desorption/ionization time - of - flight mass spectrometry
