iDRBP-ECHF

iDRBP-ECHF predicts DNA- and RNA-binding proteins from protein sequences to identify nucleic acid-binding proteins (NABPs) and support analysis of protein–nucleic acid interactions.

Key Features:

• Sequence-Based Prediction Methodology: Employs a sequence-based approach to identify DNA-binding proteins (DBPs) and RNA-binding proteins (RBPs) from protein sequences.
• Real-World Proportional Benchmark Dataset: Constructs a benchmark dataset reflecting real-world proportions of positive (NABPs) to negative samples.
• Balanced Dataset Generation through Down-Sampling: Applies down-sampling to generate three relatively balanced datasets for model training.
• Integration of Deep Learning Algorithms: Integrates deep learning algorithms to learn compact, high-level representations of input sequence features.
• Superior Performance: Demonstrates superior predictive performance on two independent datasets compared with existing sequence-based methods.

Scientific Applications:

• Protein–Nucleic Acid Interaction Discovery: Facilitates exploration of protein–nucleic acid interactions by predicting NABPs.
• Gene Regulation Analysis: Supports analysis of proteins involved in gene regulation, transcription, and translation.
• Therapeutic Target Identification: Aids identification of potential targets for therapeutic intervention by predicting DBPs and RBPs.
• Genetic Disease Research: Assists understanding of genetic diseases through prediction of NABPs implicated in disease mechanisms.
• Synthetic Biology: Supports synthetic biology by identifying nucleic acid–binding proteins relevant to engineered regulatory systems.

Methodology:

Uses a sequence-based prediction approach, constructs a benchmark dataset mirroring real-world NABP proportions, applies down-sampling to create three relatively balanced training datasets, integrates deep learning algorithms for compact feature representation, and evaluates performance on two independent datasets.