Biomarker Discovery of Colorectal Mucinous Adenocarcinoma by Fingerprint and Support Vector Machine

Abstract: To discover special serum protein of colorectal mucinous adenocarcinoma (MA) and non-mucinous adenocarcinoma (nMA) preoperatively by using the technology of surface enhanced laser desorption/ionization-time of flight-mass spectrometry (SELDI-TOF-MS) and support vector machine. The potential tumor biomarkers in serum from 12 MA patients and 41 nMA patients were screened by the technology of SELDI-TOF-MS and CM10 Protein Chip (CipherGen Company， USA). The CM10 protein chips was analyzed by PBS II protein chip reader and the protein information was transformed into the form of spectra. The ZUCI-Protein Chip Data Analyze System software package was used to analyze the results. Discrete wavelength analysis was used to eliminate noise and subtract the baseline. A linear support vector machine (SVM) classifier was used to identify peaks. MA was compared with nMA in order to search for proteomic difference between different pathological types. The intensity of 12 proteins in the two groups was significantly different. Among them the P value of the 24 297 and 23 434 m/z were 0.0067 and 0.0092， respectively. The model formed by four protein peaks of 24 297， 3 322， 3 822 and 4 353 m/z was able to distinguish MA from nMA patients with high accuracy. Ten cases in 12 MA patients and 39 cases in 41 nMA patients were classified correctly. The accuracy was 92.45% (49/53). The specific serum protein in MA patients can be preoperatively diagnosed by SELDI-TOF-MS with high accuracy. This method possesses a potential in clinical application.
    

CSTR: 32200.14.cjcb.2008.06.0027