Skip to main content
. 2022 Jan 5;50(8):e43. doi: 10.1093/nar/gkab1275

Figure 4.

Figure 4.

The performance of scMAGIC and 13 competing methods in the third situation. Panel (A) shows the heatmaps (left) and the boxplots (right) of the accuracy and the balanced accuracy of the 14 methods. The 14 methods are arranged from top to bottom according to their mean accuracy in decreasing order. In each benchmark test, the reference and the query dataset are produced from the same tissue of mouse and human, respectively. The first authors of the reference and the query dataset and the tissue are used to represent a benchmark test. For details about the datasets used in this situation, refer to Supplementary Table S1. Here, panels (B–F) are organized in the same way as (B–F). In this benchmark test, the tissue is pancreas, while the first author of the reference and the query datasets are Baron and Muraro, respectively; ASC, activated stellate cell; QSC, quiescent stellate cell.