Imputation of missing gene expressions for DNA microarray using particle swarm optimization

Abstract

While capturing gene expressions using microarray technique missing values get generated in the data set. These missing values create negative impact on downstream analysis of DNA microarray. Therefore, it is necessary to estimate them before starting further analysis. Many algorithms are proposed for imputation of missing values which are based on statistical methods. They require complete gene expression data set which is created by replacing missing values by different methods like row averaging or column averaging and later missing expressions are imputed. This may affect efficiency of algorithms. In order to deal with problem of missing values, we have proposed new method based on Swarm Intelligence which is easy to implement and apply to any kind of dataset irrespective of amount of missing values in it. This method imputes missing gene expressions in microarray data set using Particle Swarm Optimization.