Improving the performance of β-turn prediction using predicted shape strings and a two-layer support vector machine model

2011

Improving β-turn Prediction with a Two-Layer Model

Sample size: 426 publication Evidence: high

Author Information

Author(s): Tang Zehui, Li Tonghua, Liu Rida, Xiong Wenwei, Sun Jiangming, Zhu Yaojuan, Chen Guanyan

Primary Institution: Department of Chemistry, Tongji University, Shanghai, China

Can a two-layer support vector machine model improve the prediction of β-turns in protein sequences?

The proposed method significantly improves β-turn prediction accuracy compared to existing methods.

The proposed method achieved a Qtotal of 87.2% on the BT426 dataset.
The method outperformed previous best methods by 6.3% in Qtotal.
The introduction of predicted shape strings significantly contributed to the improvements.
The two-layer model better discriminates between β-turns and non-β-turns.

This study created a new way to predict parts of proteins called β-turns, which helps scientists understand how proteins work better.

The study used predicted secondary structures, predicted shape strings, and a two-layer SVM model to enhance β-turn prediction.

The accuracy of shape string prediction still has room for improvement.

p<0.05

Access the complete publication on the publisher's website