Hedjazi Lyamine, Gauguier Dominique, Zalloua Pierre A, Nicholson Jeremy K, Dumas Marc-Emmanuel, Cazier Jean-Baptiste
⊥School of Medicine, Lebanese American University, Beirut 1102 2801, Lebanon.
‡Section of Biomolecular Medicine, Division of Computational and Systems Medicine, Department of Surgery and Cancer, Faculty of Medicine, Imperial College London, Sir Alexander Fleming building, London SW7 2AZ, U.K.
Anal Chem. 2015 Apr 21;87(8):4377-84. doi: 10.1021/acs.analchem.5b00145. Epub 2015 Apr 2.
High-throughput (1)H nuclear magnetic resonance (NMR) is an increasingly popular robust approach for qualitative and quantitative metabolic profiling, which can be used in conjunction with genomic techniques to discover novel genetic associations through metabotype quantitative trait locus (mQTL) mapping. There is therefore a crucial necessity to develop specialized tools for an accurate detection and unbiased interpretability of the genetically determined metabolic signals. Here we introduce and implement a combined chemoinformatic approach for objective and systematic analysis of untargeted (1)H NMR-based metabolic profiles in quantitative genetic contexts. The R/Bioconductor mQTL.NMR package was designed to (i) perform a series of preprocessing steps restoring spectral dependency in collinear NMR data sets to reduce the multiple testing burden, (ii) carry out robust and accurate mQTL mapping in human cohorts as well as in rodent models, (iii) statistically enhance structural assignment of genetically determined metabolites, and (iv) illustrate results with a series of visualization tools. Built-in flexibility and implementation in the powerful R/Bioconductor framework allow key preprocessing steps such as peak alignment, normalization, or dimensionality reduction to be tailored to specific problems. The mQTL.NMR package is freely available with its source code through the Comprehensive R/Bioconductor repository and its own website ( http://www.ican-institute.org/tools/ ). It represents a significant advance to facilitate untargeted metabolomic data processing and quantitative analysis and their genetic mapping.
高通量¹H核磁共振（NMR）是一种用于定性和定量代谢谱分析的日益流行的可靠方法，它可与基因组技术结合使用，通过代谢型数量性状位点（mQTL）定位来发现新的基因关联。因此，迫切需要开发专门工具，以准确检测和无偏解释由基因决定的代谢信号。在此，我们介绍并实施一种组合化学信息学方法，用于在定量遗传背景下对基于¹H NMR的非靶向代谢谱进行客观和系统的分析。R/Bioconductor的mQTL.NMR软件包旨在：（i）执行一系列预处理步骤，恢复共线NMR数据集中的光谱依赖性，以减轻多重检验负担；（ii）在人类队列以及啮齿动物模型中进行稳健且准确的mQTL定位；（iii）从统计学上增强由基因决定的代谢物的结构归属；（iv）用一系列可视化工具展示结果。在强大的R/Bioconductor框架中内置的灵活性和实现方式，使诸如峰对齐、归一化或降维等关键预处理步骤能够针对特定问题进行定制。mQTL.NMR软件包可通过综合R/Bioconductor存储库及其自身网站（http://www.ican-institute.org/tools/ ）免费获取其源代码。它代表了在促进非靶向代谢组学数据处理、定量分析及其遗传定位方面的一项重大进展。
