General Introduction:
In this era of genomics,the prediction of localization of various proteins is an impotant and challenging problem in the field of molecular biology. The Prediction of Subcellular localization of protein can provide an imprtant insight about the function of protein.Here, we have designed a SVM based methods for predicting the subcellular localization of the eukaryotic proteins using various features of proteins.The three features i) physicochemical properties, amino acid compostion, dipeptide compostion of proteins are taken in consideration for the development of method.The prediction accuracy of amino acid compostion,physicochemical properties and dipeptide based modules is 78.1%, 77.8% and 82.4% respectively.We have also designed a module on the PSI-BLAST.For more relaiable prediction we have constructed a hybrid module using all features of protein and PSI-BLAST information. The prediction accuracy of hybrid module based method is 88.0%.