proteomic cluster analysis of malignant gliomas in humans

Background: gliomas are the most frequently observed primary brain tumors. thesetumors comprise a variety of different histological tumor types and malignancy grades.oligodendrogliomas typically contain a rich network of branching capillaries.approximately 50%-80% of oligodendrogliomas demonstrate a combined loss ofchromosomes 1p and 19q. oligodendrogliomas differ from neurocytomas in that theyshow a diffusely infiltrating pattern of spread that precludes surgical cure.methods: we evaluated extracted proteins from tumors and normal brain tissuesfor protein purity by the bradford test and spectrophotometry. we separated proteinsby two-dimensional gel electrophoresis. the spots were analyzed and compared usingstatistical data and maldi-tof/tof. protein clustering analyses were performed onthe list of proteins deemed significantly altered in oligodendroglioma tumor tissues.results: on each analytical two-dimensional gel, we observed an average of 1328spots. a total of 157 exhibited up-regulation of expression levels, whereas the remaining276 spots had decreased expression in astrocytoma tumors relative to normal tissue.the results demonstrated that functional clustering and principal component analysishad considerable merit in aiding the interpretation of proteomic data.conclusion: clustering methodology is a powerful data mining approach forinitial exploration of proteomic data. the clustering results depend on parameters suchas data preprocessing, between-profile similarity measurement and the dendrogramconstruction procedure.