Evaluating Double Descent in Machine Learning: Insights from Tree-Based Models Applied to a Genomic Prediction Task