IIFDTI

The 'IIFDTI' method is an advanced deep learning-based approach designed for the prediction of drug-target interactions (DTIs), a crucial aspect of drug discovery and design. This method leverages independent features of drug-target pairs and their interactive features within substructures to enhance the accuracy of DTI predictions.
Key features of IIFDTI include:
1. Extraction of interactive features between drugs and targets using a bidirectional encoder-decoder architecture.
2. Extraction of independent features of drugs and targets through graph neural networks and convolutional neural networks, respectively.
3. Fusion of all extracted features and their input into fully connected dense layers for DTI predictions.

IIFDTI is designed to consider the independent features of drugs and targets while simulating the interactive features of substructures from a biological perspective. Through various experiments, IIFDTI demonstrates superior performance compared to state-of-the-art methods when evaluated using benchmark datasets. Additionally, mapped visualizations of attention weights showcase its ability to capture biological insights, and case studies highlight its practical applications.