Identification of Marker Using Principal Component Analysis

PCA, factorial analysis and cluster analysis are important and proven techniques for complex data analysis (Li et al. 2009; Rawal et al. 2010). DART-MS fingerprinting in combination with data reduction technique like PCA is a powerful tool for distinguishing and identifying phytoconstituents as markers in foods and medicines. PCA was selected for dimensionality reduction in an attempt to distinguish the characteristic profiles from the DART-MS data and to identify marker phytochemicals that aided in grouping of samples. PCA is an unsupervised procedure that determines the directions of the largest variations in the dataset and the data are generally presented as a two-dimensional plot (score plot) where the coordinate axis represents the directions of the two largest variations. The DART-MS data of P. nigrum, P. chaba and P. longum plant parts were subjected to PCA. The 15 peaks extracted from PCA (m/z. 169, 195. 224, 238, 272, 289, 302, 312, 371, 447, 461, 559. 571. 557 and 597) were able to discriminate among the species and in between fruit, leaf and root part as shown in the score plot (PCI vs. PC2) in Figures 2.2 and 2.3, respectively. The score plot of the fruits showed clustering of the data according to the species. Similar clustering and differentiation is also clearly seen in root and leaf. Different parts of the same plant also showed similar clustering. DART-MS followed by PCA seems to be an appropriate method for easy differentiation of species as well as plant parts.

The PCs were able to explain 68.81% of the variance. The PCI vs. PC2 plot shows a distinctive discrimination between the fruits of P. nigrum, P. longum and P. chaba. The samples of P. nigrum are falling in the second quadrant having negative PC score with the main contribution from m/z. 571. The positive PC2 score is mainly due to peaks at m/z. 312 and m/z. 597.

The samples of P. longum are along the X-axis, where the main contribution to PCI score is coming from m/z. 224, m/z. 447 and m/z. 289. There is no contribution of m/z. 571 in P. longum. The P. chaba is characterized by the negative scores of both PCI and PC2. The dominance in P. chaba is mainly due to m/z 571 and to some extent m/z, 312. The PCI and PC2 scores in both are in the range from-1 to -3. But the discrimination between the two is mainly because its contribution in P. chaba is higher than that in P. nigrum. Peaks at m/z 169, m/z 195, m/z 224, m/z. 238, m/z 289, m/z 302 and m/z. 447 observed in DART mass spectrum of P. longum were contributing more indiscrimination from other Piper species, i.e., P. nigrum and P. chaba.

PC1 vs. PC2 plot shows a distinctive discrimination between the fruits P

FIGURE 2.2 PC1 vs. PC2 plot shows a distinctive discrimination between the fruits P. nigrum, P. chaba and P. longum on the basis of 15 peaks (169, 195, 224, 238, 272, 289, 302, 312, 371, 447, 461, 559, 571, 557 and 597). (Reproduced from Ref. Chandra et al. 2014 with permission from Royal Society of Chemistry.)

PCA plot of P. nigrum plant parts discrimination on the basis of nine peaks

FIGURE 2.3 PCA plot of P. nigrum plant parts discrimination on the basis of nine peaks (151.12, 224.22, 236.23, 250.15, 286.2, 447.44, 509.45, 571.42 and 597.41). (Reproduced from Ref. Chandra et al. 2014 with permission from Royal Society of Chemistry.)

Quantitative Determination of Chemical Constituents of Piper Species Using UPLC- ESI-MS/MS

 
Source
< Prev   CONTENTS   Source   Next >