Fig. 1.

An illustration of the scReClassify framework. a The initial cell type annotation. This is typically achieved by using a combination of biological knowledge and computational approach. b Correction of mislabelled cells using an AdaSampling procedure. The dimensionality of the gene expression matrix is first reduced by PCA and mislabelled cells are identified and reclassified to their correct cell types by using AdaSampling with either a support vector machine (SVM) or a random forest (RF) classifier or an ensemble of SVMs or RFs