A HYBRID PSO-GA APPROACH FOR BIOMARKER DISCOVERY

The advancement in genomic and proteomic has paved way for identifying methodologies to identify gene involved in life threatening diseases. Biomarker refers to specific gene and products with biochemical features to measure the progress of the diseases. Various soft computing techniques are applied to identify biomarkers from large micro array chips. In this paper a hybridized approach for biomarker discovery using Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) is proposed. The proposed approach is tested with microarray lymphoma dataset. Based on the experimental run twelve gene were identified as significant gene and 3 gene were found in the disease pathway which can be verified with experts to name as biomarkers using the methodology proposed. The experimental results are validated using panther tool for biological significance and verified with related literature papers. From the results it is found the performance of the algorithm is good for identifying significant gene for bio marker discovery.