DNA6mA-MINT

DNA6mA-MINT identifies N6-methyladenine (6mA) modifications in DNA sequences to map modification sites for studying epigenetic regulation across genomes.

Key Features:

• Neural Network Architecture: Employs a neural network with convolutional layers to extract high-level features from encoded binary sequence representations of DNA.
• Long Short-Term Memory (LSTM) Layer: Integrates an LSTM layer to capture long-range dependencies in sequence data and improve prediction accuracy.
• Performance Superiority: Demonstrates superior performance compared to state-of-the-art techniques in identifying 6mA modifications and is validated on Mus musculus, rice, and combined-species datasets.
• Cross-Validation: Evaluated using 5-fold and 10-fold cross-validation to assess robustness and reproducibility.

Scientific Applications:

• Epigenetic mapping: Predicts 6mA sites to support mapping of N6-methyladenine distribution and study of epigenetic regulation.
• Comparative genomics: Enables cross-species analysis using validated datasets from Mus musculus, rice, and combined-species collections.
• Biological and disease research: Facilitates investigation of the implications of 6mA modifications in biological processes and diseases.

Methodology:

The methodology follows Chou's 5-steps rule, represents DNA as encoded binary sequences, extracts features with convolutional layers combined with LSTM networks, and assesses performance using 5-fold and 10-fold cross-validation.