A Statistical Method for Significant Analysis of Comparative Proteomics Experiments Based on LCMSMS - PowerPoint PPT Presentation

Loading...

PPT – A Statistical Method for Significant Analysis of Comparative Proteomics Experiments Based on LCMSMS PowerPoint presentation | free to view - id: 8c724-ZDc1Z



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

A Statistical Method for Significant Analysis of Comparative Proteomics Experiments Based on LCMSMS

Description:

A Statistical Method for Significant Analysis of Comparative Proteomics ... Seattle, WA; &Department of Bioengineering, University of Washington, Seattle, WA. ... – PowerPoint PPT presentation

Number of Views:30
Avg rating:3.0/5.0
Slides: 2
Provided by: pwa4
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: A Statistical Method for Significant Analysis of Comparative Proteomics Experiments Based on LCMSMS


1
A Statistical Method for Significant Analysis of
Comparative Proteomics Experiments Based on
LC-MS/MS Experiments
Pei Wang, Brian Piening, Martin McIntosh,
Amanda G. Paulovich Division of Public Health
Science, Division of Clinical Research, Fred
Hutchinson Cancer Research Center, Seattle, WA
Department of Bioengineering, University
of Washington, Seattle, WA.
Model
Abstract
Consider One Protein
One of the essential challenges of comparative
proteomics using LC-MS/MS instrumentation is the
need to assess quantitative differences between
two samples using only qualitative information,
such as the presence or absence of a peptide in
different samples. We propose a statistical
method SASPECT (Significant AnalysiS of PEptide
CounTs) for quantitatively identifying proteins
differentially expressed between two groups based
on the results of tandem mass spectral
experiments.
Suppose this protein has K peptides
Key Assumption
The probability of a proteins being observed in
one LC-MS/MS experiment is proportional to its
abundance in the complex sample.
In contrast to spectral counting, SASPECT uses
the Boolean values for the presence or absence of
a positive peptide identification instead of the
raw values of spectral counts, for the latter are
more subjected to the changes of various
experimental factors. In addition, by properly
controlling the false discovery rates (FDR),
SASPECT provides quantitative guidance in
peptides and proteins selection.
Challenges and Solutions
Compared with other similar approaches in the
microarray literature (searching for
differentially expressed genes), the LC-MS/MS
problem has several additional sources of
variability
Model Fitting with Expectation-Maximization (EM)
procedure
Initialization
Initial guess of whether each peptide and the
protein appear in each LC-MS/MS experiment.
For t1,, B
Need to account for database search errors of
peptide identification and protein identification.
Make explicit use of peptide/protein
probability assignments ( for example,
PeptideProphet/ProteinsProphet scores).
Maximization
2.
Calculate the maximum likelihood estimators for
all parameters using guessed values (conditional
expectation) of unobserved variables from last
iteration.
Need to remove the artificial effect due to
different total CID numbers in different LC-MS/MS
experiment.
Employ a rescaling factor according to the
different total CID numbers in each experiment.
3.
Expectation
Need to properly control the False Discovery
Rate (FDR) when testing hundreds or thousands of
proteins simultaneously.
Calculate the conditional expectation of
unobserved variables using parameter estimators
from the maximization step.
Adapt the permutation procedures of SAM to our
test statistics. (See section of Model for more
details)
4.
Software
The approach is provided freely in an open-source
program (R code). http//peiwang.fhcrc.org/resear
ch-project.html http//proteomics.fhcrc.org/
CPL/home.html
Control False Discovery Rate (FDR)
We estimate FDR with the permutation procedure
proposed in SAM (Significant Analysis of
Microarray) (Tusher et.al. 2001, PNAS 98,
5116-5121).
PeptideProphet Keller et al., Anal Chem 74,
5383-5392, 2002 ProteinProphet
Nesvizhskii et al., Anal Chem 75, 4646-4658, 2003
About PowerShow.com