SCREEN: predicting single-cell gene expression perturbation responses via optimal transport

In this study, we propose SCREEN, a novel method for predicting perturbation responses of scRNA-seq data. Through extensive experiments on various datasets, we validated the effectiveness and advantages of SCREEN for the prediction of single-cell gene expression perturbation responses. Besides, we demonstrated the ability of SCREEN to facilitate biological implications in downstream analysis. Moreover, we showed the robustness of SCREEN to noise degree, number of cell types, and cell type imbalance, indicating its broader applicability. Source codes and detailed tutorials of SCREEN are freely accessible at Github (Califorya/SCREEN). We anticipate SCREEN will greatly assist with perturbational single-cell omics and precision medicine.