TY - GEN
T1 - A SVM-based system for predicting protein-protein interactions using a novel representation of protein sequences
AU - You, Zhuhong
AU - Ming, Zhong
AU - Niu, Ben
AU - Deng, Suping
AU - Zhu, Zexuan
PY - 2013
Y1 - 2013
N2 - Protein-protein interactions (PPIs) are crucial for almost all cellular processes, including metabolic cycles, DNA transcription and replication, and signaling cascades. However, the experimental methods for identifying PPIs are both time-consuming and expensive. Therefore, it is important to develop computational approaches for predicting PPIs. In this article, a sequence-based method is developed by combining a novel feature representation using binary coding and Support Vector Machine (SVM). The binary-coding-based descriptors account for the interactions between residues a certain distance apart in the protein sequence, thus this method adequately takes the neighboring effect into account and mine interaction information from the continuous and discontinuous amino acids segments at the same time. When performed on the PPI data of Saccharomyces cerevisiae, the proposed method achieved 86.93% prediction accuracy with 86.99% sensitivity at the precision of 86.90%. Extensive experiments are performed to compare our method with the existing sequence-based method. Achieved results show that the proposed approach is very promising for predicting PPI, so it can be a useful supplementary tool for future proteomics studies.
AB - Protein-protein interactions (PPIs) are crucial for almost all cellular processes, including metabolic cycles, DNA transcription and replication, and signaling cascades. However, the experimental methods for identifying PPIs are both time-consuming and expensive. Therefore, it is important to develop computational approaches for predicting PPIs. In this article, a sequence-based method is developed by combining a novel feature representation using binary coding and Support Vector Machine (SVM). The binary-coding-based descriptors account for the interactions between residues a certain distance apart in the protein sequence, thus this method adequately takes the neighboring effect into account and mine interaction information from the continuous and discontinuous amino acids segments at the same time. When performed on the PPI data of Saccharomyces cerevisiae, the proposed method achieved 86.93% prediction accuracy with 86.99% sensitivity at the precision of 86.90%. Extensive experiments are performed to compare our method with the existing sequence-based method. Achieved results show that the proposed approach is very promising for predicting PPI, so it can be a useful supplementary tool for future proteomics studies.
KW - binary coding
KW - local descriptor
KW - protein sequence
KW - protein-protein interaction
KW - support vector machine
UR - http://www.scopus.com/inward/record.url?scp=84882799517&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-39479-9_73
DO - 10.1007/978-3-642-39479-9_73
M3 - 会议稿件
AN - SCOPUS:84882799517
SN - 9783642394782
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 629
EP - 637
BT - Intelligent Computing Theories - 9th International Conference, ICIC 2013, Proceedings
T2 - 9th International Conference on Intelligent Computing, ICIC 2013
Y2 - 28 July 2013 through 31 July 2013
ER -