Figure 4From: A comparison of imputation procedures and statistical tests for the analysis of two-dimensional electrophoresis data Effects of imputation on the average variance. 500 simulations were performed, where each simulation generated a dataset that contained 5%, 10%, and 20% missing-ness by randomly removing spot values from the complete dataset. Missing values were imputed by row average (Row Ave), LSM, and NIPALS or k nearest neighbors (KNN) imputation with k = 3, 5 or 8. Average variances of the complete 70 protein spots without missing data (horizontal dotted-dashed line) and after imputation are shown. The NIPALS methods are summarized by "nPR" which denotes the number of principal components used to impute the missing data.Back to article page