Optimized UAV-Based Digital Phenotyping for Enhanced Breeding Efficiency in Megathyrsus maximus

Description

This dataset supports a research article on optimizing UAV-based phenotyping for a biparental population of Megathyrsus maximus. It includes trained machine learning models and phenotypic data obtained from both conventional field measurements and RGB image–derived digital features.

Steps to reproduce

Install required Python (python 3.13) packages: pip install scikit-learn pandas numpy joblib matplotlib seaborn scipy Testing Pre-trained Models: import joblib import pandas as pd import numpy as np Load a trained model # Load model (example: Random Forest for ENV1, DH trait, DAP=27, GSD=0.50) model = joblib.load('ENV1_DH/trained_models/27_0.50/RandomForest_27_0.50.pkl') Prepare your test data # Load test data with same features as training test_data = pd.read_csv('your_test_data.csv') # Remove non-feature columns (same as training) non_features = ["bloco", "gen", "d1", "d2", "date", "planta", "TDMY", "LDMY", "GMY", "DH", "location"] X_test = test_data.drop(columns=[col for col in non_features if col in test_data.columns]) # Ensure DAP and gsd match the model you loaded # For example, filter for DAP=27 and gsd=0.50 X_test = X_test[(X_test['DAP'] == 27) & (X_test['gsd'] == 0.50)] # Make Predictions predictions = model.predict(X_test) print("Predictions:", predictions) # Evaluate Performance from sklearn.metrics import mean_squared_error, r2_score from scipy.stats import pearsonr y_true = test_data.loc[X_test.index, 'DH'] # or GMY, TDMY, LDMY rmse = np.sqrt(mean_squared_error(y_true, predictions)) r2 = r2_score(y_true, predictions) pearson_r = pearsonr(y_true, predictions)[0] print(f"RMSE: {rmse:.4f}") print(f"R²: {r2:.4f}") print(f"Pearson r: {pearson_r:.4f}")

Institutions

• Universidade Estadual de Campinas Instituto de Biologia

  SP, Campinas
• EMBRAPA Centro Nacional de Pesquisas de Gado de Corte

  PB, Campina Grande
• Universidade Estadual de Campinas Centro de Biologia Molecular e Engenharia Genetica

  SP, Campinas

Categories

Funders

• Fundação de Amparo à Pesquisa do Estado de São Paulo

  São Paulo, Brazil

  Grant ID: 2023/06910-0
• Fundação de Amparo à Pesquisa do Estado de São Paulo

  São Paulo, Brazil

  Grant ID: 2019/03232–6
• Fundação de Amparo à Pesquisa do Estado de São Paulo

  São Paulo, Brazil

  Grant ID: 2022/04006-2

Licence