Abstract

Background: Apoptosis proteins have a key role in the development and the homeostasis of the organism, and are very important to understand the mechanism of cell proliferation and death. The function of apoptosis protein is closely related to its subcellular location.

Objective: Prediction of apoptosis protein subcellular localization is a meaningful task.

Methods: In this study, we predict the apoptosis protein subcellular location by using the PSSMbased second-order moving average descriptor, nonnegative matrix factorization based on Kullback-Leibler divergence and over-sampling algorithms. This model is named by SOMAPKLNMF- OS and constructed on the ZD98, ZW225 and CL317 benchmark datasets. Then, the support vector machine is adopted as the classifier, and the bias-free jackknife test method is used to evaluate the accuracy.

Results: Our prediction system achieves the favorable and promising performance of the overall accuracy on the three datasets and also outperforms the other listed models.

Conclusion: The results show that our model offers a high throughput tool for the identification of apoptosis protein subcellular localization.

Keywords: Subcellular localization, position-specific scoring matrix, second-order moving average, nonnegative matrix factorization, over-sampling, algorithm.

Graphical Abstract