The ImmuCC server has been moved to the new site,
please click here to access our online deconvolution tool.

Frequently Asked Questions

1, NA should not appeare in the first column of Gene names. Please check the data carefully as it is difficult to see NA values especially in windows systems.

2, Duplicated gene names are not allowed, please removed before submitting your data to the server. Users should note that the gene name of microarray data is Entrez IDs, while the gene name of RNASeq data is Ensemble gene.

3, The csv and txt format transcriptomal data should be separated by "," ,"/t" or " ".

4, Number of signature genes in your sample should not be too small, because we should make sure that there are enough signature genes specific for each cell type. An error called "Signature genes is not enough" will return when shared gene number is not enough. You can view the shared signature genes between the training signature matrix and your own tissue expression profile on your own computer when this error happents.

5, Other quantification units like FPKM, TPM are not suggested for seq_ImmuCC because our training data are quantified with read counts.

6, There are something wrong for our confirmation e-mail as our previous server has been moved. If the result page URL sent to you can not open or you can not receive the email, please do not close the work page and download the result directly from this page when the task is completed.

7, As our web server was placed in the computer cluster of our institute. Sometimes the server will start to run and then stop at a few seconds when there are too many works in our computer cluster. Please try it again in latter.

8, If the sample number included in your expression matrix was too large, the task will stop in a few seconds. Please split your data into some smaller files. Usually, 30 samples will not be a problem although I have not evaluated the maximum sample number for our server.

9, If you can not access our web server from, you can open it from

10, The pipeline provided in our Github was the data processing methods used for the construction of our training signature matrix. Therefore, testing data processed under this pipeline was suggested. As I have not cevaluated the consistency across different data processing ways, I can not tell you the exact bias when applying the transcriptome data processed with other pipelines on our server. However, as far as I know, it has not much impact unless you have done some special normalization methods like scale, median centered ways.

11, Signature matrix developed for RNA-Seq data can be downloaded from the linkage here. Signature matrix for Microarray data can be downloaded from the supplementary material of our paper under the linkage here.

12, For samples profiled on other platforms (Microarray platform not listed in our selection box), you should prepare your data according to the data format required above. Personally, we have not evaluate the cross platform performance of our tool on all microarray platforms, bias may existed when training data and testing data are derived from different platforms. Please carefully study your result, if you have to analysis your data with our tool when the transcriptome technology used was not included in our server.

For the usage of this server, please read our Manual

For scripts on how to process your data, please read our Github