Hi

Not knowing what type of MS data you have in your hands, I can't say anything useful. In case you have MS/MS data you can try interpret the peptide sequence with one of the peptide prediction programs (PepNovo, Lutefisk, CIDentify, Sherenga; Pepnovo is available free and perhaps the most accurate, mostly they accept CID spectra) and do a homology search.

I agree with Jana. MSMS data of your peptides will make life much easier.

There is a brand new tool, published just two weeks ago!

J Proteome Res. 2012 Mar 12. [Epub ahead of print]

Spectral clustering in peptidomics studies allows homology searching and modification profiling: HomClus, a versatile tool.

Menschaert G, Hayakawa E, Schoofs L, Van Criekinge W, Baggerman G.

Thanks for your comments. We did a quick MALDI-TOFMS and are currently trying to do exactly what you say using a nanoLC-MS/MS approach. We have very limited sample, so if that doesn't work, and peptide fragmentation is not guaranteed, I am also looking for other approaches i.e. shortlist peptide hormones around the MALDI-estimated masses and use an antibody blocking screening approach. So, back to my question, is there an on-line resource giving a comprensive list of known biologically active peptides around a particular mass, or that can be ranked according to mass?

It is better you follow this link http://pepbank.mgh.harvard.edu/. Another method is blasting by suitable methods.

But this require the sequence. Further you contact [email protected] . There is no other shortcut method available presently. lastly and better method is spectrometry

Have you tried mascot?

Pepbank appears to be for peptides smaller than 20 aa, which is not my case, but thanks for the suggestion. HomClus looks interesting although there is no obvious public access to the tool. I'll take a closer look at the paper - thanks.

Yes, I have tried Mascot, but it requires input from a tryptic or other kind of digest, and gives an identity based on peptide fingerprinting. In my case, I am looking for an identity based on one peptide mass, and Mascot doesn't work that way (as far as I am aware). I can't trypsinise each peptide because I have only a semi-purified peptide mix and have to start again with another study if I want to get enough material to obtain purified peptides (which we shall do, but will take a long time).

Hi John - any chance your MALDI-TOF is capable of post-source decay, so that you can force a bit of fragmentation of the peptide? If you know its mass and can get even a couple fragment ions, it will make your problem a bit easier.

Also, if you are planning on doing LC-MS/MS, make sure you explore using an "inclusion list", which will prioritize your peptide(s) of interest for MS/MS (versus data-dependent acquisition). If the peptide is of low abundance, you may be able to tweak the acquisition parameters (if you're using an ion trap or similar trapping instrument) to boost your signal.

Hello.......I think you should perform denovo sequencing of your peptide on Maldi-tof-tof by N-terminal derivatization. It requires very less amount of sample. We have performed the same successfully. Plz go through the link for protocol.......

http://www.springerlink.com/content/373kw18k00278160/ or chen et al (2004) in reference section.....

Unless you can dramatically narrow the list of possible peptides (prior knowledge), you will not be able to get anything useful from just mass alone. It is not just different sequences, but also different PTMs that would make the space of possible matches huge. Even with fragmentation spectra the identity of a peptide can still be unclear.

Thanks to everyone, that's very helpful. Susan, your LC-MS/MS suggestion is on board - thank you. Santosh: nice method but we lack that instrumentation in my lab. Roman: I completely agree that it's a challenge, but I do know the biological activity and localisation of the peptide, so to some extent I can eliminate candidates that do not show those characteristics. Today we got a partial (8 aa) sequence on one candidate by LC-MS/MS but it is not in any database that I have tried so far...a leap into the unknown I suspect.

John,

I don't think there is an off-the-shelf tool for what you want. In general, the mass of a peptide that is 20 aa or larger, as you've said yours is, even measured at a very high mass accuracy, will be shared by a very large number of possible amino acid sequences; of course, not all of those possible sequences will be expressed by your organism. But since any peptide mass is the sum of an integer number of C, H, N, O, S, and possibly P atoms, the mass is basically a count of the number of protons plus neutrons in the molecule, nudged a tiny bit higher or lower by the identity of the atoms in which those protons and neutrons are present in your peptide; and since the relative amount of C, H, N, O, and S in amino acids is fairly uniform, that nudge tends to average out in large peptides, with the result that measuring the mass to even higher accuracy doesn't reduce the number of peptides that have that mass by very much. As Roman said above, the mass alone just isn't terribly informative.

To amplify on Roman's comment about prior knowledge, do you know whether your peptide occurs naturallyin your organism, or has it been produced from a larger protein by e.g. digestion with trypsin in the lab? Your request for a list of biologically relevant peptides suggests to me that it might occur naturally. In that case a list of the peptides that occur naturally in your organism, and what modifications they may carry, would give you a much smaller set of candidates. Unfortunately I don't work on that kind of peptide, so I don't know where you would find one; in any event, it would need to be specific to the organism you're studying. But if/when you have such a list, the mass of each of these can be computed from their chemical formula. Note that if your observed m/z is highly accurate, it will be important to use enough significant digits for your atomic (or amino acid) masses, and for the mass of a free proton (which is different from the mass of a hydrogen atom!) for each positive charge (if your ion is negatively charged, the mass of the electron(s) can safely be ignored). Since most mass spectrometers don't provide better than 1 part per million in mass accuracy, it should be sufficient to use masses that are accurate to 1 part in 10^6, i.e. six significant digits.

If your 20+ amino acid peptide was generated by digestion with a known protease (either in vivo or experimentally), then you can use the cutting specificity of the protease to limit the set of peptides you consider; this is typically what is done in e.g. Mascot to help find a list of peptides that might be seen in a tryptic digest of all proteins expressed by a specific organism. Even with a known charge state and a highly accurate peptide mass, though, masses corresponding to peptides of at least, say, 8 amino acids or more will correspond to literally hundreds to thousands of tryptic peptides expressed by a cellular organism. However, if you can identify the masses of other peptides from the same protein, the combination of four or five peptide masses can identify the protein fairly confidently. If you're interested in that approach, look for literature on "mass mapping" tools.

Thanks very much for your thoughts on this. You raise the important point that a lot of bio-active peptides are naturally generated from larger proteins (I have not treated my samples with any protease), which means that this process is a potential target for the treatments that I am studying. We now have some fairly accurate masses from multiply charged ions of candidate peptides and, as I mentioned above, some sequence information, although so far only for one peptide of interest.

I also had a similar problem with my novel bioactive protein from a natural source as i was not able to fingerprint the identity of my protein after tryptic digestion of in-gel samples by MALDI-MS and Mascot and Prospector database searching. But the N-terminal sequence by Edman degradation was helpful to characterize the identity of my protein-or close to it. Now iam working towards deciphering the nucleotide sequence by degenerate primer construction.

have you tried TagIdent?

Thank you Avani - this is exactly what I was looking for!

Bharat: if I can purify the peptide(s) of interest, this would be the best way forward - thanks for the suggestion.

Hi John,

it seems to me you don't have a very accurate idea of what the MW of your peptide is. You could do a simple MS/MS experiments with a high resolution/accurate mass MS like an Orbitrap. Not really a cheap option but there is an Orbitrap in the Chemistry department of your university. If you ask nicely, I'm pretty sure they can help you.

I agree with Laurent's opion. in fect, MS/MS experitment with a low resolution mass spectrometer is enough. We can identified the peptides with the theoretic fragments to identify them from their isomers.