Hi,

I am having trouble with memory usage and the SWATH2stats R package. Specifically, when I attempt to split the transition groups into single transitions (in order to go on to use MSstats or MapDIA), using:

data.transition <- disaggregate(data.filtered)

I receive:

error: cannot allocate vector of size 7.0Gb 

The data has been processed through the OpenSwathWorkflow, and reports ~4,000 protein ID hits, with ~16,000 unique peptides, and ~70,000 transitions . I have filtered and reduced the data using SWATH2stats, however it is for global profiling, so I do expect relatively high numbers. 

Aside from increasing RAM (I am currently running 64-bit R, on Windows 7, with 32GB RAM), I was wondering if there are any work-arounds with SWATH2stats when dealing with data as large as this? Or whether any alternative tools are be better suited?  

Any advice or tips appreciated!

Jess

Dear Jess,

thank you for your e-mail and this is a known problem. There are three options:

1.) Try to empty your workspace as much as possible.

2.) Use it on a machine with more available RAM. One option that I typically use is to run it on a Windows Remote Desktop Server from our University (if available).

3.) Otherwise you find in the package a python script (subfolder scripts) that does the same thing (see Vignette for more details). You need to export the data from R, then run the python script, and then read it again into R.

I’m sorry for the inconvenience. I didn’t find so far a way to program it in a way that allows to process large data without requiring a lot of RAM.

Best regards

Peter