scimpute problem: error in read table

Question

[hsu]

I used scImpute to normalize my data in R.

code is :

scimpute(count_path = system.file("extdata","GeneCounts_3F-A37B12C16D21.csv", package ="scImpute" ), 
                    infile="csv", outfile="csv", out_dir="./", 
                    labeled = FALSE, drop_thre= 0.4, Kcluster=4, ncores =10)

These are errors:

Error in read.table(file = file, header = header, sep = sep, quote = quote,  : no lines available in input
In addition: Warning messages:
1: In dir.create(out_dir, recursive = TRUE) :cannot create dir '.', reason 'Permission denied'
2: In file(file, "rt") : file("") only supports open = "w+" and open = "w+b": using the former

what should I do about it ?

[Nitin Narwade]

First thing are you using GeneCounts_3F-A37B12C16D21.csv file from scImpute package?

Second, possible cause as reported in the warning is permission. You just have to change the permission of the folder where you had installed scImpute package (If you are using example data from scImpute). You can get the scImpute installation directory by typing system.file("extdata","GeneCounts_3F-A37B12C16D21.csv", package ="scImpute" ) on R terminal. Then if your working on linux environment just gives read, write and executable permission to the folder using chmod 777 -R /path/to/scImpute/installation/directory.

[hsu]

I am not using GeneCounts_3F-A37B12C16D21.csv file from scImpute package.

My work is on R-studio. GeneCounts_3F-A37B12C16D21.csv is my own gene-counts matrix. I am a little confused about the permission problem.

[igor]

system.file() is used to find files inside package directories. If you are using your own file, you should be using something like count_path = "/path/to/GeneCounts_3F-A37B12C16D21.csv"

For out_dir, it's a good practice to specify the full path. When you say ., what you think that means and what the tool thinks it means may not be the same.