Gamboa-Becerra Roberto, Ramírez-Chávez Enrique, Molina-Torres Jorge, Winkler Robert
Department of Biotechnology and Biochemistry, CINVESTAV Unidad Irapuato, Km. 9.6 Libramiento Norte Carr. Irapuato-León, 36821, Irapuato, Guanajuato, Mexico.
Anal Bioanal Chem. 2015 Jul;407(19):5673-84. doi: 10.1007/s00216-015-8744-9. Epub 2015 May 26.
In cartography, the combination of colour and contour lines is used to express a three-dimensional landscape on a two-dimensional map. We transferred this concept to the analysis of mass spectrometry imaging (MSI) data and developed a collection of R scripts for the efficient evaluation of .imzML archives in a four-step strategy: (1) calculation of the density distribution of mass-to-charge ratio (m/z) signals in the .imzML file and assembling of a pseudo-master spectrum with peak list, (2) automated generation of mass images for a defined scan range and subsequent visual inspection, (3) visualisation of individual ion distributions and export of relevant .mzML spectra and (4) creation of overlay graphics of ion images and photographies. The use of a Hue-Chroma-Luminance (HCL) colour model in MSI graphics takes into account the human perception for colours and supports the correct evaluation of signal intensities. Further, readers with colour blindness are supported. Contour maps promote the visual recognition of patterns in MSI data, which is particularly useful for noisy data sets. We demonstrate the scalability of MSI.R scripts by running them on different systems: on a personal computer, on Amazon Web Services (AWS) instances and on an institutional cluster. By implementing a parallel computing strategy, the execution speed for .imzML data scanning with image generation could be improved by more than an order of magnitude. Applying our MSI.R scripts ( http://www.bioprocess.org/MSI.R ) to low-temperature plasma (LTP)-MSI data shows the localisation of volatile and semi-volatile compounds in the cross-cut of a chilli (Capsicum annuum) fruit. The subsequent identification of compounds by gas and liquid chromatography coupled to mass spectrometry (GC-MS, LC-MS) proves that LTP-MSI enables the direct measurement of volatile organic compound (VOC) distributions from biological tissues.
在制图学中，颜色和等高线的组合用于在二维地图上表示三维景观。我们将这一概念应用于质谱成像（MSI）数据的分析，并开发了一组R脚本，以四步策略高效评估.imzML存档：（1）计算.imzML文件中质荷比（m/z）信号的密度分布，并使用峰列表组装伪主谱；（2）为定义的扫描范围自动生成质量图像并进行后续目视检查；（3）可视化单个离子分布并导出相关的.mzML谱；（4）创建离子图像和照片的叠加图形。在MSI图形中使用色相-色度-亮度（HCL）颜色模型考虑了人类对颜色的感知，并支持对信号强度的正确评估。此外，还为色盲读者提供了支持。等高线图有助于目视识别MSI数据中的模式，这对于噪声数据集尤为有用。我们通过在不同系统上运行MSI.R脚本展示了其可扩展性：在个人计算机上、在亚马逊网络服务（AWS）实例上以及在机构集群上。通过实施并行计算策略，生成图像时.imzML数据扫描的执行速度提高了一个多数量级。将我们的MSI.R脚本（http://www.bioprocess.org/MSI.R）应用于低温等离子体（LTP）-MSI数据，显示了辣椒（Capsicum annuum）果实横切面中挥发性和半挥发性化合物的定位。随后通过气相色谱和液相色谱与质谱联用（GC-MS、LC-MS）对化合物进行鉴定，证明LTP-MSI能够直接测量生物组织中挥发性有机化合物（VOC）的分布。
