DeepDTAF: a deep learning method to predict protein-ligand binding affinity.

Area of Science:

• Computational Chemistry
• Drug Discovery
• Bioinformatics

Background:

• Protein-ligand binding affinity is crucial for drug discovery.
• Experimental determination of binding affinity is time-consuming and resource-intensive.
• Existing computational methods often require protein 3D structures, which are not always available.

Purpose of the Study:

• To develop a novel deep learning approach for predicting protein-ligand binding affinity.
• To leverage sequence-level features to overcome limitations of structure-based methods.
• To accelerate the drug discovery and development process.

Main Methods:

• Developed DeepDTAF, a deep learning model integrating local and global contextual features.
• Utilized protein-binding pocket as a local input feature.
• Employed dilated convolution to capture multiscale long-range interactions.

Main Results:

• DeepDTAF demonstrated significant accuracy improvements compared to state-of-the-art methods.
• The model proved to be a reliable tool for protein-ligand binding affinity prediction.
• Analysis confirmed the effectiveness of integrating local and global features.

Conclusions:

• DeepDTAF offers a reliable and accurate method for predicting protein-ligand binding affinity.
• The approach effectively utilizes sequence-level features, reducing reliance on 3D structures.
• This advancement can significantly accelerate drug discovery pipelines.