Organelle Proteomics by Label-Free and SILAC-Based Protein Correlation Profiling - LC-MS/MS in Proteomics: Methods and Applications - page 259

4. Operate the mass spectrometer in the data-dependent mode
to acquire precursor ion and fragment ion spectra for protein
identification and quantitation.

PCP and PCP-SILAC experiments necessitate an extensive extrac-
tion of data for protein identification and quantitation. A
number of software tools fulfill these requirements such as
Mascot (MatrixScience) for protein identification and MSQuant
or MaxQuant (17) for label-free and SILAC-based protein
quantitation.
1. Extract peak lists from the mass spectra and search a
sequence database to identify proteins.
2. PCP : Quantify the abundance of each peptide ion signal in
each fraction by integrating the ion current signal intensity
extracted from the precursor ion mass spectra ( Fig. 15.1B ) .
Normalize the peptide ion abundance profile to the fraction
with maximum signal. Calculate the median of the normal-
ized abundance profiles for all peptide ion signals repre-
senting the same protein ( see Note 8 ) and plot the profile
( Fig. 15.1C ) .
3. PCP-SILAC : Quantify the lysine- and arginine-containing
peptides from the precursor ion spectra and calculate the
relative abundance of peptides in each fraction as the peak
area of the medium/light and heavy/light isotope ratios.
Compute the median of log 2 -transformed medium/light
and heavy/light isotope ratios for all peptides representing
the same protein in each fraction and use these values to
plot two relative protein abundance profiles for each protein
( Fig. 15.2D ) .
4. Compare the protein profiles for each protein with the pro-
files of organelle marker proteins. Proteins with compara-
ble profiles are likely organelle-associated candidates whereas
proteins with dissimilar profiles are likely contaminants
( Fig. 15.2E ) .

3.5.Determinationof
ProteinAbundance
Profiles

3.6.Statistical
Analysis

The data obtained from PCP and PCP-SILAC experiments can
be subjected to statistical analysis to test the likelihood of a given
protein to be annotated as an organellar protein. Various methods
have been applied for the statistical analysis of protein profile data.
The measurement of profile similarity between a given protein
and a set of known organelle proteins requires prior knowledge
whereas cluster analysis and principle component analysis ( 5 , 12 )
are unbiased. Here we describe the calculation of the Mahalanobis
distance ( see Note 9 ) between the profile for each protein and the
profiles for a group of organelle-associated proteins ( Fig. 15.2F ) .
1. Express the peptide ratios in a vector form, x =

( x 1 , x 2 , , x n ), where each dimension corresponds to a gra-

dient fraction, and n is the total number of fractions. Ratios

...