Improved quality of super(1)H NMR spectroscopic data for enhanced metabolic profiling of low molecular weight metabolites in human serum
NMR based metabolic profiling of blood samples in epidemiological studies can be used for molecular phenotyping and biomarker discovery. Often metabolic changes in blood are more subtle and demand a high quality spectrum especially when looking at low molecular weight compounds. In order to improve...
Saved in:
Published in: | Metabolomics Vol. 7; no. 2; pp. 270 - 277 |
---|---|
Main Authors: | , , , , |
Format: | Journal Article |
Language: | English |
Published: |
01-06-2011
|
Online Access: | Get full text |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Summary: | NMR based metabolic profiling of blood samples in epidemiological studies can be used for molecular phenotyping and biomarker discovery. Often metabolic changes in blood are more subtle and demand a high quality spectrum especially when looking at low molecular weight compounds. In order to improve super(1)H NMR spectroscopic data we compared different serum sample preparation methods. Application of phosphate buffer reduces chemical shift variation, enhances resolution of signal multiplicity, facilitates visual inspection of NMR spectra and annotation of signals compared to traditionally used saline. For analysis of low molecular weight compounds we found that standard 1D spectra of ultrafiltrated serum samples show enhanced spectral quality of small metabolites as compared to transverse relaxation edited spectra (also called Carr-Purcell-Meiboom-Gill, CPMG) spectra of unfiltered serum samples due to improved signal-to-noise ratio. Thus, NMR signals attributable to different amino acids and other small metabolites could readily be detected in spectra of ultrafiltrated serum, but remained invisible in the corresponding CPMG spectra. An OPLS model of fasting blood glucose showed an increase of Q super(2) when using spectra from ultrafiltrated serum (Q super(2) = 0.261) compared to using CPMG spectra (Q super(2) = 0.173). Similar results were observed for OPLS models of BMI (Q super(2) = 0.253 and Q super(2) = 0.216, respectively). Furthermore, a reduction in model dimensionality was observed when using ultrafiltrated serum data. In conclusion we recommend sample preparation of serum samples in phosphate buffer instead of saline. Ultrafiltration of serum samples prior to NMR analysis is beneficial especially for low concentrated small metabolites. |
---|---|
Bibliography: | ObjectType-Article-1 SourceType-Scholarly Journals-1 content type line 23 ObjectType-Feature-2 |
ISSN: | 1573-3882 1573-3890 |
DOI: | 10.1007/s11306-010-0248-1 |