Biology Reference
In-Depth Information
3.8 Quantification of
Identified Proteins by
Normalized Spectral
Counting
As detailed in the introduction, spectral counting is an inexpensive
and robust quantification technique for proteins in complex mix-
tures (
see
Note 11
). This type of quantification works best for com-
plex mixtures and yields reliable results only for very large datasets.
Thus, including technical replicas will make the quantification more
reliable. Spectral counting should be based on either the original
APEX procedure described by Lu and colleagues [
4
] or the modi-
fied APEX-indexing procedure nSpC that was described by
Baerenfaller and colleagues [
14
]. The difference of both procedures
is that Lu and colleagues take peptide detection probability into
account to define the set of theoretically detectable tryptic peptides
(see formula below). This probability estimate is based on a predic-
tion algorithm [
17
] and as such prone to errors and incapable of
dealing with different elemental composition of the sample. We
therefore suggest a simplification by taking all theoretically detect-
able tryptic peptides into account that are in principle accessible to
the mass spectrometric analysis. In nSpC, the expected contribution
of each individual protein to the sample total peptide and spectra
pool is calculated by assuming an equimolar distribution of proteins
and then correcting the assumption by a factor that is calculated
from the actually measured contribution of the protein. The follow-
ing formula balances between detected and expected number of
spectra assigned to all tryptic peptides in the complete sample:
-1
æ
ö
TTP
TTP
proteink
AbundanceProteinKMTP
=
´
´
MTP
ç
÷
proteink
samp
le
è
ø
sample
MTP
protein K
= Measured spectra of trypic peptides from protein
K (peptide spectrum matches).
TTP
protein K
= Theoretical tryptic peptides (TTP) of protein K.
MTP
sample
= Measured spectra of tryptic peptides (PSM) in
sample.
TTP
sample
= TTP of proteins identified in sample.
For the determination of the number of TTP
k
, we digested the
whole Arabidopsis protein database (TAIR10) with trypsin
in
silico
. If Arg or Lys was followed by Pro (KP/RP site), the site was
both cut and not cut (resulting in three tryptic peptides). If several
of these sequence pairs followed each other, we only considered
cutting of one KP/RP site per time. The resulting peptides were
labeled as TTP in case they have a mass between 400 and 6,000 Da
and comprise at least six amino acids. The list of TTP of the entire
Arabidopsis proteome is available at
www.pep2pro.ethz.ch
.
3.9 Alternative
Normalization
Schemes
Depending on the biological question to be addressed, different
normalization schemes for peptide spectrum matches have been
applied. When analyzing the adaptation of the plastid proteome to
a mutation, especially when the mutation causes a severe defect in
