Integration of two-dimensional LC-MS with multivariate statistics for comparative analysis of proteomic samples
van der Greef, J.
TNO Kwaliteit van Leven
LC-MS-based proteomics requires methods with high peak capacity and a high degree of automation, integrated with data-handling tools able to cope with the massive data produced and able to quantitatively compare them. This paper describes an off-line two-dimensional (2D) LC-MS method and its integration with software tools for data preprocessing and multivariate statistical analysis. The 2D LC-MS method was optimized in order to minimize peptide loss prior to sample injection and during the collection step after the first LC dimension, thus minimizing errors from off-column sample handling. The second dimension was run in fully automated mode, injecting onto a nanoscale LC-MS system a series of more than 100 samples, representing fractions collected in the first dimension (8 fractions/sample). As a model study, the method was applied to finding biomarkers for the antiinflammatory properties of zilpaterol, which are coupled to the β2-adrenergic receptor. Secreted proteomes from U937 macrophages exposed to lipopolysaccharide in the presence or absence of propanolol or zilpaterol were analysed. Multivariate statistical analysis of 2D LC-MS data, based on principal component analysis, and subsequent targeted LC-MS/MS identification of peptides of interest demonstrated the applicability of the approach. © 2006 American Chemical Society.
To reference this document use:
Computer aided software engineering
Multivariate statistical analysis
beta 2 adrenergic receptor
principal component analysis
Amino Acid Sequence
Molecular Sequence Data
Principal Component Analysis
Receptors, Adrenergic, beta-2
Reproducibility of Results
Analytical Chemistry, 78 (7), 2286-2296