ABRF Mass Spectrometry Peptide Sequencing Quiz Results

 

First of all

I'd like to thank those who participated. Although I neglected to ask how much time each person spent on this work, I know that it can be substantial. I hope it was worthwhile.

 

Goals of study

To reiterate, the goals of this project were 4-fold:

  1. Serve as a final exam for those who took the 1999 ABRF peptide sequencing course taught by Prof. Don Hunt.
  2. Establish whether there is a need for similar courses in the future.
  3. Allow individuals to anonymously assess their own level of proficiency at the task of sequencing peptides using mass spectral data.
  4. Allow a comparison of two types of mass spectrometers commonly used for peptide sequencing.

 

Were the goals met?

Relevant to goal #1 above, only one person who had attended the Don Hunt course actually submitted answers to this quiz. Hence, 99% of the attendees of the course failed the exam. The person who had taken the course (secret code #1) did a nice job of it, though.

As for goal #2, there were a total of twenty responses, which is perhaps an insufficient number from which to judge the need for additional sequencing courses. My guess is that more formal courses are not necessary. In addition to a certain degree of apathy (as judged by the low number of responses), it seems that those who did submit answers generally did a decent job of it. This includes those who said that they had '0' or '1' years of experience in sequencing peptides by mass spectrometry. It appeared as though those with three or more years of experience were no less error-prone than the novices. Furthermore, development of this type of skill lends itself well to web-based instruction, which holds some advantage over formal courses that occur over a limited time and space.

Goal #3 was successfully achieved, and the correct answers are

Of course, I didn't expect anyone to be able to rationally distinguish between Leu and Ile using low energy CID data; however, most people familiar with US popular culture from several decades back were able to correctly guess the Leu's and Ile's in Problems #1 and #3. A compilation of the answers, along with other information, provided by the 20 respondents can be found here. Problems #1 and #2 were comprised of MS/MS and MS/MS/MS data for these two peptides and were obtained using a Thermoquest LCQ Classic ion trap mass spectrometer. Problems #3 and #4 were comprised of MS/MS spectra obtained using a Micromass Qtof quadrupole / time-of-flight hybrid mass spectrometer. Synthetic peptides #1 and #3 had sequences that were expected to give data that were equally easy to interpret. Likewise, peptides #2 and #4 had sequences that were expected to yield data that would be equally difficult to interpret. Specifically, the latter peptides contained prolines near the N-terminus, which can result in extensive (and confusing) internal fragment ion series. Also, these peptides contained Gly-Gly, Ala-Gly, and Gly-Ala in their sequences, which can often be confused with Asn (= Gly + Gly) and Gln/Lys (= Gly + Ala).

Goal #4 was also met, in the sense that data for the two types of mass spectrometers has been presented for people to examine. However, the data set (n = 20) is too limited to conclusively answer the question of which mass spectrometer is best suited for the purpose of peptide sequencing. Of course, from a practical standpoint other issues come into play (i.e., "Is the difference in data quality worth the 3-fold difference in price?" or "Can I even get that kind of money?"). Also, this study did not address some other practical issues such as sensitivity, although I personally find both instruments to have similar sensitivities. Nevertheless, when asked to comment on the relative merits of the Qtof versus the ion trap, twelve of the twenty had a preference for the data obtained from the Qtof (full scan w/ higher mass accuracy). Three respondents said they were about equal, and one preferred the ion trap data (ability to obtain MS/MS/MS spectra). Some of the preferences could be attributed to what each respondent uses in their own labs.

 

Other comments about the study

Eleven of the respondents said that they have been required to correctly sequence a peptide de novo using tandem mass spectrometry at least once. Most of the rest were afraid that some day they might be called upon to do so. Regarding the use of O-18 labeling of the C-terminus, nine said that it was helpful, and seven felt that it was not necessary. Eleven respondents were involved with academic facility laboratories, three were from pharmaceutical companies, one was from a biotech company, one was a graduate student, and one was an unemployed drifter.

 

Computer algorithms for de novo sequencing

Given the tedious and repetitive nature of manually interpreting tandem mass spectrometry data, one might wonder if computers would find this task more enjoyable. The history of automated de novo sequencing using mass spectrometry has had a checkered past (and, perhaps, present), but there are nevertheless a few programs available that one might consider using. A program can be purchased from Micromass for use on their instruments. The one I wrote for sequencing tryptic peptides is called "Lutefisk1900"; executables for Macintosh and Windows, along with the source code, can be downloaded (http://www.hairyfatguy.com/). "PEPSEQ" is another freeware program that can be downloaded. Another one is called "SeqMS", but I haven't had a chance to look into either of these last two. I probably overlooked or forgotten some others. For fun, you can see the results obtained using the Micromass program and Lutefisk1900 here.