accuEnhancer

accuEnhancer predicts active enhancers by integrating DNase-seq data and primitive sequences with deep learning to identify H3K27ac-marked regulatory elements across cell types.

Key Features:

• Deep Learning Integration: Employs a deep learning model that uses DNase-seq data to predict H3K27ac peaks indicative of active enhancers.
• Cross-Cell Type Joint Training: Performs joint training across multiple cell types to capture complex regulatory patterns and improve prediction accuracy in unstudied cell types.
• Data-Driven Predictions: Uses DNase data and primitive sequences as input features to predict enhancer activities, emphasizing chromatin accessibility and sequence information.
• Application to Novel Cell Types: Predicts active enhancers in cell types lacking H3K27ac modification data by leveraging information from other cell types.

Scientific Applications:

• Gene Expression Regulation Studies: Enables exploration of how enhancers influence gene expression across different cell types, aiding studies of cellular differentiation and function.
• Epigenetic Research: Predicts H3K27ac peaks to support studies of histone modifications and chromatin dynamics.
• Cross-Species Comparisons: Facilitates comparative studies across species using joint learning to assess evolutionary conservation of regulatory elements.

Methodology:

Integrates DNase-seq data and primitive sequences as inputs to a deep learning model trained jointly on datasets from multiple cell types to predict H3K27ac peaks marking active enhancers.