Cell-cell communication (CCC) across multiple cell types and tissues orchestrates organismal development, homeostasis and single-cell functions. The signaling events behind CCC, mostly mediated by ligand-receptor pairs, create a highly connected signaling network and regulate the biological behavior of cells within their communities. Numerous computational tools have been developed to infer CCC from Single-cell RNA-seq at both the individual cell and cell cluster levels(1-2).

We developed CellCom for prediction, visualization, and evaluation of cell-cell communication mediated by ligand-receptor pairs. CellCom integrated rich prior knowledge resources and 6 state-of-the-art methods (CellChat(3), CellPhoneDB(4), SingleCellSignalR(5), NATMI(6), ICELLNET(7), and CytoTalk(8)) for predicting cell communications, supporting user-defined Combination of methods and resources. Furthermore, CellCom proposed a Visualization pipeline for users to comprehensively analyze intercellular communication networks from five perspectives: interaction, expression, communication, contribution, and downstream. CellCom also offered two Evaluation programs to compare different methods from the performance metrics and display the similarities and differences between these methods.

CellCom collected and integrated 24 databases with different focuses to build a prior knowledge database of 71,381 ligand-receptor pairs covering ten species. Additionally, we enriched prior knowledge resources by integrating information on multi-subunit complexes, pathways, and so on.