Integrating machine learning and mechanistic modeling towards adaptation of plants to climate change

This repository contains data, pre-processing scripts, evaluation standards, and results from my collaborative MSc thesis conducted between the School of Engineering at the University of Minho and the Bioinformatics Group at Wageningen University & Research (WUR). Our research focused on the integration of Machine Learning and mechanistic modelling techniques to predicting gene regulatory interactions associated with Arabidopsis' response to ambient temperature.

During this study, we evaluated three tools: BIONIC (Biological Network Integration using Convolutions), PyWGCNA (Python-based Weighted Gene Correlation Network Analysis), and BADDADAN (currently under development by Ben Noordijk at WUR). Among them, BIONIC stands out as a Deep Learning framework with a Graph Convolutional Network architecture. This tool reportedly integrates biological networks and extracts features that have an improved performance at downstream tasks, like module detection.