Database Search Strategies for Proteomic Data Sets Generated by Electron Capture Dissociation Mass Spectrometry



research articles

Sweet et al.


Figure 7. (A) ECD identification with a Mascot peptide score of 0.36. (B) ECD identification with a Mascot peptide score of 1.96. In both
cases, all fragments are identified with mass accuracy better than 10 ppm. Mass spectra are shown as acquired, i.e., prior to trimming.


ported.9 ETD and ECD result in similar fragmentation patterns
and it is therefore likely that the described ETD algorithm
would be useful for ECD spectral processing. We compared the
two data processing methods directly, using the same set of
DTAs (generated by the Good algorithm from raw files). The
results are shown in Table 5. The ETD algorithm allows removal
of the precursor window, neutral losses up to 60 Da from the
charge-reduced precursor and the charge-reduced precursor
ions themselves. Both processing options result in an increase
in identifications compared to the unprocessed DTAs, with the
ECD processing algorithm resulting in the most identifications.
These searches resulted in fewer identifications than the
equivalent searches using DTAs generated by Bioworks. Bio-
works automatically corrects the precursor mass to the monoiso-
topic value even if the second isotopic peak was selected for
fragmentation. The DTAs generated using the ETD algorithm
may be out by 1 Da if the second isotopic peak was selected
for fragmentation. While these identifications are recovered by
filtering two windows, one around 0 Da error and the second
around 1 Da error, the likelihood of a false-positive is now
increased.

Table 5. Comparison with Good9 Algorithma

presearch

forward
hits (2
+)

reverse

hits      I

D rate

ECD search (3341 DTAs); Good Algorithm; Mascot

DTAs generated using
Good algorithm.

1423 (848)

14

42.6

DTAs generated and processed
using Good algorithm.

1470 (873)

14

44.0

DTAs generated using Good
algorithm; processed using
ECD algorithm (as in Table
4).

1507 (915)

15

45.1

a Identifications from doubly-charged precursors are shown in
parentheses. To achieve the estimated FDR of 1%, results were filtered
according to Mascot scores (scores of 4.09, 0.79 and 8.4, respectively).

Applicability to Phosphoproteomic Data Set. In earlier
work, we identified over 900 phosphopeptides from a similar
mouse whole-cell lysate sample after TiO
2-based phosphopep-
tide enrichment.5 From a total of 6080 ECD DTAs, 1087
phosphopeptides identifications were made (with redundancy).
These identifications were made using the same search (OMSSA)
and postsearch strategies (filtering by precursor mass error and

5482 Journal of Proteome Research Vol. 8, No. 12, 2009



More intriguing information

1. Urban Green Space Policies: Performance and Success Conditions in European Cities
2. A Theoretical Growth Model for Ireland
3. The name is absent
4. Palvelujen vienti ja kansainvälistyminen
5. The Value of Cultural Heritage Sites in Armenia: Evidence From a Travel Cost Method Study
6. The name is absent
7. The name is absent
8. International Financial Integration*
9. The name is absent
10. The effect of classroom diversity on tolerance and participation in England, Sweden and Germany
11. The name is absent
12. Voting by Committees under Constraints
13. The Impact of EU Accession in Romania: An Analysis of Regional Development Policy Effects by a Multiregional I-O Model
14. Ventas callejeras y espacio público: efectos sobre el comercio de Bogotá
15. LAND-USE EVALUATION OF KOCAELI UNIVERSITY MAIN CAMPUS AREA
16. Disturbing the fiscal theory of the price level: Can it fit the eu-15?
17. Brauchen wir ein Konjunkturprogramm?: Kommentar
18. The name is absent
19. A Principal Components Approach to Cross-Section Dependence in Panels
20. The name is absent