Khan Yaser Daanial, Rasool Nouman, Hussain Waqar, Khan Sher Afzal, Chou Kuo-Chen
Department of Computer Science, School of Systems and Technology, University of Management and Technology, Lahore, Pakistan.
Department of Life Sciences, School of Science, University of Management and Technology, Lahore, Pakistan; Dr Panjwani Center for Molecular Medicine and Drug Research, International Center for Chemical and Biological Sciences, University of Karachi, Karachi, 75270, Pakistan.
Anal Biochem. 2018 Jun 1;550:109-116. doi: 10.1016/j.ab.2018.04.021. Epub 2018 Apr 25.
Among all the post-translational modifications (PTMs) of proteins, Phosphorylation is known to be the most important and highly occurring PTM in eukaryotes and prokaryotes. It has an important regulatory mechanism which is required in most of the pathological and physiological processes including neural activity and cell signalling transduction. The process of threonine phosphorylation modifies the threonine by the addition of a phosphoryl group to the polar side chain, and generates phosphothreonine sites. The investigation and prediction of phosphorylation sites is important and various methods have been developed based on high throughput mass-spectrometry but such experimentations are time consuming and laborious therefore, an efficient and accurate novel method is proposed in this study for the prediction of phosphothreonine sites. The proposed method uses context-based data to calculate statistical moments. Position relative statistical moments are combined together to train neural networks. Using 10-fold cross validation, 94.97% accurate result has been obtained whereas for Jackknife testing, 96% accurate results have been obtained. The overall accuracy of the system is 94.4% to sensitivity value 94% and specificity 94.6%. These results suggest that the proposed method may play an essential role to the other existing methods for phosphothreonine sites prediction.
在蛋白质的所有翻译后修饰（PTM）中，磷酸化是真核生物和原核生物中已知最重要且最常发生的PTM。它具有一种重要的调节机制，在包括神经活动和细胞信号转导在内的大多数病理和生理过程中都是必需的。苏氨酸磷酸化过程通过向极性侧链添加一个磷酰基来修饰苏氨酸，并产生磷酸苏氨酸位点。磷酸化位点的研究和预测很重要，基于高通量质谱已经开发了各种方法，但此类实验既耗时又费力。因此，本研究提出了一种高效准确的新方法来预测磷酸苏氨酸位点。所提出的方法使用基于上下文的数据来计算统计矩。将位置相对统计矩组合在一起以训练神经网络。使用10折交叉验证，获得了94.97%的准确结果，而对于留一法测试，获得了96%的准确结果。该系统的总体准确率为94.4%，灵敏度值为94%，特异性为94.6%。这些结果表明，所提出的方法可能对其他现有的磷酸苏氨酸位点预测方法发挥重要作用。
