Computational and Statistical Methods for Protein Quantification by Mass Spectrometry

by Eidhammer, Ingvar; Barsnes, Harald; Eide, Geir Egil; Martens, Lennart

• ISBN13: 9781119964001
• ISBN10: 1119964008
• Edition: 1st
• Format: Hardcover
• Copyright: 2013-02-11
• Publisher: Wiley

Summary

The definitive introduction to data analysis in quantitative proteomics. This book provides all the necessary knowledge about mass spectrometry based proteomics methods and computational and statistical approaches to pursue the planning, design and analysis of quantitative proteomics experiments. The author's carefully constructed approach allows readers to easily make the transition into the field of quantitative proteomics. Through detailed descriptions of wet-lab methods, computational approaches and statistical tools, this book covers the full scope of a quantitative experiment, allowing readers to acquire new knowledge as well as acting as a useful reference work for more advanced readers. With clear and thorough descriptions of the various methods and approaches, this book is accessible to biologists, informaticians, and statisticians alike and is aimed at readers across the academic spectrum, from advanced undergraduate students to post doctorates entering the field. Computational and Statistical Methods for Protein Quantification by Mass Spectrometry: Introduces the use of mass spectrometry in protein quantification and how the bioinformatics challenges in this field can be solved using statistical methods and various software programs. Is illustrated by a large number of figures and examples as well as numerous exercises. Provides both clear and rigorous descriptions of methods and approaches. Is thoroughly indexed and cross-referenced, combining the strengths of a text book with the utility of a reference work. Features detailed discussions of both wet-lab approaches and statistical and computational methods. With clear and thorough descriptions of the various methods and approaches, this book is accessible to biologists, informaticians, and statisticians alike and is aimed at readers across the academic spectrum, from advanced undergraduate students to post doctorates entering the field.

Author Biography

Ingvar Eidhammer, Department of Informatics, University of Bergen, Norway

Harald Barsnes, Department of Biomedicine, University of Bergen, Norway

Geir Egil Eide, Centre for Clinical Research, Haukeland University,Norway

Lennart Martens, Department of Biochemistry, Faculty of Medicine and Health Sciences, Ghent University, Belgium

Table of Contents

Preface

1 Introduction

1.1 The composition of an organism

1.1.1 A simple model of an organism

1.1.2 Composition of cells

1.2 Homeostasis, physiology and pathology

1.3 Protein synthesis

1.4 Site, sample, state and environment

1.5 Abundance and expression - protein and proteome profiles

1.5.1 The protein dynamic range

1.6 The importance of exact specification of sites and states

1.6.1 Biological features

1.6.2 Physiological and pathological features

1.6.3 Input features

1.6.4 External features

1.6.5 Activity features

1.6.6 The cell cycle .

1.7 Relative and absolute quantification

1.7.1 Relative quantification

1.7.2 Absolute quantification

1.8 In vivo and in vitro experiments

1.9 Goals for quantitative protein experiments

1.10 Exercises

2 Correlations of mRNA and Protein Abundances

2.1 Investigating the correlation

2.2 Codon bias

2.3 Main results from experiments

2.4 The ideal case for mRNA-protein comparison

2.5 Exploring correlation across genes

2.6 Exploring correlation within one gene

2.7 Correlation across subsets

2.8 Comparing mRNA and protein abundances across genes from

two situations

2.9 Exercises

2.10 Bibliographic notes

3 Protein-level Quantification

3.1 Two-dimensional gels

3.1.1 Comparing results from different experiments - DIGE

3.2 Protein arrays

3.2.1 Forward arrays

3.2.2 Reverse arrays

3.2.3 Detection of binding molecules

3.2.4 Analysis of protein array readouts

3.3 Western blotting

3.4 ELISA - Enzyme-Linked Immunosorbent Assay

3.5 Bibliographic notes

4 Mass Spectrometry and Protein Identification

4.1 Mass spectrometry

4.1.1 Peptide mass fingerprinting (PMF)

4.1.2 MS/MS - Tandem MS

4.1.3 Mass spectrometers

4.2 Isotope composition of peptides

4.2.1 Predicting the isotope intensity distribution .

4.2.2 Estimating the charge

4.2.3 Revealing isotope patterns

4.3 Presenting the intensities - the spectra

4.4 Peak intensity calculation

4.5 Peptide identification by MS/MS spectra

4.5.1 Spectral comparison

4.5.2 Sequential comparison

4.5.3 Scoring

4.5.4 Statistical significance

4.6 The protein inference problem

4.6.1 Determining maximal explanatory sets

4.6.2 Determining minimal explanatory sets

4.7 False discovery rate for the identifications .

4.7.1 Constructing the decoy database

4.7.2 Separate or composite search

4.8 Exercises

4.9 Bibliographic notes

5 Protein Quantification by Mass Spectrometry

5.1 Situations, protein and peptide variants

5.1.1 Situation

5.1.2 Protein variants - peptide variants

5.2 Replicates

5.3 Run - experiment - project

5.3.1 LC-MS/MS run

5.3.2 Quantification run

5.3.3 Quantification experiment

5.3.4 Quantification project

5.3.5 Planning quantification experiments

5.4 Comparing quantification approaches/methods

5.4.1 Accuracy

5.4.2 Precision

5.4.3 Repeatability and reproducibility

5.4.4 Dynamic range and linear dynamic range

5.4.5 Limit of blank - LOB

5.4.6 Limit of detection - LOD

5.4.7 Limit of quantification - LOQ

5.4.8 Sensitivity

5.4.9 Selectivity

5.5 Classification of approaches for quantification using LC-MS/MS

5.5.1 Discovery or targeted protein quantification .

5.5.2 Label-based vs. label-free quantification

5.5.3 Abundance determination - ion current vs. peptide identification

5.5.4 Classification

5.6 The peptide (occurrence) space

5.7 Ion chromatograms

5.8 From peptides to protein abundances

5.8.1 Combined single abundance from single abundances

5.8.2 Relative abundance from single abundances

5.8.3 Combined relative abundance from relative abundances

5.9 Protein inference and protein abundance calculation

5.9.1 Use of the peptides in protein abundance calculation

5.9.2 Classifying the proteins

5.9.3 Can shared peptides be used for quantification?

5.10 Peptide tables

5.11 Assumptions for relative quantification

5.12 Analysis for differentially abundant proteins

5.13 Normalization of data

5.14 Exercises

5.15 Bibliographic notes

6 Statistical Normalization 82

6.1 Some illustrative examples

6.2 Non-normally distributed populations

6.2.1 Skewed distributions

6.2.2 Measures of skewness

6.2.3 Steepness of the peak - kurtosis

6.3 Testing for normality .

6.3.1 Normal probability plot

6.3.2 Some test statistics for normality testing .

6.4 Outliers

6.4.1 Test statistics for the identification of a single outlier

6.4.2 Testing for more than one outlier

6.4.3 Robust statistics for mean and standard deviation

6.4.4 Outliers in regression

6.5 Variance inequality

6.6 Normalization and logarithmic transformation

6.6.1 The logarithmic function

6.6.2 Choosing the base to use

6.6.3 Logarithmic normalization of peptide/protein ratios

6.6.4 Pitfalls of logarithmic transformations

6.6.5 Variance stabilization by logarithmic transformation

6.6.6 Logarithmic scale for presentation

6.7 Exercises

6.8 Bibliographic notes

7 Experimental Normalization

7.1 Sources of variation and level of normalization

7.2 Spectral normalization

7.2.1 Scale based normalization

7.2.2 Rank based normalization

7.2.3 Combining scale based and rank based normalization

7.2.4 Reproducibility of the normalization methods

7.3 Normalization at the peptide and protein level

7.4 Normalizing using sum, mean and median

7.5 MA-plot for normalization .

7.5.1 Global intensity normalization

7.5.2 Linear regression normalization

7.6 Local regression normalization - LOWESS

7.7 Quantile normalization

7.8 Overfitting

7.9 Exercises

7.10 Bibliographic notes

8 Statistical Analysis

8.1 Use of replicates for statistical analysis

8.2 Using a set of proteins for statistical analysis

8.2.1 Z-variable

8.2.2 G-statistic

8.2.3 Fisher-Irwin exact test

8.3 Missing values

8.3.1 Reasons for missing values

8.3.2 Handling missing values

8.4 Prediction and hypothesis testing

8.4.1 Prediction errors

8.4.2 Hypothesis testing

8.5 Statistical significance for multiple testing .

8.5.1 False positive rate control

8.5.2 False discovery rate control

8.6 Exercises

8.7 Bibliographic notes

9 Label-based Quantification

9.1 Labeling techniques for label-based quantification

9.2 Label requirements

9.3 Labels and labeling properties

9.3.1 Quantification level

9.3.2 Label incorporation

9.3.3 Incorporation level

9.3.4 Number of compared samples

9.3.5 Common labels

9.4 Experimental requirements

9.5 Recognizing corresponding peptide variants

9.5.1 Recognizing peptide variants in MS spectra

9.5.2 Recognizing peptide variants in MS/MS

Spectra

9.6 Reference-free vs. reference-based

9.6.1 Reference-free quantification

9.6.2 Reference-based quantification

9.7 Labeling considerations

9.8 Exercises

9.9 Bibliographic notes

10 Reporter-based MS/MS Quantification

10.1 Isobaric labels

10.2 iTRAQ

10.2.1 Fragmentation

10.2.2 Reporter ion intensities

10.2.3 iTRAQ 8-plex

10.3 TMT - Tandem Mass Tag

10.4 Reporter-based quantification runs

10.5 Identification and quantification

10.6 Peptide table

10.7 Reporter-based quantification experiments

10.7.1 Normalization across LC-MS/MS runs - use of a reference

sample

10.7.2 Normalizing within an LC-MS/MS run

10.7.3 From reporter intensities to protein abundances

10.7.4 Finding differentially abundant proteins

10.7.5 Distributing the replicates on the quantification runs

10.7.6 Protocols

10.8 Exercises

10.9 Bibliographic notes =

11 Fragment-based MS/MS Quantification

11.1 The label masses

11.2 Identification

11.3 Peptide and protein quantification

11.4 Exercises .

11.5 Bibliographic notes

12 Label-based Quantification by MS-spectra

12.1 Different labeling techniques

12.1.1 Metabolic labeling - SILAC

12.1.2 Chemical labeling

12.1.3 Enzymatic labeling - 18O

12.2 Experimental setup

12.3 MaxQuant as a model

12.3.1 HL-pairs .

12.3.2 Reliability of HL-pairs

12.3.3 Reliable protein results

12.4 The MaxQuant procedure

12.4.1 Recognize HL-pairs

12.4.2 Estimate HL-ratios

12.4.3 Identify HL-pairs by database search

12.4.4 Infer protein data

12.5 Exercises

12.6 Bibliographic notes

13 Label-free Quantification by MS spectra

13.1 An ideal case - two protein samples

13.2 The real world

13.2.1 Multiple samples

13.3 Experimental setup

13.4 Features

13.5 The quantification process

13.6 Feature detection

13.7 Pairwise retention time corrections

13.7.1 Determining potentially corresponding features

13.7.2 Linear corrections

13.7.3 Nonlinear corrections

13.8 Approaches for feature-tuple detection

13.9 Pairwise alignment

13.9.1 Finding an optimal alignment

13.10Using a reference run for alignment

13.11Complete pairwise alignment

13.12Hierarchical progressive alignment

13.12.1Measuring the similarity or the distance of two runs

13.12.2Constructing static guide trees

13.12.3Constructing dynamic guide trees

13.12.4Aligning subalignments

13.12.5SuperHirn

13.13Simultaneous iterative alignment

13.13.1Constructing the initial alignment in XCMS .

13.13.2Changing the initial alignment

13.14The end result and further analysis

13.15Exercises

13.16Bibliographic notes

14 Label-free Quantification by MS/MS spectra

14.1 Abundance measurements

14.2 Normalization

14.3 Proposed methods

14.4 Methods for single abundance calculation

14.4.1 emPAI

14.4.2 PMSS

14.4.3 SI

14.5 Methods for relative abundance calculation

14.6 Comparing methods

14.6.1 An analysis by Griffin

14.6.2 An analysis by Colaert

14.7 Improving the reliability of spectral count quantification

14.8 Handling shared peptides

14.9 Statistical analysis

14.10Exercises

14.11Bibliographic notes

15 Targeted Quantification - Selected Reaction Monitoring

15.1 Selected Reaction Monitoring - the concept

15.2 A suitable instrument

15.3 The LC-MS/MS run

15.3.1 Sensitivity and accuracy

15.4 Label-free and label-based quantification

15.4.1 Label-free SRM-based quantification

15.4.2 Label-based SRM-based quantification

15.5 Requirements for SRM transitions

15.5.1 Requirements for the peptides

15.5.2 Requirements for the fragment ions

15.6 Finding optimal transitions

15.7 Validating transitions

15.7.1 Testing linearity

15.7.2 Determining retention time

15.7.3 Limit of detection/quantification

15.7.4 Dealing with low abundant proteins

15.7.5 Checking for interference

15.8 Assay development

15.9 Exercises

15.10Bibliographic notes

16 Absolute Quantification

16.1 Performing absolute quantification

16.1.1 Linear dependency between the calculated and the real

abundances

16.2 Label-based absolute quantification

16.2.1 Stable isotope-labeled peptide standards

16.2.2 Stable isotope-labeled concatenated peptide standards

16.2.3 Stable isotope-labeled intact protein standards

16.3 Label-free absolute quantification

16.3.1 Quantification by MS spectra

16.3.2 Quantification by the number of MS/MS spectra

16.4 Exercises

16.5 Bibliographic notes

17 Quantification of Posttranslational Modifications

17.1 PTM and mass spectrometry

17.2 Modification degree

17.3 Absolute modification degree

17.3.1 Reversing the modification

17.3.2 Use of two standards

17.3.3 Label-free modification degree analysis

17.4 Relative modification degree

17.5 Discovery-based modification stoichiometry

17.5.1 Separate LC-MS/MS experiments for modified and unmodified

peptides

17.5.2 Common LC-MS/MS experiment for modified and unmodified

peptides

17.5.3 Reliable results and significant differences

17.6 Exercises

17.7 Bibliographic notes

18 Biomarkers

18.1 Evaluation of potential biomarkers

18.1.1 Taking disease prevalence into account

18.2 Evaluating threshold values for biomarkers

18.3 Exercises

18.4 Bibliographic notes

19 Standards and Databases

19.1 Standard data formats for (quantitative) proteomics

19.1.1 Controlled vocabularies (CVs)

19.1.2 Benefits of using CV terms to annotate metadata

19.1.3 A standard for quantitative proteomics data .

19.1.4 HUPO PSI

19.2 Databases for proteomics data

19.3 Bibliographic notes

20 Appendix A: Statistics

20.1 Samples, populations and statistics

20.2 Population parameter estimation

20.2.1 Estimating the mean of a population

20.3 Hypothesis testing

20.3.1 Two types of errors

20.4 Performing the test - test statistics and p-values

20.4.1 Parametric test statistics

20.4.2 Nonparametric test statistics

20.4.3 Confidence intervals and hypothesis testing .

20.5 Comparing means of populations

20.5.1 Analyzing the mean of a single population

20.5.2 Comparing the means from two populations

20.5.3 Comparing means of paired populations

20.5.4 Multiple populations

20.5.5 Multiple testing

20.6 Comparing variances

20.6.1 Testing the variance of a single population

20.6.2 Testing the variances of two populations

20.7 Percentiles and quantiles

20.7.1 A straightforward method for estimating the percentiles

20.7.2 Quantiles

20.7.3 Box plots

20.8 Correlation

20.8.1 Pearson’s product-moment correlation-coefficient

20.8.2 Spearman’s rank correlation coefficient

20.9 Regression analysis

20.9.1 Regression line

20.9.2 Relation between Pearson’s correlation coefficient and

the regression parameters

20.10Types of values and variables

21 Appendix B: Clustering and Discriminant Analysis

21.1 Clustering

21.1.1 Distances and similarities

21.1.2 Distance measures

21.1.3 Similarity measures

21.1.4 Distances between an object and a class

21.1.5 Distances between two classes

21.1.6 Missing data .

21.1.7 Clustering approaches

21.1.8 Sequential clustering

21.1.9 Hierarchical clustering

21.2 Discriminant analysis

21.2.1 Stepwise feature selection

21.2.2 Linear discriminant analysis using original features

21.2.3 Canonical discriminant analysis

21.3 Bibliographic notes

Bibliography

Index