| Objective The aim of this study was to analyze the altematios of serum biomarkes in patients with Myelodysplastic syndrome(MDS) compared with healthy, and establish decision tree pattern of serum marker for diagnosis of MDS. And try to screen distinct proteins of different types of MDS to developed a classification algorithm named artiflcal neural network (ANN).Methods Sera were collected from 21 patients with MDS, Amony them, 6 RA, 6 RAEB, 2 RAEBt and 7 CMML samples, 6 patients with Acute myelocytic Leukemia (AML), and 12 from healthy. All of serum were added to CM10proteinchip Arrys( a weak cation exchange) and perfomed on a proteinchip PBSⅡReaders of the proteinchip Biomarker system surface-enhanced Laser desorption ionization time-of-flight mass spectrometer(SELDI-TOF-MS) manufactured by Cipergen Biosystems Inc. (Fermont, CA, USA). All date and mass spectra were analyzed using the software of Biomarker Wizard and BPS5.0(Biomarker patterns systems).Results About 362 spectras of each sample were discovered in the range of M/Z (mass/chang) values 0-50000Da(except for2 RAEBt). Among which 9 ones were significantly different between MDS and controls(p＜0.01), Five proteins of 9 with M/Z 2738 Da,2949 Da,3157 Da,3239 Da,3397 Da were up-regulated, and four ones with M/Z 6441 Da,6640 Da,8160Da,8957Da were down-regulated in the serum of MDS, two of them M/Z2738Da, 6441Da were used to build a proteomic decision tree pattern, the results yielded a predictive sensitivity of 100%and predictive specificity of 100%. Furthermore, A total of 22 distinguished proteomic peaks were detected among the surm of different types of MDS, AMLand control. 14of them were use to developed a classification algorithm.Conclusions SELDI-TOF-MS combined with proteinchip technology could greatly facilitate the discovery of biological markers of MDS and it show the diagnostic patterm established by distinguished proteomic peaks can discriminate MDS from healthy control and different types of MDS can be identified. |
PDF Full Text Request