Exploring plant protein functions through structure-based clustering.

Area of Science:

• Plant biology
• Bioinformatics
• Computational biology

Background:

• The rapid increase in plant protein sequences exceeds experimental functional characterization capabilities.
• Sequence homology-based protein function prediction is often unreliable for proteins with low sequence similarity.

Purpose of the Study:

• To utilize AI-predicted protein structures for enhanced plant protein function annotation and discovery.
• To establish a structure-based clustering approach for identifying evolutionary relationships and novel protein families.

Main Methods:

• Leveraging AI tools like AlphaFold for accurate protein structure prediction across diverse plant proteomes.
• Implementing structure-based protein clustering to group proteins with similar structural features.
• Applying this clustering to annotate protein functions and identify new protein families.

Main Results:

• Demonstrated the effectiveness of structure-based clustering in overcoming sequence similarity limitations for function prediction.
• Enabled the identification of distant evolutionary connections and previously unrecognized protein families within plant genomes.
• Provided a robust strategy for exploring plant protein functions and bridging the sequence-function data gap.

Conclusions:

• Structure-based protein clustering, powered by AI predictions, is a potent strategy for plant protein function annotation.
• This approach facilitates the discovery of novel protein families and evolutionary insights, aiding in protein design.
• It effectively addresses the challenge of functional characterization for the vast number of unannotated plant protein sequences.