Fig. 4: Pearson coefficient between the reference and imputed data by eight different techniques for different datasets (cellular taxonomy-row 1, mammalian brain-row 2, mouse intestinal epithelium-row 3, and 3D neural tissue data-4th row).
From: Leveraging data-driven self-consistency for high-fidelity gene expression recovery
The sampling efficiency (0.1–10%) to create the observed data are shown in x-axis. The center of the violin (denoted with a white circle) denotes the median value, and the spread of the violin denotes the standard deviation of the coefficient values across different cells. Source data are provided as a Source Data file.
