PABST

From SPCTools

PABST overview

 This page describes the Peptide Atlas Best SRM Transition tool or PABST, which is a Peptide Atlas functionality that uses
various sources of information to produce lists of peptides and (optionally) fragment ions for use in selected reaction 
monitoring (SRM) assays.  This information is compiled into 'builds' that allow for fast querying via a web interface, 
which can be found at following URL:  https://db.systemsbiology.net/sbeams/cgi/PeptideAtlas/GetTransitions

 If you are interested in accessing the transitions programmatically you can find more information [here].

 The PABST build process has 3 discrete steps which are described in more detail in the following paragraphs.  First of all,
a list of the the top N (currently 10) most suitable peptides is determined for each protein in the target organism.  
Secondly, a list of the top M (currently 8) potential fragment ions from each selected peptide is determined.  Finally
this information in loaded into the Peptide Atlas database and forms are provided to query and retrieve list of potential
assays. 

Peptide list selection

 The criteria used by PABST for selecting the list of most suitable peptides from a target protein are primarily based on the 
likelihood of observation and the presence or absence of various sequence features.  To address the former, we first query 
the Peptide Atlas and find all peptides mapping to a subject protein that have been observed more than once.  We also 
generate a list of all possible trypic peptides from this same peptide, and apply two separate peptide detectability 
algorithms, Peptide Sieve and Peptide Detectability Predictor.  The information is then merged, with the empirical (EPS)
proteotypic score being used if available, else the theoretical proteotypic score is used.  

 The sequence is then evaluated based on length, amino acid composition, and redundancy (peptide maps to more than one 
protein and/or peptide maps to more than one region on genome), and a final score is calculated.  This is done by
applying weighting factors which can be modified by editing the configuration file fed to the command line peptide
selector.  This is described in some detail on [this page].

  The proteotypic score obtained from step one is then adjusted by the various weighting factors in step two, and a final
score from each peptide is determined.  The list of peptides is sorted by this score in descending order, and the 
desired number of peptides are selected.

Fragment ion list selection

  The best fragment ions for a given spectrum are determined by querying a series of sources in order of preference for potential 
ions.  Once the desired number of fragment ions is collected the process goes on to the next peptide, otherwise it continues to 
the next source in order.  

 The first source considered is PATR, the Peptide Atlas Transition Resource, which stores lists of SRM transitions obtained 
from the literature and from user submission.  

 Next, the process looks at data from consensus spectral libraries, as these have information about which fragment ions 
are known to be generated in CID, as well as their  relative intensities.  If a QQQ (triple quadrupole) library available, it 
is scanned first for the peptide of interest.  Next, a QTOF library is used if available, and finally an ion-trap library,
generally the most complete, is consulted.  For a given spectrum that matches the subject peptide, peaks are taken in order 
of descending relative intensity.

 Finally, theoretical fragments are determined and collected based on the following rules.  Ions with m/z > 2500 or m/z < 500
are omitted, as are fragments within +/- 25 units of precursor ion m/z.  Of the ions that are not excluded, we first take any
singly charged y-ions with m/z > precursor m/z in order of ascending m/z.  Next we take any Y ions < precursor m/z in order of 
descending m/z.  Finally, the list is augmented with B ions if necessary.