Make sure that the following precuations are taken care of-

1) The FASTA database used for search should have the protein sequence of interest

2) Correct parameters pertaining to the instrument (MALDI-TOF/TOF or OrbiTrap or LTQ ion trap etc) and experiment should be used(no. of missed cleavages, enzyme specificity, label used, fixed and variable modifications etc.

3) A sensitive algorithm should be used for search (Mascot/Sequest/OMSSA/XTandem/MassWiz/Crux etc)

I assume you may have already taken care of these things but just in case...

Amit we expect this is a novel protein and we already perform your suggestions.

In such a case, try against swissprot database or NCBInr, but again- your experimental design will have a profound effect on downstream analysis method (and their pros and cons). So it would help if you can elaborate on what methods were followed.

Amit basically I have not a strong knowledge in bioinformatics, so please tell me in details for further study of the protein (novel) for its identification through m/z ratios

If you can provide details of your protocol, this will help me in giving a directed suggestion instead of generalized solutions that may or may not fit your case. For eg- if u used MALDI TOF/TOF, the answer would differ from the same experiment done on ESI-Trap or Orbitrap or LTQFT. Was it MS1 or MS2?(I assumed MS2 all the way).

Even if protein is novel, the sequence for peptides can still be deduced by de novo approach which is more diffcult, laborious and subjective but will provide the answer.

It may sound stupid, but have you checked the performance of your MS? I am asking this because previously we have a lot of difficulties in protein ID, and it turns out the main problem is with our instrument e.g. mass accuracy, sensitivity, etc. Try running some standard proteins through your protocol to see how good the protein ID workflow is doing.

Once you have a good quality control in your workflow, the remaining problem will be how to ID a novel protein. I believe your "novel" protein still share some conserved region with known proteins. In this case you should check your peptide ID result, but not protein hits.

If you have high-quality fragment spectra, de-novo sequencing of your fragment spectra using either open source (e.g. PepNovo from Ari Frank´s lab, http://proteomics.ucsd.edu/Software/PepNovo.html) or commercial software (e..g PEAKS) can provide you with candidate peptide sequences which you can the use in combination with MS-BLAST(dove.embl-heidelberg.de/Blast2/msblast.html) to identify homologs of your protein that are present in the database.

Hope this helps,

Stefan

Hi Soni! which Software and which protein database did you use for MS search? If you only use "bovine" as taxonomy you can relax your parameter and set "Mammalia". You can identify the protein form another sequenced organism.

Dear Soni,

everybody else has already provided you with excellent advice. If you are confident that your mass spectrometry data is of good quality (control proteins are identified using the same workflow with good scores), it might also be worth including the common repository of adventitious proteins (cRAP, pronounce: cee-RAP) in your search (http://www.thegpm.org/crap/index.html) before you go off and try de novo sequencing.

Good luck,

Achim

Thanks to all, basically we have not mass spectroscopy facility in our department, we purify the protein and it send to a private company, they send us a collection of m/z data then we go to the mascot

If it is indeed a novel protein (i.e., not in the databases you are searching your mass spectra against), then you will of course not find an ID for the protein no matter how high the quality of the MS and MS/MS spectra are. However, with high quality spectra de novo sequencing may be informative as suggested by others, though not trivial for non-mass spec experts (and even for lots of mass spec experts, although the suggested algorithms for giving good candidate sequences can be useful). If you expect that the protein may have homologs in other mammalian organisms, then expanding your search database from Bovine to Mammalia may be useful, as suggested by others. Use of Error Tolerant searching (Mascot) or algorithms like ProteinPilot which do a better job of finding e.g., matches with single amino acid changes such as what one might find in protein homologs from other species may also be helpful.

Bovine system has complete database with SwissProt accession numbers. It will provide you the sequence. I prefer you visit matrixsciences.com to identify the m/z values and confirm it by scoring with Mascot online database.

AFAIK Bos taurus genome is completely sequenced and you should have no problem identifying bovine protein in, for example, Uniprot database. Maybe, your sample is contaminated by protein from different organism? Try searching the complete DB, not to limit the search to Bovine only.

I am suspecting that something went wrong. A protein purified from bovine system would at least give you some impurities because the modern mass specs have very good sensitivity and a dynamic range of about 4 orders of magnitude. I would suggest that you test drive the whole procedure (digestion, LC-MS/MS, database search) with some standard protein for trouble shooting. Good luck.