Prosit offers high quality MS2 predicted spectra for any organism and protease as well as iRT prediction. When using Prosit is helpful for your research, please cite "Gessulat, Schmidt et al. 2019" DOI 10.1038/s41592-019-0426-7
Thank you for your support! I have a question: can you provide more detail on how to perform the rescoring of predictions made by Prosit from a peptidelist.csv file (derived from a MaxQuant msms.txt file) via Percolator? Especially, given the fact that Prosit doesn't output any cosine similarity to the MS2 spectra (which I understand is obvious since the RAW file is not an input of Prosit).
I'm referring here to a previous question from another user to whom you provided a very short answer. It would be great if you can pinpoint me to some tutorial or some detailed explanation. Thanks!
Best,
Victor
Hi Marc,
you can download the code here and run it locally, but you would need a server with an NVIDIA GPU. Here we only provide the prediction tool though. For the rescoring you would need build prediction based features and run a percolator on that files to rescore them.
I'd like to use Prosit for rescoring locally as you was trying to.
Have you found a way to build features from Prosit output, that you can feed to Percolator ?
Dear Siegfried or Tobias,
Thank you for your support! I have a question: can you provide more detail on how to perform the rescoring of predictions made by Prosit from a peptidelist.csv file (derived from a MaxQuant msms.txt file) via Percolator? Especially, given the fact that Prosit doesn't output any cosine similarity to the MS2 spectra (which I understand is obvious since the RAW file is not an input of Prosit).
I'm referring here to a previous question from another user to whom you provided a very short answer. It would be great if you can pinpoint me to some tutorial or some detailed explanation. Thanks!
Best, Victor
Hi Marc,
you can download the code here and run it locally, but you would need a server with an NVIDIA GPU. Here we only provide the prediction tool though. For the rescoring you would need build prediction based features and run a percolator on that files to rescore them.
Originally posted by @gessulat in https://github.com/kusterlab/prosit/issues/16#issuecomment-507152063