Help interpreting MCPcounter output

[masaver]

Hi All,

I tested the MCPcounter implementation in immunedeconv, using the following command:
res_MCPcounter = deconvolute( expData , "mcp_counter" , feature_types = "HUGO_symbols" ) .
And you can see the results in the following plot:

I have two questions regarding the output scores:

1. As you can see, there are 11 groups/cell-types, while in the original publication have only 10. You can see that there are two groups, Macrophage/Monocyte & Monocyte , that seem to have the same values. Are they in fact the same or different from one another ?
2. One of the groups is the cytotoxicity score. Is this the same as the cytotoxic lymphpocites form the original MCPcounter publication.

Best,

-Mathias

[grst]

Hi Mathias,

1. As you can see, there are 11 groups/cell-types, while in the original publication have only 10. You can see that there are two groups, Macrophage/Monocyte & Monocyte , that seem to have the same values. Are they in fact the same or different from one another ?

yes, they are exactely the same. See also #52. In brief, the duplicate mapping is a relict from the benchmarking project, and I'll fix it in a future version.

2. One of the groups is the cytotoxicity score. Is this the same as the cytotoxic lymphpocites form the original MCPcounter publication.

Correct! To check how cell-types have been mapped, have a look at the mapping sheet of this excel file: https://github.com/icbi-lab/immunedeconv/blob/master/inst/extdata/cell_type_mapping.xlsx

Hope that helps,
Gregor

[masaver]

Thanks a lot ! That helped to answer my questions.

-Best,

Mathias