Biology Reference
In-Depth Information
searches. A single peptide (a single MS/MS spectrum) may even cor-
rectly identify a protein depending on the number of amino acids in its
sequence. However, the downside of PFF identification lies in the high
proportion of unassigned MS/MS spectra collected during an experi-
ment. The usual explanations for these nonmatches are the presence of
experimental contaminants appearing before or during MS acquisition,
bad-quality spectra with noise and unusual fragmentation for certain
mass spectrometers, spectra derived from proteins not present in the
database or products of alternatively spliced genes not annotated in the
database, etc.
Some algorithms have been developed to deal with most of these
issues, although none of them currently handle all problems at once.
Some have opted to reduce the number and complexity of MS/MS
spectra while increasing their quality (e.g. NoDupe
37
), but most avail-
able software have adopted a classical search strategy. Most classical
search programs split the identification process into two stages. The
first stage is aimed at building a list of candidate proteins from confi-
dently identified spectra, and the second one is aimed at matching
unidentified spectra against this list with more combinatorial parame-
ters (e.g. taking into account a larger number of modification types or
increasing the mass error tolerance). The main idea behind this strat-
egy is to increase the sequence coverage by loosening constraints.
Popular programs of that category are Mascot
28
; Sequest
38
; Phenyx,
39
a
software platform developed by GeneBio S.A.
c
in collaboration with
the PIG; and X!Tandem.
40
Classical database search algorithms aim to identify the best
sequence match given a list of fixed and/or potential modifications in a
database for each spectrum analyzed. The advantage of classical search
tools is the speed with which larger data sets can be processed in a rea-
sonable amount of time. The drawback is their conceptual limitation of
identifying only spectra with expected modifications. Open modification
search tools, also known as tag approach or blind search, address this
problem. This search strategy is also based on detecting matches
c
Phenyx runs publicly through a Web interface (http://phenyx.vital-it.ch/); its
commercial version is installed locally.
