Statistical analyses were performed using the programming software R 3.4.4 (R Core Team, 2018). Principal component analysis (PCA) using samples as individuals and based on all the quantification values for each protein was performed in order to identify the main factors explaining the differences between samples. PCA was computed using Z-score transformed values and established from the correlation matrix.

At the individual protein level, analysis of the explanatory factors for wheat protein abundance variations was carried out, after verification of the absence of any block/repeat effect, using a nested analysis of variance (ANOVA) test based on the following linear model:

where Yijkl refers to the individual values, μ is the general mean of the variable considered, Cvi is the effect of the wheat cultivar (i.e., cv. Recital, cv. Cadenza, and cv. Renan), Tj is the effect of the treatment (i.e., F. graminearum-inoculated or water-inoculated), Cvi × Tj is the interaction of the cultivar effect by the treatment, Cvi × Tj {Sk} is the effect attributable to the interaction of the two main factors (Cvi × Tj) taking into account the inoculated fungal strain (Sk) as a nested factor in the main treatment factor (Tj), and εijkl is the residual.

For each individual wheat protein, the p values obtained for each effect (Cvi, Tj, Cvi × Tj, and Cvi × Tj {Sk}) were adjusted to control the FDR for independent test statistics (Benjamini and Hochberg, 1995). Only proteins with an FDR < 0.01 corresponding to p values < 0.00026, < 0.00045, < 0.00002, and < 0.000003 were deemed significant for the Cvi, Tj, Cvi × Tj, and Cvi × Tj {Sk} effects, respectively. Following the methodology described by Kumar and Futschik (2007), fuzzy C-means clustering of wheat proteins showing significant abundance changes according to each effect tested was performed from Z-score transformed values and a fuzzification parameter of 2, with the exception of the Cv × T{S}_effect proteins for which a hierarchical clustering was realized using the Euclidean distance as dissimilarity metric and the Ward’s method as aggregation criteria.

Based on the results of the ANOVA, regularized generalized canonical correlation analysis (rCCA) were used to assess the canonical relationships between the abundance changes of wheat proteins and the accumulation of fungal effector proteins that are supposed to putatively control host biological processes. This rCCA was computed from all wheat proteins harboring an interaction effect of the two main factors (i.e., Cv × T_effect and Cv × T{S}_effect proteins) and the F. graminearum putative effectors identified from the same biological samples and described in Fabre et al. (2019a). More specifically, all F. graminearum proteins displaying abundance patterns significantly impacted by the host cultivar and/or the fungal genetics have been primarily selected (Fabre et al., 2019a). Among these, only fungal proteins predicted as effector using EffectorP2.0 (Sperschneider et al., 2018) and/or secreted according to the predicted F. graminearum secretome described in Brown et al. (2012) were chosen. These structural features used to select the fungal proteins have been extracted from Fabre et al. (2019a) and are provided in Supplementary Table 1. The rCCA was performed following the methodology described in Gonzalez et al. (2008) and using the mixOMICS r-package v5.2 (Rohart et al., 2017). Since the number of subjects was lower than the number of variables in both datasets, the regularization parameters λ1 and λ2 estimated following the methodology described in Budzinski et al. (2019) were used for the covariance matrices X and Y. The pairwise association matrix was computed for the first eight dimensions, and all canonical correlations have been plotted using the network function with a threshold set to 0.95.

Note: The content above has been extracted from a research article, so it may not display correctly.



Q&A
Please log in to submit your questions online.
Your question will be posted on the Bio-101 website. We will send your questions to the authors of this protocol and Bio-protocol community members who are experienced with this method. you will be informed using the email address associated with your Bio-protocol account.



We use cookies on this site to enhance your user experience. By using our website, you are agreeing to allow the storage of cookies on your computer.