A Deep Learning Approach for Modelling the Complex Relationship between Environmental Factors and Biological Features

Environmental factors play a pivotal role in shaping the genetic and phenotypic diversity among organisms. Understanding the influence of the environment on a biological phenomenon is essential for deciphering the mechanisms resulting in trait differences among organisms. In this study, we present a novel approach utilizing an Artificial Neural Network (ANN) model to investigate the impact of environmental factors on a wide range of biological phenomena. Our proposed workflow includes hyperparameter optimization using model-based methods such as Bayesian and direct-search methods such as Random Search, and a new approach combining random search and linear models (RandomSearch+lm) to ensure a robust ANN architecture. Moreover, we employed a generalized version of the variable importance method to generate the feature importance metric using estimated weights from ANN. By applying this comprehensive ANN-based approach to functional genomics, we can gain valuable insights into the mechanisms underlying trait differentiation in organisms, while simultaneously enabling prediction and feature selection tasks. This methodology provides a robust and efficient framework for studying the complex relationships between environmental factors and biological features in biological systems.