A New Data Imputing Algorithm
DNA microarray analysis has become the most widely used functional genomics approach in the bioinformatics field. Microarray gene expression data often contains missing values due to various reasons. Clustering gene expression data algorithms requires having complete information. This means that there should not be any missing values. In this paper, a clustering method is proposed, called "Clustering Local Least Square Imputation method (ClustLLsimpute)", to estimate the missing values. In ClustLLsimpute, a complete dataset is obtained by removing each row with missing values. K clusters and their centroids are obtained by applying a non-parametric clustering technique on the complete dataset. Similar genes to the target gene (with missing values) are chosen as the smallest Euclidian distance to the centroids of each cluster. The target gene is represented as a linear combination of similar genes. Undertaken experiments proved that this algorithm is more accurate than the other algorithms, which have been introduced in the literature.
Keywords: Missing Values, Imputation, Microarray, Regression
Download Full-Text
