MCP counter's problems

Question

[Neil]

Hello. I need to produce analysis on this study http://www.cbioportal.org/study?id=skcm_tcga#summary I downloaded the datasets and among then found the file

data_RNA_Seq_v2_expression_median.txt

Next I need to count infiltrating cells with R package MCP-counter. So I have RNA-expression file (on the picture)

I entered the command:

MCPcounter.estimate(expression = my_data,featuresType="HUGO_symbols")

and got the error

Warning message: In MCPcounter.estimate(expression = my_data, featuresType = "HUGO_symbols") : Found no markers for population(s): T cells, CD8 T cells, Cytotoxic lymphocytes, NK cells, B lineage, Monocytic lineage, Myeloid dendritic cells, Neutrophils, Endothelial cells, Fibroblasts

What did I wrong?

[David_emir]

hi maksjytov.nail, I got the same Warning message, did you solve this error?

[jwang293]

Hi, I am Wondering that if Rsem data download from cbioportal can be used in MCPcounter directly? Or rpkm, tpm are more recommended? Or anyone of them (RSEM, TPM, RPKM) is okay for analysis?

Answer 1

[Etienne Becht]

Hi,

Sorry about the delay I hadn't seen this post until it was signaled to me.

The warning states that there is no overlap between rownames of the input data and the signatures we defined.

Could you try following the instruction in the second post of https://github.com/ebecht/MCPcounter/issues/4 and see if it helps? Thank you

[jwang293]

Hi, I am Wondering that if Rsem data download from cbioportal can be used in MCPcounter directly? Or rpkm, tpm are more recommended? Or anyone of them (RSEM, TPM, RPKM) is okay for analysis?

Answer 2

[inspiron user]

Hi, I wonder if using TPM (transcript per million) values as input of MCP counter is not advisable, but rather KRPM input is more appropriate as the MCP counter output is more descriptive of inter-sample differences. Could you please advise? Thank you!