Overview of SC2MeNetDrug.

(Left) Illustration of sc2MeNetDrug. (Right) The detailed data analysis pipeline of sc2MeNetDrug: (1) Raw data uploading: Users need to first upload raw-count data along with an optional design file (cell group); (2) Preprocessing: Then, the preprocessing is applied on the raw data to perform the quality control, normalization, and imputation; (3) Dimension reduction: the dimension reduction algorithm applied on the normalized data; (4) Clustering: Cluster cell sample into different group based on dimension-reduced representation for each cell sample; (5) Cell annotation: annotate each cell cluster with the best matching cell type given cell candidate and corresponding marker genes (The cell annotation result can also be uploaded along with raw data to directly perform the following analyses); (6) Cell distribution: Visualize cell type distribution for each cell group; (7) GO enrichment analysis: Gene ontology enrichment analysis to reveal the activated/inhibited GO process for selected cell type and test/control groups. (8) Proliferation and EMT: Compute proliferation and EMT score for selected cell type (mainly used for cancer dataset); (9) Ligand & Receptor communication: Identify up-regulated ligands and receptors for each cell type and potential ligand-receptor interactions between different cell types given the selected test/control groups; (10) Dysfunctional signaling pathway: Identify dysfunctional cell-cell communication and signaling pathway between two cell types given the selected test/control groups; (11) Drug discovery: Identify possible drugs to inhibit the discovered cell-cell communication network. Note that the steps of pre-analysis (1)-(5) need to be done sequentially (indicated by blue color in the figure). All downstream analyses like (6)-(11) can be performed based on the interest after sc2MeNetDrug obtain the cell annotation results (indicated by orange color in the figure).