Potential of classification of cancer by multiple discriminant analysis for relationship between cancer and...

This is the first report to describe the potential for classification of cancer using anti-phosphoprotein monoclonal antibodies (PPmAbs) and multiple discriminant analysis. Over 150 hybridoma clones producing monoclonal antibodies were generated against a human phosphoprotein mixture derived from a human leukemia cell line. The expression profiles of 22 cell lines from 9 different types of cancer using PPmAbs were examined. The relationship between cancer cells and the expression of human phosphoprotein in the cells was analyzed by multiple discriminant analysis and was used to construct a diagnostic system for cancers. Multiple discriminant analysis was able to successfully classify the cell lines into the correct cancer group by using the diagnostic system for cancers. These results show that multiple discriminant analysis based on phosphoprotein expression in cells or tissues may be a potentially valuable method for assisting in the classification of several cancers.