The Use of Principal Component Analysis in MALDI-TOF MS - a Powerful Tool for Establishing a Mini-optimized Proteomic Profile

Abstract

Background Recently, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) technology has been applied to the exploration of biomarkers for early cancer diagnosis, but more effort is required to identify a single sensitive and specific biomarker. For early diagnosis, a proteomic profile is the gold standard, but inconvenient for clinical use since the profile peaks are quantitative. It would therefore be helpful to find a minimized profile, comprising fewer peaks than the original using an existing algorithm and compare it with other traditional statistical methods. Methods In the present study, principal component analysis (PCA) in the ClinProt-Tools of MALDI-TOF MS was used to establish a mini-optimized proteomic profile from gastric cancer patients and healthy controls, and the result was compared with t-test and Flexanalysis software. Results Eight peaks were selected as the mini-optimized proteomic profile to help differentiate between gastric cancer patients and healthy controls. The peaks at m/z 4212 were regarded as the most important peak by the PCA algorithm. The peaks at m/z 1866 and 2863 were identified as deriving from complement component C3 and apolipoprotein A1, respectively. Conclusions PCA enabled us to identify a mini-optimized profile consisting of significantly differentiating peaks and offered the clue for further research.

Intracellular proteome expression during 4-n-nonylphenol biodegradation by the filamentous fungus Metarhizium robertsii

“…There are increasing number of publications that use PCA for complex proteome analyses in different species and different data types using PC loadings (Verhoeckx et al 2004;Rao and Li 2009;Shao et al 2012). BLAST searches did not result in any reasonable results for the unidentified proteins; therefore, a principal components analysis (PCA) was employed to propose a role for the unknown proteins ( fig.…”

Biomarker Characterization by MALDI–TOF/MS

“…Combined with MALDI-TOF/MS, each mass spectrum is represented by a point in 2D or 3D PCA plots, where mass spectra with similar variation characteristics are clustered. This approach has also been applied in many research areas including food safety, metabolomics, proteomics, and IMS [188][189][190][191][192][193][194][195]. Hierarchical clustering analysis (HCA), a grouping method that functions by creating a cluster tree or dendrogram, is another multivariate analysis method [196].…”

Comparative proteomic analysis of membrane microdomains isolated from two hyperlipidemic animal models

“…In addition, Principal Component Analysis (PCA) algorithm was applied within Sieve 2.1. Thus, a reduction in dimension of the different sets of data corroborated with quantitative spatial profile determination [21,22] was performed for both the mice and hamster samples. The 3D representations revealed a distinct spatial scattering relative to their respective controls (N, C) of the [4] hyperlipidemic and statin treated hyperlipidemic sets of DRMs isolated from the Golden Syrian hamsters and ApoE deficient mice respectively.…”