G2P models: Predicting phenotypes from genomes

In this work, we train a set of tree models with genome-wide information, aiming to compare with models trained by integrating genotype and environment. In order to improve the prediction accuracy of these tree models, we use satcking ensemble learning method to integrate them into a strong learner.

basic models

• RF.pkl
• GBDT.pkl
• XGBoost.pkl
• lightGBM.pkl

stacking ensemble

• layer2model.pkl

Instructions for use

• These models are used in two places, one in the independent prediction part and the other in the model interpretation part. It should be noted that we have integrated the model-independent prediction into scripts Basic_model.cp38-win_amd64.pyd, and each training will generate a new set of models, The original saved models will be overwritten. In the interpretable part, you only need to call these models, please refer to script XAI.pkl for details.