Hi Maria,

it very much depends on what kind of MS data you have and how it was generated.

Have you done any tandem MS of your peptides (CID, ETD, ECD)? This is what you need for identifying the position of your phosphorylation sites.

There's quite a few options in terms of automated data analysis, and selecting the right software is not always straightforward, particularly for phosphoprotemics due to the chances of misidentifying the phosphorylation sites.

Protein Prospector (http://prospector.ucsf.edu/prospector/mshome.htm) is an on line software you could try to search your data against a selected database (Uniprot, Swissprot and so on...). It will take data generated from pretty much any instrument (Q-ToF, low resolution ion trap, FT-ICR, with either ESI or MALDI as ionization source) as long as you can convert your raw data into an mgf file or equivalent to be uploaded on the software. This is something you should be abel to do by using your instrument-specific software for data analysis, which normally runs a script to generate the .mgf file (or equivalent).

Mascot is another very popular search engine which used MS data for peptides (and proteins) identification, although the free online version is limited to very small raw data files.

There are also more 'sophisticated' tools like Ascore which are particularly useful for distinguishing between isomeric phosphopeptides (it was originally designed for Thermodata).

Either way, a confident identification most of the times requires manual inspection of the raw tandem MS, as the search engine can sometime misidentify the right position of the modification. This is more than an issue with CID of phosphorylated peptides containing more than one S, T or Y residues which sometimes fails to generate enough sequence specific b and y ions for a confident identification.

I would also strongly recommend some background reading to have a feeling of what different search engines and different MS data can actually tell you.

As for the database, I normally refer to phosphosite plus http://www.phosphosite.org , which also includes whether a particular phosphosite was identified by MS.

Kind regards,

Francesco

Hi Maria,

You can try the following from cell signaling:

http://www.phosphosite.org/homeAction.do;jsessionid=4BF69E55841DDE5EF01292AFC89FA747

I am sure there are other ones, but this is pretty easy to use.

Thanks for the link, I checked it. The site looks very useful, but unfortunately, it doesn't comprise the yeast genome, only mammals.

I have encountered the same problem. Have anyone fixed it?

Hi Maria,

You can have a look at PHOSIDA... http://www.phosida.com/

Regards,

Srikanth

Thank you, Srinivas Srikanth, this looks indeed useful.

Hi Maria,

it very much depends on what kind of MS data you have and how it was generated.

Have you done any tandem MS of your peptides (CID, ETD, ECD)? This is what you need for identifying the position of your phosphorylation sites.

There's quite a few options in terms of automated data analysis, and selecting the right software is not always straightforward, particularly for phosphoprotemics due to the chances of misidentifying the phosphorylation sites.

Protein Prospector (http://prospector.ucsf.edu/prospector/mshome.htm) is an on line software you could try to search your data against a selected database (Uniprot, Swissprot and so on...). It will take data generated from pretty much any instrument (Q-ToF, low resolution ion trap, FT-ICR, with either ESI or MALDI as ionization source) as long as you can convert your raw data into an mgf file or equivalent to be uploaded on the software. This is something you should be abel to do by using your instrument-specific software for data analysis, which normally runs a script to generate the .mgf file (or equivalent).

Mascot is another very popular search engine which used MS data for peptides (and proteins) identification, although the free online version is limited to very small raw data files.

There are also more 'sophisticated' tools like Ascore which are particularly useful for distinguishing between isomeric phosphopeptides (it was originally designed for Thermodata).

Either way, a confident identification most of the times requires manual inspection of the raw tandem MS, as the search engine can sometime misidentify the right position of the modification. This is more than an issue with CID of phosphorylated peptides containing more than one S, T or Y residues which sometimes fails to generate enough sequence specific b and y ions for a confident identification.

I would also strongly recommend some background reading to have a feeling of what different search engines and different MS data can actually tell you.

As for the database, I normally refer to phosphosite plus http://www.phosphosite.org , which also includes whether a particular phosphosite was identified by MS.

Kind regards,

Francesco

Thank you all for the valuable suggestions. In fact, I am not an expert on MS technology. I had some complex samples analyzed by LC-MS/MS by a collaborator, who has left the lab since. I have a list of all detected peptides in every experimental condition, as well as the corresponding proteins in an EXCEL file, and this data unfortunately does not include the exact information of the position of the phosphosite within the matched protein.

I was just wondering whether there are simple tools available to look for the P-site positions myself - of course I can ask the collaborator's lab to use the tools that Ashwin suggested above.

In principle what I need is not a software to interpret raw MS data (since this has already been done by my collaborator), but a software that takes my phosphopeptides as input, aligns it with the matching protein sequence of interest and gives me back the P-site position.

In my group, when given only a peptide list (possibly with modifications), we usually run a FASTA file of detected underlying peptides (ie stripped of any mods) through BLASTP against an organism-specific database of unmodified protein sequences. The resulting matches then require manual curation. Of course this assumes you have already localized mods on the detected peptides.

More often, we have raw LC-MS/MS data. In case you can access these data, our tool of choice has been InsPecT, which includes an interesting blind search feature. The TransProteomic Pipeline (TPP) is also a great resource. It integrates many tools that cover several steps of proteomics data analysis, most recently including a version of the Phosphate Localization Score (PLS).

To do anything more sophisticated than my first suggestion (like getting any solid statistics to back your conclusions), you probably need access to the raw data and/or a proteomics bioinformatician. Of course all of this depends on precisely what resources and kind of data you have to work with and what type of experiment was used to obtain it.

If any of that was helpful, let me know if you have further questions, though the best I may be able to do is fine pointers to relevant proteomics literature.

Best of luck!

'In principle what I need is not a software to interpret raw MS data (since this has already been done by my collaborator), but a software that takes my phosphopeptides as input, aligns it with the matching protein sequence of interest and gives me back the P-site position.'

This is the basic principle of automated data analyses by search engines like Mascot, Protein Prospector, OMSSA, Sequest etc, they will give you a matching of theoretically predicted and experimental b and y ions (if you're dealing with CID data), neutral losses (as specified in your searching settings), ion scores (measurement of the quality of your annotated tandem MS, based on mass difference between theoretic and experimental ions and percentage of 'annotated' ion current), an expectation value or similar (how likely is your ID to be a random event), as well as info as to scan number, retention time etc so you can track down the raw file and manually confirm the ID (assuming you want to do that).

BW,

Francesco

Hey Maria,

1) Where your phosphopeptide residues within a matched protein

Given that you have excel files, a quick way to possibly map your phosphosites to your proteins to to look at the start position of the peptide relative to the protein (there is very likely to be a column in the excel file which says 'starts from' or something similar) and add this number to the phosphosite position on the peptide e.g.

P-Peptide is phosphorylated on site 5 and starts from 123 so the phosphosite relative to the protein is 128.

2) Cross-referencing with a public phosphopeptide repository to see whether it's previously been recorded etc

This is much more tricky and would be very labour-intensive if you plan on doing it manually. I had a quick scan of the literature and there seems to be a nice tool called 'PhosphoBLAST' where you input your phosphopeptide query sequences (from your MS) and the phosphosite databases.

If you want something that runs on your own computer and you have full control, this looks like a nice start but you'll need to be careful about the format it requires to work.

There is a webserver on PhosphoELM (another phosphosite database) which has a 'PhosphoBlast' (I don't think it's the same one) but this database hasn't been updated for a little while I think.

Word of warning about phosphosite databases though -> They'll contain both sites that have been determined by high-throughput (e.g. discovery type MS/MS) and low-throughput (e.g. mutating specifically that site) means -> The latter as a general rule are probably more trust-worthy as your real sites.

These two sites are the best out there, flexible, just imput your secuence or search the protein name, with show mice and humans. Have fun.

http://www.phosphosite.org/homeAction.action

http://www.hprd.org/PhosphoMotif_finder