mzAlign causing crashes

[Andrew Goodenough]

I have been attempting to calibrate some of my images to reference peaks using the mzAlign function. I thought at first this might be an issue with memory, so I tried it on small groups and single pixels. It worked for most pixels, but would consistently crash on others. There is no error message to go off of, as R crashes completely. 

I'm not sure if this is happening because those peaks aren't found in those pixels or what. If anyone knows what is causing this or how to work around this, it would be much appreciated. Thanks,

Andrew

[kbemis]

Do any other preprocessing methods fail on those spectra, or is it only mzAlign()?

-Kylie

[Andrew Goodenough]

Hi Kylie,

peakBin() also fails on those same pixels. Some more information on the image: it is from a low (20k) resolution MALDI-ToF (mass accuracy is not fantastic when imaging) and I read in an Analyze file for the data variable below. Here is the processing code:

dataProc = data %>%
  normalize() %>%
  smoothSignal() %>%
  reduceBaseline() %>%
  peakPick() %>%
  process()

dataFilt = dataProc %>%
  peakAlign(tolerance = 0.2, units = "mz") %>%
  peakFilter() %>%
  process()

dataBinned = data %>%
  normalize() %>%
  smoothSignal() %>%
  reduceBaseline() %>%
  peakBin(mz(dataFilt), tolerance = 0.2, units = "mz") %>%  # Running this block individually (normalize, then smoothSignal, etc.), this step causes a crash.
  process()