Create Training dataset – Training Prosit on my own data.

[ekvall93]

Hi, I would like to train Prosit on my dataset, but I have some problems transforming my data to be compatible with Prosit. I have tried to extract the data from the mzml file and then used "match.augment" to generate a dataframe containing data compatible with Prosit, which I then can create my hdf5-files. However, I seem to have trouble reproducing the data you have provided in https://figshare.com/articles/dataset/ProteomeTools_non_tryptic_-_Prosit_fragmentation_-_Data/12937092
i.e., the dataset you provide seems to get different intensities and find more ions than what I'm getting using match.augment.

Do you use further preprocessing on the data besides normalizing the intensities (I'm using the "ITMS" option) before calling match.augment?

Any guidance on how to generate the hdf5 files from a parsed mzml-file would be appreciated.

Best regards,
Markus Ekvall