Biomedical Engineering Reference
In-Depth Information
of two spectra. The result of this global alignment procedure is illustrated
in Fig. 2(c) and Fig. 2(d) for two dierent mass regions. It is clear that
after the linear operation on spectrum 2, peaks in two spectra have been
brought into alignment.
It has been shown that this alignment procedure is eective and easy to
carry out, as long as two spectra share some known common peaks. After
aligning two spectra based on these known common peaks, one may identify
if there are other common peaks that are not obvious before alignment. This
is performed by comparing the shift in peak positions in all possible peak
pairs with the estimated peak position uncertainties. If the shift is large
relative to the uncertainties, they will be treated as separate peaks that
show in one spectrum but not in the other. Otherwise, they are identied
as common peaks in both spectra. After aligning two spectra, this alignment
can be extended to multiple spectra in the same fashion.
4. Summary
The peak detection algorithm discussed here is based on a physical under-
standing of a TOF-MS spectrum and utilizes Bayes' theorem. The peak
detection algorithm is automated and can be applied to a variety of TOF-
MS spectra, from counting experiments (TOF-SIMS) to instruments like
MALDI-TOF which integrate the ion signal rather than counting individual
ion. The noise characteristics are dierent in these devices. The algorithm
automatically provides estimates of uncertainties in the peak positions and
intensities. This information is used to verify that aligned spectra collected
at dierent spatial positions or at dierent times are properly aligned.
Acknowledgments
The authors are greatful to Dr. M.W. Trosset for his helpful comments. Fi-
nancial supports from National Cancer Institute Phase I and II SBIR grant
(CA101479) and the Virginia Commonwealth Technology Research Fund
(IN2003-04) are also acknowledged. This work was performed in part using
computational facilities at the College of William and Mary which were en-
abled by grants from Sun Microsystems, the National Science Foundation,
and Virginia's Commonwealth Technology Research Fund.
References
1. M. Mann, R.C. Hendrickson, and A. Pandey, Analysis of proteins and pro-
teomes by mass spectrometry, Annu. Rev. Biochem., 70 (2001), 437-473.
Search WWH ::
Custom Search