Skip to main content
Fig. 6 | BMC Bioinformatics

Fig. 6

From: Distribution based nearest neighbor imputation for truncated high dimensional data with applications to pre-clinical and clinical metabolomics studies

Fig. 6

Comparison of the true missing values with missing values imputed from the three methods based on a single simulated dataset (N = 50 X M = 400). The values for the first 20 metabolites are shown. The x-axis represents the metabolites, and the y-axis represents the intensity values. The open black circles represent observed values, while the black stars represent missing observations. Blue triangles, red squares, and green diamonds represent missing values imputed by KNN-TN, KNN-CR and KNN-EU, respectively. The region below the LOD is shaded in light red. In most cases, the KNN-TN algorithm is able to impute missing values below the LOD better than the other two methods (e.g., metabolites 1, 3, 4, 7, 8, 12, and 13). In other cases, the KNN-TN imputations are similar to KNN-CR (e.g., for metabolite 5, for which the missing below the LOD was too high and the NR algorithm was unable to converge)

Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us