Classification of Human Cancer Diseases Gene Expression Profiles using Genetic Algorithm by Integrating Protein Protein Interactions Along with Gene Expression Profiles

The classifications of different cancer types are most important in cancer diagnosis. The several researches are employed to diagnosis the cancer disease. In the existing research work, the classification of cancer disease is performed based on gene expression data. And it is initially presented Information Gain (IG) for Feature selection and Genetic Algorithm (GA) is used for feature reduction to remove the irrelevant features and noise data present in the given data set. Iteratively the Genetic Algorithm (GA) is classified the types of human cancer. The existing approach has some limitation to classify the cancer disease such as high computation time and required large amount of computational resource. In order to overcome this limitation, the integrated gene expression data with Protein- Protein interaction (IGEPPIN) is proposed for predicting cancer disease by prioritizing disease candidate proteins to rank each protein. In this approach initially find the beginning proteins of human and constructed the PPIN of seed proteins from human PPI data. Second, combine the function similarity of the gene expression profiles and protein to protein interactions confidence for calculating weight of each interacting protein for classification of human cancer disease. Subsequently, measure the disease relevance score for each protein by adding the interaction confidence, functional similarity of its neighbours. Finally calculated the weight of proteins in the specific disease and rank them based on descending order of their scores. The proposed system is estimated based on classification Accuracy, Precision, Recall, F- Measure and Specificity.