With regard to the fact that unknowns are not in the data bases, the
advances in molecular biology allows us to go from the small amount of
sequence obtained by ms to longer stretches in clones. The strategy of
combining sequencing of the clone, with comparison of the peptides already
identified in the original ms analysis quickly confirms that the correct
clone has been identified, as well as ensuring that frame shifts and other
cloning artifacts have not occurred. The incomplete clone can then be used
in data base searches. The chance of getting a related protein or protein
family is actually quite high. This information can be used to guide the
remainder of the cloning strategy.
The real key is the availability of experienced mass spectrometrists who
can interpret sequence data, do data base searches and sequence alignments
for homology studies to guide the cloning.
Katheryn Resing
>Dear ABRFers,
>
>I would like to add one comment to the discussion of peptide mass
>fingerprinting. Ken wrote in his last comment that he feels that mass
>mapping is a bit oversold. I certainly agree with him in this respect. We
>have seen a flood of papers (very impressive ones, I must say) which imply
>that almost any protein can be identified if it is in some data base.
>Whether this is achieved by mass mapping alone or by searching data banks
>with CAD spectra is not being discussed here. However, I am under the
>impression that the technology is stuck in the 'identification corner'.
>However, how many organisms are sequenced completely? 12 or 15? I don't
>know. If you think how many other organisms are not sequenced the
>available sequences for identification of a given protein are SMALL!
>Especially the mass spec companies are trying to sell us the idea that
>once you have the proper technology and software at hand, you can identify
>almost anything. Yes, that's true, but only if something is in a data
>bank. Let's face it: if you are dealing with an unknown, mass spectrometry
>is pretty bad to deduce a sequence (except those nice peptides which
>happen to fragment very well, but our experience is against that, namely
>that you can read a tiny bit of sequence, the rest is garbage). And I
>think out of such inability to produce sequence data from any given
>peptide came the idea to turn this into an advantage, namely to identify
>proteins rather than going through dedicated structure elucidation of real
>unknowns.
>
>Don't get it wrong: we've been given extremely versatile MS methods for
>protein identification. However, it is foreseeable that the sequencing
>projects which produce the basis for all the identification methods will
>stop at some point. And then what? Don't forget the aspect that there are
>still so many unknown proteins in various organisms that I find it
>personally a bit sad if such an elegant method as mass spectrometric
>structure elucidation is being kept on the level of mere protein
>identification.
>
>Best regards
>
>Paul
>
>=========================================================================
>
>Paul Jenoe, PhD
>Department of Biochemistry
>Biozentrum of the University of Basel
>Klingelbergstrasse 70
>CH-4056 Basel
>Switzerland
>
>Tel. ++41 61 267 21 68
>Fax. ++41 61 267 21 48
>e-mail jenoe@ubaclu.unibas.ch
>
>=========================================================================