在集成的工作环境中用创新的谱图数据分析法和NNM的多变量分析进行生物标志物鉴定

会议Poster

Biomarker Identification Combining Multivariate Analysis of NMR Spectra with an Innovative Spectral Data Analysis Approach
in an Integrated Working Environment

Abstract

An integrated informatics solution, compatible with multiple instrument formats, covering the full range  of steps from raw data processing to biomarker identification, has been long-anticipated by researchers performing metabonomics studies. Such a solution now exists, and we will demonstrate its successful application  to  the  analysis  of  1 H  NMR  spectra  of  human  serum  samples  from  37  diabetic  and  non-diabetic  subjects.  This  integrated  informatics  approach  also  includes  an  innovative  patent-pending
spectral Overlap Density Heatmap (ODH) tool.  By relating statistical findings to spectral observation in a supervised approach, ODH complements Principal Component Analysis (PCA) and offers meaningful direction for biomarker identification.

The integrated informatics approach to metabolomics research includes the following steps:
•  Batch processing of raw FIDs, phasing, baseline correction, and cross-spectral alignment.
•  Batch import of the processed spectra.
•  Include/exclude spectral ranges, optional binning and bucketing, pre-processing, Y-transforms, and PCA.
•  Visualization as an Overlap Density Heatmap, a novel method to quantitatively evaluate the similarity/dissimilarity among multiple overlaid spectra.
•  Comparative use of loadings plots resulting from PCA and peaks resulting from Overlap Density Heatmaps as search queries against a database of known metabolites.
•  Link to the KEGG database for metabolic pathways of identified metabolites.

Conclusions

This NMR-based metabolomics study demonstrates the potentials of utilizing a software platform that combines the full range of applications from raw data analysis  to  biomarker  identification  in  one  single  environment.  The  well-established  PCA  and  the  novel  Overlap  Density  Heatmap  technology  provided  within  the KnowItAll ®  Informatics System complete each other for the identification of key metabolites or biomarkers: ODH provides a visualization tool that “talks” to the chemist and allows direct examination of the data, either in a supervised or non-supervised way. Such an advanced tool can help define the best spectrum processing and data pre-processing options, and lead researchers to identify spectrum areas on which to focus in their analyses. PCA, on the other hand, provides easier automation and reproducibility, as well as easier access to time evolutions via the display of trajectories in toxicity studies.