Prediction

CellCom integrates six state-of-the-art algorithms to perform the prediction. To facilitate the selection of algorithms, we provide practical guidance from two aspects: By Method Type, and By Biological Basis.

In the By Method Type section, we divide the six algorithms into two categories: expression permutation-based and network-based according to the type of prediction methods. For the expression permutation-based algorithms, CellChat, CellphoneDB, and ICELLNET consider multi-subunit structure of ligands and receptors. For the network-based algorithms, CytoTalk accounts for intracellular interactions in receiver cells that arise from CCC to improve inference.

In the By Biological Basis. section, we characterizes these algorithms by the biological basis on which they are based. CellChat and SingleCellSignalR calculate interaction scores using nonlinear modeling approaches, while other methods, such as CellPhoneDB, NATMI, and ICELLNET, simply calculate the product of ligand and receptor expressions. If users consider the modulation of interactions by cofactors (agonists and antagonists), CellChat can be a good option. If users consider not only intercellular ligand-receptor interactions, but also intracellular signaling pathways, CytoTalk can be chosen.

Please fill the next steps to run the prediction
Step1 : Upload your scRNA-seq data
Tips:

Ⅰ. Here you need to upload a .rds file which is pre-processed by the R package Seurat. See the example for details.

Ⅱ. It is recommended that the uploaded file size should not exceed 300MB.

Ⅲ. If the file is large, the running time will be long. It is recommended to fill in your email address in Step 4. We will send the complete results to your email address after the prediction is completed.

Step2: Construct database of ligand-receptor pairs
Option1
Select species
Select ligand-receptor resources
Option2
Upload your ligand-receptor pairs
Tips:

Ⅰ. You can choose to use the prior knowledge resources provided by CellCom or upload a file with ligand-receptor pairs.

Ⅱ. The uploaded file should include at least four columns: the first column named Ligand_Symbol, the second column named Receptor_Symbol, the third column named Annotation, and the fourth column named Pathway. See the example for details.

Ⅲ. The format of the uploaded file should be csv, txt, xls, or xlsx.

Step3 : Select methods to predict cell-cell communication
Tips:

Ⅰ.You can choose multiple methods by checking multiple checkboxes for predicting cell-cell communication.

Step4 : Enter your email (optional)

CSU-Li Group

鲁ICP备2021005642号-1