Statistical Analysis of NMR Metabolic Fingerprints: Established Methods and Recent Advances

Abstract

In this review, we summarize established and recent bioinformatic and statistical methods for the analysis of NMR-based metabolomics. Data analysis of NMR metabolic fingerprints exhibits several challenges, including unwanted biases, high dimensionality, and typically low sample numbers. Common analysis tasks comprise the identification of differential metabolites and the classification of specimens. However, analysis results strongly depend on the preprocessing of the data, and there is no consensus yet on how to remove unwanted biases and experimental variance prior to statistical analysis. Here, we first review established and new preprocessing protocols and illustrate their pros and cons, including different data normalizations and transformations. Second, we give a brief overview of state-of-the-art statistical analysis in NMR-based metabolomics. Finally, we discuss a recent development in statistical data analysis, where data normalization becomes obsolete. This method, called zero-sum regression, builds metabolite signatures whose estimation as well as predictions are independent of prior normalization.

NMR-Based Metabolomics in Metal-Based Drug Research

“…Each PC is a linear combination of the original data (the NMR variables), explaining the maximum amount of variance, not accounted by the previous PCs. Conversion of the spectral data to PCs results in two matrices, known as scores (corresponding to samples) and loadings (corresponding to NMR variables) [52]. By this unsupervised approach, it is possible to describe the general trend of samples and their potential clustering, depending on specific loadings (spectral variables) that may cause any cluster separation.…”

“…The further-supervised multivariate data analytical tools, such as projection to latent structure (PLS) and orthogonal projection to latent structure (OPLS), are tightly focused on the effects of interest, using any sample of a priori knowledge (e.g., class membership). Specifically, these methods relate the data matrix containing independent variables, such as spectral intensity values (defined as the X matrix), to a matrix containing dependent variables (e.g., measurements of response, such as toxicity or drug treatment) for those samples (Y matrix) [52,53]. This last (Y matrix) can be defined by continuous (pH values or the time course of the progression of a pathological effect), or discrete variables (group identity, health condition).…”

PLS2 in Metabolomics

“…Data pre-processing includes all the procedures that allow the transformation of raw data into a data table. With regards to NMR data, the procedures most commonly used for data pre-processing are: phasing, baseline correction, peak-picking, spectra alignment and binning [34,35]. For GC/U(H)PLC-MS data, data extraction is the tricky step [36].…”

Toxicogenomics and New Advances in Toxicology

“…Örnekler çözücü ekstraksiyonundan sonra, bozulmamış dokular (magic angle spinning NMR), sıvı veya yarı katı materyaller (NMR ve FT-IR) veya kuru materyaller (FT-IR) analizleri gerçekleştirilmektedir. 2,20 Metabolomikler, H-NMR (proton NMR) veya bir kromatografik seperasyonla (LC veya GC) ikili olarak, MS tayin teknolojilerini kullanmaktadır. NMR spektroskopisi, analitlerin ayrılmasına dayanmayan ve yıkıcı olmayan tek tespit tekniğidir.…”

“…Ayrıca, ilgili metabolitlerin tanımlanması bir sorun teşkil edebilmektedir. 1,2,20 Metabolik görüntülemede sıklıkla kullanılan diğer teknolojiler ise LC-MS veya GC-MS'dir. Kompleks örnekler sıvı veya gaz kromatografisiyle ayrılmakta ve her fraksiyonun bileşenlerinin kütleleri MS ile belirlenmektedir.…”