Publicação
Integrating machine learning and mechanistic modeling towards adaptation of plants to climate change
| Resumo: | This dissertation focuses on the pressing issue of climate change and its impact on agricultural yield. Crop species, particularly polyploids, present challenges when investigating these effects. To overcome these challenges, our research centres on Arabidopsis thaliana, known for genes homologous to those linked to crop yield, found in various crops like wheat, rice, and tomato, contributing to similar traits and enhancing tolerance to adverse conditions. Furthermore, prior studies have revealed connections between these yield-associated genes and flowering genes in Arabidopsis. The primary aim of this study is to improve the predictive power of Arabidopsis’s flowering time in response to variations in ambient temperature by harnessing the potential of integrated machine learning and mechanistic modelling techniques. Synergizing machine learning’s data-driven pattern recognition with mechanistic modelling’s knowledge-based causal understanding enhances predictability and deepens our understanding of plant responses to temperature. To achieve this goal, we employ three methods: Biological Network Integration using Convolutions (BIONIC), Python-based Weighted Gene Correlation Network Analysis (PyWGCNA), and Bioinformatics Approach to Describe Dynamical Activations of a Dimension-reduced Arabidopsis Network (BADDADAN), for the robust detection of gene modules, the inference of relationships between module expression and relevant traits, as well as the mathematical formulation of found regulatory interactions. This endeavour is further strengthened by validating these relationships through Gene Set Enrichment Analysis (GSEA), ensuring a coherent functional grouping of modules within biologically relevant processes. Our research uncovers 14 modules that exhibit significant associations with temperature. Notably, five of these modules are intricately linked to the flowering process, while two modules exhibit dual associations with cold response and flowering regulation. This work deepens our understanding of how temperature influences the regulatory interactions within Arabidopsis, offering valuable insights into the influence of climate on plant development. |
|---|---|
| Autores principais: | Babo, Ana Camila Ribeiro de |
| Assunto: | Machine learning Mechanistic modelling Climate change Arabidopsis thaliana Temperature stress Flowering time Aprendizagem automática Modelação mecanística Alterações climáticas Stress térmico Tempo de floração |
| Ano: | 2023 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade do Minho |
| Idioma: | inglês |
| Origem: | RepositóriUM - Universidade do Minho |
| Resumo: | This dissertation focuses on the pressing issue of climate change and its impact on agricultural yield. Crop species, particularly polyploids, present challenges when investigating these effects. To overcome these challenges, our research centres on Arabidopsis thaliana, known for genes homologous to those linked to crop yield, found in various crops like wheat, rice, and tomato, contributing to similar traits and enhancing tolerance to adverse conditions. Furthermore, prior studies have revealed connections between these yield-associated genes and flowering genes in Arabidopsis. The primary aim of this study is to improve the predictive power of Arabidopsis’s flowering time in response to variations in ambient temperature by harnessing the potential of integrated machine learning and mechanistic modelling techniques. Synergizing machine learning’s data-driven pattern recognition with mechanistic modelling’s knowledge-based causal understanding enhances predictability and deepens our understanding of plant responses to temperature. To achieve this goal, we employ three methods: Biological Network Integration using Convolutions (BIONIC), Python-based Weighted Gene Correlation Network Analysis (PyWGCNA), and Bioinformatics Approach to Describe Dynamical Activations of a Dimension-reduced Arabidopsis Network (BADDADAN), for the robust detection of gene modules, the inference of relationships between module expression and relevant traits, as well as the mathematical formulation of found regulatory interactions. This endeavour is further strengthened by validating these relationships through Gene Set Enrichment Analysis (GSEA), ensuring a coherent functional grouping of modules within biologically relevant processes. Our research uncovers 14 modules that exhibit significant associations with temperature. Notably, five of these modules are intricately linked to the flowering process, while two modules exhibit dual associations with cold response and flowering regulation. This work deepens our understanding of how temperature influences the regulatory interactions within Arabidopsis, offering valuable insights into the influence of climate on plant development. |
|---|