Multi-view local hyperplane nearest neighbor model based on independence criterion for identifying vesicular...

Area of Science:

• Biochemistry
• Computational Biology
• Genomics

Background:

• Vesicular transport proteins are vital for intracellular substance movement and implicated in various human diseases.
• Accurate identification of these proteins is essential for understanding disease mechanisms and developing therapeutic strategies.

Purpose of the Study:

• To develop a novel, accurate computational strategy for identifying vesicular transport proteins.
• To introduce the graph-regularized k-local hyperplane distance nearest neighbor (HSIC-GHKNN) model for this identification task.

Main Methods:

• Utilized pseudo-position-specific scoring matrix (PsePSSM) and AATP for protein evolution information extraction.
• Applied the Edited Nearest Neighbors (ENN) algorithm to handle dataset imbalance.
• Developed a multi-view classifier (HSIC-GHKNN) combining Hilbert-Schmidt independence criterion (HSIC) and local hyperplane distance nearest neighbor.

Main Results:

• Achieved an accuracy of 85.8% and a Matthew correlation coefficient of 0.548 on an independent test set.
• The HSIC-GHKNN model demonstrated superior performance compared to existing methods across most evaluation metrics.
• Analysis identified key influencing factors within the PsePSSM and AATP feature sets.

Conclusions:

• The proposed HSIC-GHKNN strategy offers a robust and accurate method for identifying vesicular transport proteins.
• The multi-view classification model is adaptable for similar protein identification challenges.
• This work contributes to a better understanding of vesicular transport proteins and their roles in health and disease.