Description

Parser to convert MSI imzML files into HDF5 files.

Preconditions

Python is needed for execution. We recomment to install anaconda from https://www.anaconda.com/distribution/#download-section . Afterwards, use the predefined anaconda environment. Therefore, open your console, navigate into the imzML-to-HDF5 package and call conda env create -f .\environment.yml. Activate the environment with activate msiparse (Windows users should use the CMD instead of powershell). Activation needs to be done every time before the parser is called. Deactivate the environment with deactivate.

Execution

1. Call activate msiparse.
2. Call python msiproc.py to convert from imzml + ibd to h5 data. This requires a lot of memory for bigger imzml/ibd files and should be executed on a cluster. For parameter details call msiproc.py -h

Optional

(a). Provide a mir file (peak selection file) to reduce data sets generated to the specified peak selection.

(b). Call python consensus_peaklist.py to align multiple peak selected data sets. Peaks with a user specified percentual overlap will be merged. For parameter details call python consensus_peaklist.py -h.
Important: This method will only change the representative m/z value of peaks. The intensity will remain as the sum over the original peak range, i.e. a shifting will be simulated without changing the peaks in the original data. The new peak values should not be used in relation with the original data set!

(c). Provide a mis file to generate position mapping images and scaled images.

Hints and issues:

• The following files have to be named equally: imzml, ibd(, mis).
• Computing consensus peaks can cause some issues with due to peak chaining effects and multiple peak overlapping, e.g. two peaks of D₁ overlap with one peak of D₂.
  None of these Problems should raise with enough spacing between selected peaks.