Fig. 1

Computed orthogonal partial least squares discriminant analysis (OPLS-DA) and partial least squares discriminant analysis (PLS-DA) models. (A) and (B) are computed OPLS-DA models showing the separated metabolic features of the control (red) and Pst-infected (green) samples of Koonap and Gariep wheat cultivars, respectively. (C) and (D) further show time-dependent metabolic changes in both control (control in blue demarcation) and infected (SRT in red demarcation) samples, starting from week 1 (W1) post-inoculation to week 4 (W4) post-inoculation, left to right. The OPLS-DA and PLS-DA models show differential separation of data based on pathogen infection as well as plant development over time as an indication of pathogen-induced metabolic reprogramming. The data projected above were median-normalised, log transformed and Pareto-scaled in MetaboAnalyst for correlation and predictability scores of R2 = 0.921 and Q2 = 0.600 (A); R2 = 0.609 and Q2 = 0.256 (B); R2 = 0.944 and Q2 = 0.702 (C); R2 = 0.890 and Q2 = 0.528 (D)