SKCM MAF from GDC can not be read about maftools

[tiagochst]

You might be able to use TCGAbilinks to download the data and use with maftools, an example is in the bottom page.
https://www.bioconductor.org/packages/devel/bioc/vignettes/TCGAbiolinks/inst/doc/mutation.html

library(TCGAbiolinks)
library(maftools)
maf <- GDCquery_Maf("SKCM", pipelines = "muse") %>% read.maf(removeSilent = TRUE, useAll = FALSE)

[tiagochst]

@PoisonAlien I'm not sure you were going to create a GUI, but I just added your tool to TCGAbiolinksGUI
Your package is awesome, nice work!

Youtube Video

[PoisonAlien]

Hi,

I just downloaded SKCM from GDC (I believe this is the one : https://gdc-portal.nci.nih.gov/files/49540f53-3228-4a65-8854-ebd2f4a8f5ce ) and it works fine on my Mac. However I guess you're on Windows and the input file is gzipped ? Can you uncompress gz file to plain text and try again ?

[rolfhaut]

Hi,

That was indeed the problem.
Thanks
David

[krcoombes]

I'm not convinced that this issue should be closed without changing either the code or the documentation. The man page for 'read.maf' says that the 'maf' argument is a
"tab delimited MAF file. File can also be gz compressed."
But (at least on Windows as far as I can determine), whenever I pass in a gzipped MAF file, I get the same error about more columns than column names. If I first manually gunzip the file, it works just fine. I view the failure of the performance to match the documentation as a still-open bug.

[PoisonAlien]

Hi,
Yes you are right about it. I will try to fix it by weekend. I guess its because the first line of input file is a comment (generally indicating MAF version) and read.csv is struggling to handle it.

[krcoombes]

Not sure why you would use read.csv instead of read.table for a tab-separated file. But even less sure of why you get different results when gunzip-ing in R instead of with an external tool. Is it as simple as making sure you include the "skip=1" parameter?

…

On Feb 14, 2017 7:52 PM, "PoisonAlien" ***@***.***> wrote: Hi, Yes you are right about it. I will try to fix it by weekend. I guess its because the first line of input file is a comment (generally indicating MAF version) and read.csv is struggling to handle it. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#40 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFp9yC6LS9F2hS2epWBuJ6g0Nt7tVSIOks5rckxOgaJpZM4L2pbk> .

[PoisonAlien]

There is not much difference between table and csv reader, hence I used read.csv. Regarding second point yes, even I am not sure about whats going on. skip=1 will skip header if there are no comment lines, better way is to use comment.char = "#". Most of the maf files I checked are on either linux or MacOS, I will check again on windows machines and fix it soon.

[zayni]

Hi,
getting errors this morning. I am a Mac User

reading maf..
Error in data.table::fread(input = maf, sep = "\t", stringsAsFactors = FALSE, :
Input is either empty or fully whitespace after the skip or autostart. Run again with verbose=TRUE.
It was working well yesterday. Not sure what is going on?

[rafalcode]

worth mentioning you need
library(magrittr)
for the "%>%"

[PoisonAlien]

wow this is amazing! Thank you for the effort and making maftools a part of TCGABiolinks :)