Ballgown Error in file(file, "rt") : invalid 'description' argument

Question

[vinayjrao]

Hi, I am trying out the new tuxedo protocol (hisat, stringtie and ballgown). While calling the ballgown function with bg_chrX = ballgown(dataDir = "ballgown", samplePattern = "ERR", pData=pheno_data), I get Error in file(file, "rt") : invalid 'description' argument

Any help in this regard would be appreciated.

Thanks in advance :)

[Mithil Gaikwad]

Hello Mr. Vinayrao, This error occurs when your files are not accessible

Just make sure you have set your path correctly using command setwd() and also mention the path while accessing your stringtie results:

bg_chrX = ballgown(dataDir = "path/to/folder/where/your/StrinTie/results/are/stored", samplePattern = "ERR", pData=pheno_data)

Another possibility could be the different and sophisticated name Keep the name of folders simple, and same as .gtf files have for eg., SRRXXX.gtf must be present in SRRXXX folder.

[Morris_Chair]

Hello I have a simple question, I want to reproduce this pipeline with my own data but it's not clear to me what the "ERR" in samplePattern is, I mean, is it the filepath of the single folder containing the .gtf created?

I would appreciate this clarification

thank you

[James Ashmore]

I can't find an explicit reference to a description argument in the ballgown constructor function so it must be an error from a function called within the ballgown constructor function. When you run R what is your working directory? The ballgown constructor function will look for data files in a subfolder of that directory called "ballgown" and then will look for files with the "ERR" pattern in the filename. Instead try explicitly giving the samples argument a vector of file paths to folders containing sample-specific ballgown data.

[vinayjrao]

Firstly, thanks for the reply. My working directory is the parent directory of ballgown, and as mentioned in the paper "Transcript-level expression analysis of RNA-seq experiments with HISAT, StringTie and Ballgown", the ballgown directory has a folder for each sample, within which my sample file with the ERR pattern is contained.

[Matteo Schiavinato]

Are you using a GFF file anywhere? If yes, how do mRNA or transcript lines look like, especially in the 9th field?

[vinayjrao]

I am using gtf files, whose field 9 contains gene_id ; transcript_id ; cov ; FPKM ; TPM

[Matteo Schiavinato]

I think the problem might be dataDir = "ballgown". What happens if you use the fullpath?

[vinayjrao]

I used the absolute path. It's still the same :(

[Matteo Schiavinato]

I have no idea how to solve it but googling a bit I found that Error in file(file, "rt") : invalid 'description' argument happens when a file is, for some reason, not accessible or not existing when trying to open it from R (and in other situations that are not R-related).

How is the permissions situation of the file? -rwxr--r-- ?

[vinayjrao]

The permissions to the gtf file is -rw-rw-r--

[Matteo Schiavinato]

It probably won't work, but try chmodding it to 777!

[vinayjrao]

You were right. That doesn't work either.

[Matteo Schiavinato]

Ok, then my job ends here, I can't help further ^^ However:

Let doctor Google help me, please!

Answer 1

[wood]

Do not use "samplePattern", use "samples" instead. For example

samples.c <- paste('ballgown', pheno_data$ids, sep = '/')
> samples.c
[1] "ballgown/ERR188044" "ballgown/ERR188104" "ballgown/ERR188234" "ballgown/ERR188245" "ballgown/ERR188257" "ballgown/ERR188273" "ballgown/ERR188337" "ballgown/ERR188383"
[9] "ballgown/ERR188401" "ballgown/ERR188428" "ballgown/ERR188454" "ballgown/ERR204916"
bg <- ballgown(samples = samples.c, meas='all', pData = pheno_data)