Mass Spectrometry Metabolomic Data Handling for Biomarker Discovery - Proteomic and Metabolomic Approaches to Biomarker Discovery - page 414

CHAPTER
27

Mass Spectrometry Metabolomic Data
Handlin g for Biomarker D iscovery

Julien Boccard, Serge Rudaz
School of Pharmaceutical Sciences, University of Geneva, University of Lausanne, Switzerland

OUTLINE

Metabolomics for Biomarker Discovery 426 Data Modeling 434

Exploratory Analysis with Unsupervised
Methods

Mass Spectrometry-Based Metabolomics 426

434

Direct MS Methods 426

Principal Component Analysis 435

Hyphenated MS Methods 427

Cluster Analysis 435

Targeted versus Untargeted Strategies 428 Regression and Classi cation with Supervised

Methods 436

Data Handling 428

Partial Least Squares Regression 437

Data Preprocessing 429

Decision Trees 437

Centroid Data and Format Conversion 429

Kernel Methods 437

Resolution Tuning, Noise Filtering,
and Mass Features Detection

Other Supervised Methods 438

430

Conventional Statistical Analysis and ROC
Curves

430

Spectral Features Alignment

438

Data Pretreatment: Normalization
and Scaling 431 Model Validation 439

Software Packages 431 Statistical Validation 439

Metabolite Identi cation: Biological

Variable Selection 432

Validation 440

Selection by Modeling/Prediction 432

Data Redundancy/Correlation Suppression 433 Conclusions 440

Integration of Biological Knowledge 434

References 441