Novel Approach for Decoding Olfactory Receptors Interactions With Molecules Based on Multimodal Feature and Deep...

Area of Science:

• Biochemistry
• Computational Biology
• Neuroscience

Background:

• Olfaction involves complex odorant molecule-receptor interactions, challenging current understanding.
• Predicting these interactions is crucial for human olfaction research and drug discovery.
• Existing methods struggle with the nonlinear and diverse nature of molecule-receptor binding.

Purpose of the Study:

• To develop a novel approach for accurately predicting molecule-receptor interactions in olfaction.
• To precisely identify specific olfactory receptors (ORs) for given odorant molecules using advanced AI.
• To enhance the understanding of olfaction mechanisms and support pharmaceutical development.

Main Methods:

• Utilized multimodal deep learning networks to model molecule-receptor binding.
• Developed a new method for predicting specific ORs for odorant molecules.
• Created a space distribution map to analyze receptor structural intricacies and clustering patterns.

Main Results:

• Achieved a prediction accuracy of 95.1% on a newly curated dataset.
• Demonstrated superior performance compared to existing deep learning models and recognition approaches.
• Showcased robustness against discontinuities in molecule structure-to-OR mapping.

Conclusions:

• The proposed deep learning method significantly improves the prediction of molecule-receptor interactions.
• The findings provide a foundation for the digitization of receptor reactions and a deeper understanding of olfaction.
• This work paves the way for advancements in synthetic biology and drug design.