How to tell the CEMiTool that the input file is an expression matrix? It treats gene names as part of the matrix and gives an error that it only accepts numeric data

, ,
1

I try to use CEMiTool for gene coexpression analysis. the tool works perfectly if I removed the (first column= row names = gene names) from the input expression matrix. because it can only deal with numeric data and nor character.

image
image950×396 22.8 KB

But doing so will result in an error if I used the optional parameter of the CEMiTool which is (Pathway list) to perform pathway enrichment analysis, because the tool will not find genes in the expression matrix to relate to the pathway list.

image
image1442×344 37.1 KB

So How to tell the tool that the first column is the gene names? the tool only accepts tabular format. .mtx files are not accepted

2

Welcome, @Menna_Youssef

I can reproduce the problem, and am exploring a solution. I’ll get back to you here today. Thanks for sharing the history – very helpful! :slight_smile:

3

Hi again @Menna_Youssef

So … the expression matrix input has a special characteristic

  • The header only includes enough values to equal the number of samples.
  • Then the sample lines have an extra value with the gene identifier

An example is here: https://usegalaxy.eu/u/jenj/h/test-cemitool

Note that the example had quotes around the values, and that is needed when using the tool directly in R but it doesn’t appear to be required in Galaxy (due to the way the data is parsed). But confirm this with your own data and add in quotes as needed.

For the other files, examples are in the test directory here → tool form link to ToolShed (upper right corner in the Options menu) Galaxy | Tool Shed → link to Development repository tools-iuc/tools/cemitool at main · galaxyproject/tools-iuc · GitHub → then navigate into the test-data directory.

This works for all tools if you ever get stuck again. You can copy the raw data link into the Upload tool to get the files into Galaxy to play around with them, compare to your own data, and similar.

Hope this helps! :scientist:

4

That’s so helpful, thanks a lot!

1 Like