3.5.Principal component analysis (PCA) of carbon source metabolization
The differences in metabolic characteristics of the intercropping system were directly reflected by the position of the points in the principal vector space. As shown in Figure 3, the multivariate vectors were transformed into two uncorrelated principal component vectors; PC1 accounted for 55.64% of the total variability and PC2 for 13.24%. For the intercropping system, the point positions of MN0, MNE, M0, ME, AN0, ANE, AE, and A0 were significantly different.
The sample distribution along the PC axis was relevant to the carbon source use. Based on the results of the PCA, 17 carbon sources constituted PC1, including six amino acids, three carbohydrates, three carboxylic acids, two polymers, two miscellaneous substances, and one amide. Only four carbon sources accounted for PC2, including two carboxyl acids, one polymer, and one carbohydrate. The carbon sources used by mulberry microbes were greatly related to PC1, including amino acids, carbohydrates, and carboxylic acid, while the alfalfa community mainly used carbon sources related to PC2.