StackEPI

StackEPI predicts cell line-specific enhancer-promoter interactions to identify gene regulatory relationships from enhancer and promoter sequences.

Key Features:

• Cell Line-Specific Prediction: Predicts enhancer-promoter interactions in specific human cell lines to capture cell-type regulatory contexts.
• Stacking Ensemble Learning Strategy: Uses a stacking ensemble learning approach that combines multiple machine learning models to improve prediction accuracy and reduce training time.
• Comprehensive Information Extraction: Integrates diverse encoding schemes to extract features from enhancer and promoter sequences.
• Improved Performance and Speed: Demonstrates higher prediction performance and faster computation than other state-of-the-art models in experimental evaluations.

Scientific Applications:

• Gene Regulation Studies: Identifies cell line-specific EPIs to support elucidation of regulatory mechanisms underlying gene expression.
• Cell Differentiation Research: Enables analysis of interactions specific to different cell lines to inform studies of cellular differentiation processes.
• Comparative Genomics and Epigenetics: Facilitates comparison of EPIs across cell types to investigate epigenetic regulation and cell-type-specific regulatory networks.

Methodology:

Implements a stacking ensemble learning framework that combines multiple machine learning models, applies diverse sequence encoding schemes to enhancer and promoter sequences for feature extraction, and benchmarks prediction performance against state-of-the-art models.