| Proteomics, a new advent branch of science, was coming into being on the basis of human genome project. The aim of proteomics was not only to unveil the mystery of life, but also to provide substantial foundation for disease mechanism and diagnostics. The study of proteomics was from the end of last century abroad and from 2001 domestically and the content of which focused mainly on the study of serious diseases of human beings, new biomarkers of diseases and so on.After the invention of Surface Enhanced Laser Desorption Ionization-Time of Flight Mass Spectrometry (SELDI-TOF MS ) , some biomarkers of caner such as ovarian cancer, galactophore cancer, lung cancer, prostate cancer and so on were validated partly. But the SELDI-TOF MS data of cancer serum which were rich of biological and medicinal information needs to be explored to elucidate the pathological mechanism.Ovarian cancer is worldwide malignancy which imperils women's health. It presents at a late clinical stage in more than 80% of patients, and is associated with a 5-year survival of 35% in this population. By contrast, the 5-year survival for patients with stage I ovarian cancer exceeds 90%. Because the ovary is in the inside of women's body, there was no obvious symptom in early stage and there were no effective diagnose methods till now, it is crucial to diagnose the disease in early stage.In this context, the ovarian cancer data of SELDI-TOF MS spectrum afforded by scientists from FDA and NCI was studied. After the feature selection by statistical test and GA-PLS, 10 feature m/z values and the corresponding feature matrix were obtained. The 10 m/z values were all below 500, the region of which was called low molecular mass (LMM) range. This region, rich of important information, was poorly studied now unfortunately.The selected feature matrix (253x10) was studied by various pattern recognition classification methods including non-linear methods like Artificial Neural Networks and Support Vector Machines, linear methods such as Bayes Discrimination Analysis, Fisher Linear Discrimination, K-nearest neighbor classification method. Methods of BDA, KNN, ANN and SVM could reach 100% when applied to the selected feature matrix (253×10). The results indicated that it is linear classified and the linear classifier can satisfy the classification requirement of this data.The pattern constituted by the 10 feature m/z values was well classified and if we could validate the m/z values from the view of biology and medicine, maybe we could find new ovarian cancer biomarkers and to enlighten the mechanism of ovarian cancer.SELDI-TOF MS based serum proteomics, is a high throughput, complete, and dynamical methodology when applied to study change of protein molecule with particular superiority in the finding early biomarkers of cancer. Through mining its information, we could understand the signal transduction, apoptosis, infiltration and resistance in chemical therapy and find new measure to early diagnostic, treatment and prognosis of cancers.The ovarian cancer data of SELDI-TOF MS spectrum were studied in this paper and some important features were selected by proposed method. It has very important sense for the study of the mechanism of ovarian cancer if we could find the detailed protein information.
|
PDF Full Text Request