Skip to main content
Fig. 6 | Genome Biology

Fig. 6

From: Evaluation of some aspects in supervised cell type identification for single-cell RNA-seq: classifier, feature selection, and reference construction

Fig. 6

Computation performance of each method. The horizontal dotted red line denotes 1 and indicates a linear relation. The star denotes the p-value of the estimates (**p-value < 0.01; ***p-value < 0.001). A Regression coefficients of each method describe the relationship between training time and reference data size. As shown in the figure, the training time of SVM and random forest grows faster than the increase of reference data size, and all others are slower. Among all classifiers, the coefficient estimation of GEDFN is not significant. B Regression coefficients of each method describe the relationship between training time and number of cell types. The training time of GEDFN and SVM with linear kernel grows faster than the increase of the number of cell types. Coefficient estimations of scmap and MARS are not significant

Back to article page