DEEPSEN

DEEPSEN predicts super-enhancers using a convolutional neural network (CNN) to integrate genomic features and identify enhancer clusters that regulate gene expression and are implicated in cell identity and diseases such as cancer and Alzheimer's disease.

Key Features:

• Convolutional Neural Network Architecture: DEEPSEN leverages a CNN to learn hierarchical patterns from genomic data for super-enhancer prediction.
• Comprehensive Feature Integration: The model incorporates 36 distinct features to capture multifaceted aspects of super-enhancer biology.
• Performance Superiority: DEEPSEN outperforms existing methods for genome-wide super-enhancer prediction, improving prediction accuracy.

Scientific Applications:

• Oncogene Identification: By identifying super-enhancers, DEEPSEN aids in pinpointing oncogenes driven by these regulatory elements in cancer research.
• Disease Association Studies: DEEPSEN highlights super-enhancer regions associated with diseases such as Alzheimer's disease, facilitating discovery of disease-associated mutational sites.

Methodology:

DEEPSEN employs a CNN-based framework that integrates multiple genomic features to predict super-enhancers across the genome and identifies important features contributing to super-enhancer activity.