Identification of intelligence-related proteins through a robust two-layer predictor.

Area of Science:

• Bioinformatics
• Computational Biology
• Molecular Biology

Background:

• Identifying proteins associated with intelligence is crucial for understanding cognitive functions.
• Existing methods lack a cross-phyla approach for intelligence-related protein identification.

Purpose of the Study:

• To develop a robust computational methodology for identifying intelligence-related proteins across diverse phyla.
• To create a novel two-layer predictor, Intell_Pred, for classifying proteins and transcripts related to intelligence, learning, and memory.

Main Methods:

• Utilized a support vector machine-based classifier to predict intelligence-related proteins based on distinctive features.
• Developed Intell_Pred, a two-layer computational predictor for classifying sequences as intelligence-related or non-intelligence-related.
• Validated the predictor using five-fold cross-validation and an independent blind test.

Main Results:

• Intell_Pred achieved an average accuracy of 87.48% (cross-validation) and 88.89% (independent test).
• A prediction score >0.75 was identified as a reliable threshold for predicting intelligence-related proteins across kingdoms.
• Proteins involved in plant seismonastic movements and associate learning showed high similarity to intelligence-related proteins.

Conclusions:

• Intell_Pred provides a reliable method for identifying and classifying intelligence-related proteins and transcripts.
• The tool demonstrates potential in uncovering the molecular basis of cognitive functions, even in organisms like plants.
• This methodology advances the study of intelligence at a molecular level across the biological spectrum.