Random Forest and Gene Ontology for functional analysis of microarray data

With the development of DNA microarray technology, scientists can now measure gene expression levels. However, such high-throughput microarray technologies produce a long list of genes with small sample size and high noisy genes. The data need to be further analysed and interpreting information on biological process requires a lot of practice and usually is a time consuming process. Most of the traditional frameworks focus on selecting small subset of genes without analysing the gene list into a useful biological knowledge. Thus, we propose a model of Random Forest and GOstats. In this research, two datasets were used which included Leukemia and Prostate. This model was capable to select a small subset of genes that were informative with relevant significant GO terms which can be used in clinical and health areas. The experimental results also validated that the subset of genes selected was functionally related to carcinogenesis or tumour histogenesis.