Initial MSiMass List

[Markus Stoeckli]

The motivation for us to start the MSiMass list reaches back about 14 years when we started identifying the few proteins we could measure off tissue. I had then our informatics guy program a software which would go through SwissProt and calculate all potential masses with the known modifications. This proved to be very powerful, as it allowed to tentative identification based on the mass.

I realize that there are many databases on proteins and MS and that I’m probably not aware of most of them. But to my knowledge there is not one DB which focusses on MALDI MS(I) directly from tissue sections. The bases for this initiative is the assumption, that there is a limited number of proteins which can be seen by this approach, and that many of these are already published. We asked ourselves what people would most likely query from such a database. We came up with observed mass and protein name as the major fields.

Based on this, the aim of this database (or simple list if you will) is to collect observed masses (of proteins) in MSI experiments and list the corresponding reference. The aim was to make this a simple process, without a formal review process. So, by design the database will be incomplete and have wrong entries. If you see errors, please report them and we’ll fix them. But most important, this allows us to collect data very rapidly from many users. This design flaw does in my view not hamper the usefulness of the database: If you observe a mass in you tissues, you can check this list and if you find a match, you can then validate it by checking the primary literature.

Any feedback is welcome. Changes to the scope and design… just anything which moves this along. Please involve also other people you would like to give input.

This is a community effort and I hope you’ll contribute as well.

Markus

[Liam McDonnell]

The MSi list is looking very good indeed, and will be of enormous help to the field. Well done! The lists are likely to get quite long, especially for on-tissue digestion (and in the future lipids) so I would recommend splitting the lists into the different biomolecular classes and ionization method. Even if there is no immediate scope for filling lists for lipids etc the presence of protein and tryptic peptide lists should spur similar efforts for imaging MS of other molecular classes, and using other ionization methods.

If we consider MALDI imaging MS of proteins I would separate the lists into one for imaging of tryptic peptides and one for (non-tryptic) peptides and proteins. The reason for this is that there are different considerations for the two identifications. For example, as seen in the current list tryptic peptides can be non-specific, and so will have multiple hits. For tryptic peptides it would be useful to indicate which tryptic peptide was detected along with its modifications and missed cleavages (which increase the peptide search space and the risk of false positives).

On the other hand MALDI imaging MS of (non-tryptic) peptides and proteins can detect the products of proteolytic processing. Several reported biomarker proteins (e.g. Cazares et al. in Cancer Res) are non-tryptic fragments of larger proteins. Many of the peptides and small proteins detected by protein imaging experiments of resected tissues could be protein fragments because of post-mortem degradation during the resection procedure and subsequent tissue handling steps. The database should include which fragments were detected.

In short having different lists for the different molecular classes will enable the tables to include the most relevant information (for that class) within the available space.

Please let me know what you think.

With best regards
Liam

[Liam McDonnell]

Another point, it would be helpful if the list differentiated between peptides/proteins that had been assigned only on the basis of mass accuracy and those which have been validated, either by on-tissue MS/MS or IHC. The link to the original article would then also provide the validation method.

[Liam McDonnell]

New entry for the database
m/z 2778 (linear MALDI ToF) in human chondrosarcoma (bone tumor)
identified as a nontryptic, C- terminal fragment of the protein vimentin, LIKTVETRDGQVINETSQHHDDLE using ETD on an Orbitrap Velos.
DOI of paper: DOI: 10.1021/pr301190g
Uniprot ID: P08670
On-tissue validation method: synthesis of isotopically labeled variant and comparison of on-tissue MS/MS

The same paper also reports another 350 top down ID’s but these have not been confirmed in the MALDI data.

[Liam McDonnell]

Hi Markus

My postdoc Benjamin Balluff has just indicated there is an error in the database, which mirrors an error in the cited paper.

In the database it is current listed as

3371 DEFA1
3442 DEFA2

However according to Uniprot (http://www.uniprot.org/uniprot/P59665) DEFA1 is the heavier peptide.

[Markus Stoeckli]

Thanks, changed accordingly. Markus

[Author]

Posts