Spatial and Spectral Correlations in MALDI Mass Spectrometry Images by Clustering and Multivariate Analysis

详细信息    查看全文

• 作者：Gregor McCombie ; Dieter Staab ; Markus Stoeckli ; and Richard Knochenmuss
• 刊名：Analytical Chemistry
• 出版年：2005
• 出版时间：October 1, 2005
• 年：2005
• 卷：77
• 期：19
• 页码：6118 - 6124
• 全文大小：981K
• 年卷期：v.77,no.19(October 1, 2005)
• ISSN：1520-6882

文摘

Matrix-assisted laser desorption/ionization mass spectrometry imaging is a technique for direct analysis oftissue sections without the use of molecular tags orcontrast agents. The combination of spatial and massresolution results in large and complex data sets thatrequire powerful and efficient analysis and interpretationtools. Conventional images, derived from a specific analytemass, do not identify the spatially localized correlationsbetween analytes that are latent in the data. A newapproach to find and visualize these correlations ispresented. Clustering methods are used to classify pixelsby spectral similarity, facilitating definition of distinctspatial regions. Principal component and discriminantanalyses are combined to comprehensively identify changesin the mass spectra between regions. Images are generated by projecting the spectra of each pixel on thediscriminant spectra; contrast is then a function of multiple correlated peaks.