Optimizing Metabonomic Spectral Replacement

TL;DR

This paper introduces a new protocol for spectral replacement in metabonomics that improves the accuracy of signal estimation, addressing issues of overestimation caused by previous scaling methods.

Contribution

A novel spectral replacement protocol that accurately estimates replaced regions, enhancing analysis of complex metabonomic spectral data.

Findings

Improved accuracy in spectral signal replacement

Reduction of overestimation in signal intensity

Enhanced reliability of metabonomic data analysis

Abstract

Metabonomics, the measure of the fingerprint of biochemical perturbations caused by disease, drugs or toxins, recently has become a major focus of research in various areas especially indications of drug toxicity. Two types of technology (known by the initials NMR and MS) are employed and both produce massive data in form of spectra. Sophisticated statistical models, known as pattern recognition techniques, are commonly applied for summarizing and analyzing these multidimensional data. However, strong signals from compounds that are administered during toxicological trials interfere with these models. So called 'spectral replacement' is a method to eliminate these signals by replacing them with the signals in their corresponding regions in control spectrum. The replaced regions are subsequently scaled. However, this scaling is not accurately measured and often results in overestimation of integrated intensity of the replaced signals. Here, a novel protocol is proposed which provides an accurate estimation of the replaced regions.