3.3 | Metabolomic comparison of fenced and unfencedN. sylvatica plants
Comparison of the global metabolome of fenced and unfenced N. sylvatica plants resulted in the identification of 1,025 metabolite features. To assess the overall metabolomic accumulation pattern of the samples from the fenced and open areas, PCA was conducted using the 1025 metabolite features, which clustered the N. sylvatica samples into two separate groups that correspond with the treatment (presence or absence of fence); the first two principal components (PC1 and PC2) explained 47.5% of the total variability (Figure 4A ). The top 15 metabolite features that contributed to the separation of the samples into the two distinct clusters were identified using PLS-DA (Figure 4B ), and their accumulation pattern varied based on treatment (fence or no fence) (Figure 4C ). Hierarchical cluster analysis (HCA) was conducted to identify the 25 top significant metabolite features and to cluster the samples based on the top features (Figure 4D ). The PLS-DA (Figure 4C ) and HCA (Figure 4D ) analyses identified similar metabolite features whose accumulation patterns were consistent with the presence/absence of fences (Figure 4D ).
We obtained putative prediction on the chemical identity of the 1,025 metabolites and their associated metabolic pathways using the GSEA functional analysis tool. Among the identified metabolic pathways are the pentose phosphate pathway (P = 0.01), starch and sucrose metabolism (P = 0.02), pyrimidine metabolism (P = 0.054), cyanoamino acid metabolism (P = 0.054), riboflavin metabolism (P = 0.054) and monoterpenoid biosynthesis (P = 0.058) (Appendix Table 3 ).
Out of the identified metabolic pathways, the monoterpenoid biosynthetic pathway produces metabolites that mediate indirect defenses in many plant species , while the cyanoamino acid pathway is implicated in detoxification .