Interface-aware molecular generative framework for protein-protein interaction modulators.

Area of Science:

• Biochemistry and Structural Biology
• Computational Chemistry
• Drug Discovery

Background:

• Protein-protein interactions (PPIs) are vital in biological processes but challenging to target with drugs.
• Existing molecular generative models struggle with the unique properties of PPI interfaces.

Purpose of the Study:

• To develop a novel molecular generative framework, GENiPPI, specifically for targeting PPI interfaces.
• To address the challenge of designing compounds that modulate PPIs using structure-based approaches.

Main Methods:

• Constructed a dataset of PPI interfaces with active/inactive compound pairs.
• Employed Graph Attention Networks and Convolutional Neural Networks to analyze interface and compound features.
• Utilized a Conditional Wasserstein Generative Adversarial Network for molecular generation.

Main Results:

• GENiPPI effectively captures relationships between PPI interfaces and active molecules.
• The framework generates novel, structurally diverse compounds targeting PPIs.
• GENiPPI demonstrates successful few-shot molecular generation, producing compounds similar to known disruptors.

Conclusions:

• GENiPPI is the first structure-based molecular generative model focused on PPI interfaces.
• This framework facilitates the structure-based design of novel PPI modulators.
• GENiPPI advances the field of molecular generation for challenging drug targets.