error in fun(x((i)), ...) : only defined on a data frame with all numeric alike variables

Question

[Azra]

Hello everyone,

I'm new to Rstudio, and i'm a little bit stuck. I'm trying to run the code of cibersort for the deconvolution of RNAseq samples using the LM22 signature matrix provided. I did a previous differential analysis with DESeq2, and used the normalized matrix of my analysis to run the cibersort script. here is my code:

if (!require(CIBERSORT))devtools::install_github("Moonerss/CIBERSORT") library(CIBERSORT) data(LM22) mydata<-read.table("C:/Users/dell/Desktop/cibersort rstudio/normdata radiosensible vs radioresistantgenes.txt", header = TRUE, sep= '\t') myresults<-cibersort(sig_matrix = LM22, mixture_file = mydata)

Also, here is a snippet of my data: mydata[1:5,1:5]

I keep getting this error: error in fun(x((i)), ...) : only defined on a data frame with all numeric alike variables

Can anyone help me? Thanks in advance for your replies.

Answer 1

[ATpoint]

To the best of my knowledge, CIBERSORT is and has always and only been available via the official website application from Stanford. Whatever this GitHub is, it is (by best knowledge) neither official, hence not reliable. I suggest you abandon it.

[Mathew]

I looked up a publication using CIBERSORT to see how they got the package, and indeed they acquired it through official website application from Stanford.

Here is the publication: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5895181/

Relevant methods section:

3.1 Installation

CIBERSORT can be run online (http://cibersort.stanford.edu/) or downloaded for local use, and is freely available for academic non-profit research. R and Java implementations are available for download. While the current R script can be used to run the CIBERSORT deconvolution engine, users wishing to create a custom signature matrix will need to use the website or the Java executable.

I would agree to avoid this GitHub version, and apply through the official website application through Stanford.. if you are doing analysis for research purposes and use this GitHub version, a reviewer might be extremely skeptical.

[Azra]

I see, thank you!