DiffBind: error with dba.count

[blakeoft]

Hello,

I've been using DiffBind to analyze some ATAC-seq peak data. Briefly, peaks in these datasets represent sites of open chromatin. DiffBind seems to do a very good job, but I've encountered an error in my most recent comparison. Here's the R code:

> library(DiffBind)
> diffpeaks <- dba(sampleSheet = "metadata.csv")
# suppressed output
> diffpeaks <- dba.count(diffpeaks)
# suppressed output
[B]Error in if (sum(tokeep) < length(tokeep)) { : 
  missing value where TRUE/FALSE needed[/B]
In addition: Warning message:
In dba.multicore.init(DBA$config) :
  Parallel execution unavailable: executing serially.

I've bolded the actual error. I've looked through the dba.count() code, and I can't tell where the error is coming from. My best guess is that it's from the pv.count() function, but I can't seem to find anything about it anywhere. This is really a mystery to me (and I thought I knew what was going on in R too )...

I’ve selected subsets of the metadata with varying sizes--sometimes it works, and sometimes I get the same error. This leads me to believe that it isn’t a memory issue.

Does anyone have any ideas about this?

Thanks for any help,

Blake

[rory]

It appears that DiffBind can't handle the "-" character (minus sign/hyphen) in the file names. Remove them and it should work!

Cheers-
Rory

[lime gin]

Hi Blake, just curious if this has been resolved. I'm getting the same exact error with my data; while I've pinpointed the exact subset where the error is coming from, my file names all have alphanumeric characters.

[blakeoft]

Unfortunately, I have not resolved the issue yet. I'm just curious, what kind of machine are you running DiffBind on? I've used it several times without error on a Unix based HPC. But ever since I've tried using it on my windows computer, I get this error. The HPC has been down for maintenance, so I haven't been able to see if running it there solves the problem. I'll post my results whenever the cluster is back up.

[lime gin]

Thanks for your response. I'm running it on my Windows computer through RStudio. We do have a Unix based HPC, but I've never tried running R there. I'll give it a shot and see how it goes.

[lime gin]

Update: I ran the same set of data on our Unix based HPC and it worked!

[blakeoft]

That's good to hear. I really don't have any sort of an idea of why DiffBind (specifically dba.count()) does not seem to work on Windows. I look forward to when I can use our HPC again! Thanks for the update.