I have only experience with Mascot and i think this is the most used. But have u used MS/MS? Otherwise you measured a mixture of proteins which will give you no result.

A really important question is on what organism are you working.

If it is human, for exemple, with MS/MS data, a mascot search on the swissprot database should be enough (NCBInr would be great too).

First question: did you do a LC separation before your MALDI analysis? Generally, Co-IP results are too complex to be analyzed by direct spotting. For the database search: if you have access to a computer running ProteinPilot with the Paragon algorithm (ABSciex), you might be able to identify several additional peptides / proteins due to the number of modifications which can be search simultaneously.

I agree with Mario and Sebastien. Mascot is working pretty well with MALDI data. But obviously you need to perform MSMS to confirm your identifications. It depends also on your MALDI performance (resolution, accuracy): if the id are obtained with an error bigger than 50ppm, I will doubt the results, especially without MSMS. SwissProt is well cured and it is not redundant, NCBI is probably to large.

Hope it helps

Stella

Hi,Mascot database is the best option a MALDI TOF and MS/MS analyses .

Thank you very much everyone!

@Mario, I did not use MS/MS. I guess we do not have MS/MS option. We have MALDI-TOF machine from Bruker microflex.

@Sabastein, Yes, I am working with human retinal pigment epithelial cells.

@Chris, I didnt do LC separation before MALDI analysis. Is that ProteinPilot with the Paragon algorithm (ABSciex), is a free software? I did not use this before. Could you tell me little more details?

@ Stella, I am using more than 100 ppm some times. So What is the main criteial to confirm the peptide/ protein? Is sequence coverage? or peptide match or score?

Thanks you Poomalai.

Hi, I guess MASCOT is one of the the best search engine for MALDI MS analysis using Swissprot or NCBInr databases. but if you have a simple mixture of few proteins, it is possible to identify up to 3-4 proteins using second pass search where in the subsequent search, you will remove peptide masses matched with first protein identified and so on..better to keep peptide mass tolerance at below 50ppm..

Dear Srinivas,

many of the above posters are absolutely right, when they point out the importance of the database (in addition to the search engine that you are using). You are likely to get the best results, when you are searching a genomic database for the organism that your samples were derived from. I would also agree with everybody who points out that Mascot is an excellent search engine which has been successfully used for peptide mass fingerprinting in many published papers. A great resource are the publicly accessible Mascot help pages, which tell you a lot about how you should interpret the results that you get from a search.

See: http://www.matrixscience.com/help/pmf_help.html

Have also a look at the 'peptide mass fingerprinting' section on the ms-utils web page (a fantastic resource of different free software packages for the processing and analysis of mass spectrometry data):

http://www.ms-utils.org/wiki/pmwiki.php/Main/SoftwareList

Hope this helps,

Achim

ProteinPilot is a commercial software from ABSciex (see http://www.absciex.com/products/software/proteinpilot-software or for a free trial license: https://licensing.absciex.com/download/index ) which uses a different algorithm than Mascot. The Paragon algorithm has the big advantage that a wide number of modifications can be search without compromising the search space (see Mol Cell Proteomics. 2007 Sep;6(9):1638-55.). I don´t know if the algorithm already works for single spot searches.

Did you separate your Co-IP products by SDS-PAGE? Or did you digest the whole set of proteins and spot it directly?

Christoph,

ProteinPilot is a great piece of software, but it will only work with MSMS data. However, you and a few others absolutely correctly point out that the peptide mass fingerprint data is pointless, unless the proteins in the Co-IP have been separated prior to the tryptic digest (I hadn't spotted earlier that Srinivas might not have done this). It is likely that 2D electrophoresis would be necessary to do this (but: be aware of all the 2DE-associated problems, such as losing membrane proteins, large proteins, small proteins, alkaline proteins and acidic proteins in the process).

Alternatively it will be necessary to re-do the experiment using an LCMSMS based approach.

Be aware that even once you have an experiment that will allow you to identify proteins, it is by no means an easy thing to understand what this list means. Not all proteins that you identify in a Co-IP experiment are bait interactors!

Angus Lamond's lab in Dundee have published a few excellent papers over the past few years that deal with the biggest problem in this kind of experiments: how do you distinguish between proteins that are 'real interactors' and the very large number of contaminants that you will pull down and identify as well in this kind of experiment. A nice one to start with is here:

ten Have S, Boulon S, Ahmad Y, Lamond AI: Mass spectrometry-based immuno-precipitation proteomics - the user's guide. (http://www.ncbi.nlm.nih.gov/pubmed/21365760)

Hope this helps,

Achim