Quantification of Proteins by Label-Free LC-MS/MS - LC-MS/MS in Proteomics: Methods and Applications - page 225

all search results of all analyzed samples. The identity of the pep-
tides is attached based on the accurate mass and retention time.

Normalization should be applied to each experiment indepen-
dently. The normalization should account for experimental vari-
ation leaving biological variation unchanged. The choice of nor-
malization method depends on the experimental design. There
are two basic options. The first includes normalization based on
an exogenous internal standard which is spiked into each sample
either before LC-MS analysis, in which case it should be digested
first, or in the initial sample preparation stage, in which case it
should be spiked as an intact protein. During the normalization
step in the data analysis only the signal from the internal standard
peptides are used to calculate the normalization factor.
The second method is total ion current normalization. In this
method the intensity measurements of all detected peptides in a
given sample are used to calculate the normalization factor.
Whichever method is used, the sum of intensities of all pep-
tides in each sample is used to calculate the normalization factor.
The sample with the largest sum is set to 1 and all other sums
are divided by the intensity of the highest sample ( see Fig. 13.2 ) .
Each intensity value is then multiplied by the appropriate normal-
ization factor.

3.3.2.Normalization

The summarized data matrix includes all detected peptides and
their corresponding intensity in all replicates of all samples, as
mentioned previously. However, for various reasons not all pep-
tides are detected in all replicates or samples. The number of times
a given peptide was detected reflects the reliability of the measure-
ment. For example, a peptide that was detected in only one repli-
cate of three can be regarded not reliable and should be excluded.
The same is true for peptides that are not detected in several sam-
ples. Thus a data set can be filtered to exclude those peptides that
are not detected in at least two of three replicates for each sample.
The second filtering is on the sample level such that a peptide will
be included in the down-stream analysis only if it replicates in the
majority of samples in each treatment group (i.e., diseased and
control). This method of filtering ensures that a qualitative differ-
ence, e.g., detection only in one treatment group, will be noted
for further analysis. This is shown for peptide 5881 in Fig. 13.3 .

3.3.3. Filtering

The database search results for all injections are used to annotate
the data set. This is an automated process in which the identity
of the peptides is matched to the entries in the data matrix based
on the accurate molecular weight and retention time measure-
ments. The peptide with the highest identification score of all
search results is used for annotation ( see Fig. 13.4 ) . Since not all
peptides are fragmented efficiently, identification will be available

3.3.4.Annotation