Fig. 1
From: DrImpute: imputing dropout events in single cell RNA sequencing data
DrImpute has significantly better performance on discriminating dropout zeros from true zeros than existing methods. (a) Overview of DrImpute pipeline: (1) data cleansing, normalization, and log transformation; (2) calculating the distance matrix among cells; (3) imputing the dropout entries based on the clustering results; and (4) averaging all imputation results to determine the final imputation. b-c Three scRNA-seq imputation algorithms DrImpute, scImpute and MAGIC were used to discriminating the dropout zeros from the true zeros in the simulation studies. The full scRNA-seq datasets from (b) Pollen et al. and (c) Usoskin et al. were down-sampled at 10, 15%, 25, 40 and 63% of the total number of reads. The discriminative performance was measured by F1 score (the harmonic mean of precision and recall)