Imputation of missing values in DNA microarray gene expression data

Abstract

Most multivariate statistical methods for gene expression data require a complete matrix of gene array values. In this paper, a imputation method based on least squares formulation is proposed to estimate missing values. It exploits local similarity structures in the data as well as least squares optimization process. The proposed local least squares imputation method (LLSimpute) represents a target gene that has missing values as a linear combination of similar genes. This algorithm showed better performance than the other imputation methods such as k-nearest neighbor imputation and an imputation method base on Bayesian principal component analysis.