| Popis: |
Today, it is possible to monitor a gene expression on a genomic scale using hierarchical clustering, DNA micro-arrays and k-means partitioning which are being the most popular methods. Several tools make use of the GO ontologies or the gene associations provided by consortium members or even individuals. While some progress has been made in addressing the gene classification, current methods are restricted by the limitations of the clustering and visualizations techniques. For example, Avadis, BiNGOb and DAVID tools are based on visualization for gene expression data. In visualization, gene annotations are visualized in as a table view and so the granularity of the GO DAG can be viewed freely by the user or use CLASSIFI (Cluster Assignment for Biological Inference) which is a data-mining tool that can be used to identify significant co-clustering of genes with similar functional properties such as cellular response to DNA damage. Furthermore, Current research is generally more concerned with the clustering and visualizations techniques for gene expression data analysis. To enhance the bioinformatics, many researchers and technicians have preferred to match the clustering to the specifications of biomedical applications. In this papered, we have reviewed a number of clustering algorithms for different approaches and data types. In addition, a proposed solution is presented. The objective of the expected solution is to predict various diseases that could be occurred. Our Solution idea is to study the correlations between the genes in the same classes and between the different classes. The results illustrated that the proposed solution is fine in terms of accuracy and performance. However, the features and parameters need to be developed further. |
