SecretP: a new method for predicting mammalian secreted proteins

Area of Science:

• Bioinformatics
• Proteomics
• Cell Biology

Background:

• Distinguishing between classically secreted proteins (CSPs), non-secreted proteins (NSPs), and non-classically secreted proteins (NCSPs) is difficult due to limited experimental data for NCSPs.
• Existing methods cannot simultaneously classify these three protein types, hindering comprehensive secretome analysis.

Purpose of the Study:

• To develop a computational method for simultaneously classifying mammalian proteins into CSPs, NSPs, and NCSPs.
• To provide a tool that aids in the identification and study of proteins utilizing non-classical secretory pathways.

Main Methods:

• Data mining of mammalian proteins exported via ER-Golgi-independent pathways from literature.
• Development of a Support Vector Machine (SVM)-based ternary classifier, SecretP.
• Utilizing pseudo-amino acid composition (PseAA) and five additional features for prediction.

Main Results:

• SecretP achieved an accuracy of 88.79% in distinguishing the three protein types.
• On an independent test set of 92 human proteins, SecretP correctly predicted 76 as NCSPs.
• Performance on a public dataset was comparable to existing computational methods.

Conclusions:

• SecretP is a valuable supplementary tool for secretome research.
• The developed method facilitates the identification of proteins secreted through non-classical pathways.
• The SecretP web server is available for public use to advance secretome studies.