Integrative statistical methods for decoding molecular responses to insect herbivory in Nicotiana attenuata

Gulati, Jyotasana GND

This work focuses on the development of statistical methods to select features (genes and metabolites) exhibiting induced local and systemic defense responses to insect attack in Nicotiana attenuata along with the extraction of additional information regarding their timing of action. To characterize the dynamics of activation in time and space of herbivory-induced responses, I designed a framework by combining methods previously developed for feature selection and extraction to identify activated network motifs. These motifs are the set of features that are differentially perturbed in local and systemic tissues in response to herbivory. The extraction of multifactorial statistical information in terms of time response variable simultaneously captured the dynamic response of a gene/metabolite in more than one tissue and therefore helped in identifying tissue-specific activation of biochemical pathways during herbivory, their transition points and shared patterns of regulation with other physiological processes and gene-metabolite interactions at the level of isolated motifs. I utilized this framework to evaluate the transcriptional and metabolic dynamics in the roots to investigate their role in aboveground stress responses. I discovered an emergent property of an inversion in root-specific semidiurnal (12h) rhythms in response to simulated leaf herbivory. In addition, I illustrated the benefits of our statistical framework, used for generating spatio-temporally resolved transcriptional/metabolic maps, by visualizing the chronology of the activation of pathways central to signaling, tolerance and defense in N. attenuata. The research described in this thesis, in addition to being valuable in deciphering dynamic responses to insect attack in a whole plant context, lays the foundation for future analyses in which statistical modeling of these networks assisted with experimental data could predict the logical rules governing these dynamic interactions.


Citation style:
Could not load citation form.