The role of protuberant data analysis in selection of certain genes having distinctive level of activities between conditions of interest i.e diseased gene and normal genes is very significant. Nowa- days it is become a standard in gene analysis that microarray of DNA is a crucial data preparation step in systemization and other biological analysis. We consider the problem of constructing an accurate prediction rule for separating the different labels of genes in microarray gene expression data. Use of SVM in such data analysis is not new but it is not up to the mark we desire. So in this manuscript, we have tried to modify Support Vector Machine (SVM) for better accuracy in cancer genes systemization. Here we have modified SVM to account for gene redundancy and keep a check on it. In the other approach, instead of keeping bias a constant in SVM, we have tried to modify SVM by bias variation which we call as Orthogonal Vertical Permutator (OVP).