Rai Lavanya, Kumar Pratap, Mahato Krishna K, Kartha Vasudevan B, Santhosh Chidangil
Manipal University, Centre for Atomic and Molecular Physics, Manipal, India.
J Biomed Opt. 2008 Sep-Oct;13(5):054062. doi: 10.1117/1.2992166.
High performance liquid chromatography with high sensitivity laser-induced fluorescence detection is used to study the protein profiles of serum samples from healthy volunteers and cervical cancer subjects. The protein profiles are subjected to principal component analysis (PCA). PCA shows that the large number of chromatograms of a given class of serum samples--say normal/malignant--can be expressed in terms of a small number of factors (principal components). Three parameters--scores of the factors, squared residuals, and Mahalanobis distance--are derived from PCA. The parameters are observed to have a narrow range for protein profiles of standard calibration sets formed from groups of clinically confirmed normal/malignant classes. Limit tests using match/no match of the parameters of any test sample with parameters derived for the standard calibration sets give very good discrimination between malignant and normal samples with high sensitivity (approximately 100%) aand specificity (approximately 94%).
采用具有高灵敏度激光诱导荧光检测的高效液相色谱法,研究健康志愿者和宫颈癌患者血清样本的蛋白质谱。对蛋白质谱进行主成分分析(PCA)。PCA表明,给定类别的血清样本(如正常/恶性)的大量色谱图可以用少量因子(主成分)来表示。从PCA得出三个参数——因子得分、平方残差和马氏距离。观察发现,由临床确诊的正常/恶性类别组形成的标准校准集的蛋白质谱,这些参数具有较窄的范围。使用任何测试样本的参数与为标准校准集导出的参数进行匹配/不匹配的限度测试,在高灵敏度(约100%)和特异性(约94%)下,能对恶性和正常样本进行很好的区分。