easyPheno

easyPheno predicts phenotypes from genotypic data using diverse predictive models and evaluation procedures to support genomic selection and comparative model assessment.

Key Features:

• Model Diversity: Supports traditional genomic selection methods, machine learning algorithms, and deep learning models for phenotype prediction.
• Bayesian Optimization for Hyperparameter Tuning: Uses Bayesian optimization to automatically search hyperparameter spaces for predictive models.
• Extensibility and Integration: Provides a modular framework to integrate novel predictive models and extend computational workflows.
• Benchmarking Capabilities: Enables consistent benchmarking of new prediction models against existing methods within a unified experimental setup.
• Simulated Data Assessment: Allows evaluation of prediction models under predefined settings using simulated datasets.

Scientific Applications:

• Genomic Selection: Facilitates genomic selection by predicting complex traits from genotypic data.
• Personalized Medicine: Supports modeling genotype–phenotype relationships relevant to personalized medicine applications.
• Evolutionary Biology: Enables genotype-to-phenotype inference and trait prediction for evolutionary biology studies.
• Agricultural Genetics: Supports agricultural genetics through prediction and benchmarking of trait prediction models.

Methodology:

Implements Bayesian optimization for hyperparameter tuning, a modular architecture for integration of diverse predictive models, and evaluation using simulated data within unified benchmarking workflows, with analysis workflows spanning data preprocessing to result interpretation.