search

kusterlab / prosit

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
https://www.proteomicsdb.org/prosit/
Apache License 2.0
85 stars 45 forks source link

Web site spectral library generation - cysteine residue mass #13

Closed karlssoc closed 5 years ago

karlssoc commented 5 years ago

Hi, I have generated a Prosit spectral library via https://www.proteomicsdb.org/prosit/ and compared the synthetic library (in spectronaut format) to a DDA/OpenSwathWorkflow-based counterpart (n ~5500 peptides).

For most peptides with modifications (oxidized methionine or carbamidomethylation) the PrecursorMz for the Prosit version is lower than OpenSwath. The most common delta PrecursorMz of OpenSwath - Prosit when considering peptide charge state is 57.02. It appears as the web service is not considering all C as treated C with carbamidomethylation.

I am not sure if this is an issue or if I misunderstood something.

An example:

OpenSwath ADC(UniMod:4)VQTLLLNQQR, PrecursorCharge 2 779.9039 ADC(UniMod:4)VQTLLLNQQR, PrecursorCharge 3 520.2717

Prosit ADC[Carbamidomethyl (C)]VQTLLLNQQR, PrecursorCharge 2 751.3932 ADC[Carbamidomethyl (C)]VQTLLLNQQR, PrecursorCharge 3 501.2646

https://www.proteomicsdb.org/prosit/task/E62AFBA12ECBF8F4FEAF3D0D530A31BF

Thanks, Christofer

tkschmidt commented 5 years ago

Thanks for the notice. You were/are completely right and I made a small mistake at the precursor calculation step (just for the csv format). The fragment masses were not affected by that.
I pushed a new version of our website right now and it looks right, could you check if it is fine for you as well?

karlssoc commented 5 years ago

Thanks! The problem is resolved: I generated a second version of library today and the delta-PrecursorMz:s are now ranging between -0.000072549 to 0.000061701.

tobigithub commented 5 years ago

@karlssoc good catch. tnx!

jmueller95 commented 4 years ago

Hey all, is it possible that this has only been resolved in the online version of Prosit, but not in the version in this git repo? I just started using it today, and for any peptide that contains C, the precursor m/z seems to be too small. E.g. for 3+ charged "AAALLCQLLL", the precursor m/z is reported as 343.543939, however, if the Cysteine is carbamidomethylated, I believe it should be around 362.54398. Since it's my first time using Prosit, I might be misunderstanding something, though.

Best wishes, Julian

tobigithub commented 4 years ago

@jmueller95 Could be, maybe run 10-20 relevant molecules and compare? Its all about validation.

jmueller95 commented 4 years ago

@tobigithub Thanks for the fast reply! I ran some peptides containing Cysteine and calculated unmodified and C+57 precursor m/zs for each of them (peptides_w_cysteines.txt). The m/z reported by Prosit always agrees with the unmodified m/z, even though the fragment m/zs account for the mass shift.

tkschmidt commented 4 years ago

hey, we will investigate