TY - CHAP
T1 - Identifying protein short linear motifs by position-specific scoring matrix
AU - Fang, Chun
AU - Noguchi, Tamotsu
AU - Yamana, Hayato
AU - Sun, Fuzhen
N1 - Funding Information:
We gratefully thank Dr. Catherine Mooney for providing us with the datasets of SLiMPred, and Professor Xiaohong Liu of Shandong University of Technology for providing a lot of convenience for our study. This work was supported by a grant from Natural Science Foundation of Shandong Province (NO. ZR2014FQ028). We also thank the anonymous reviewers for his/her helpful comments, which improved the manuscript.
Publisher Copyright:
© Springer International Publishing Switzerland 2016.
PY - 2016
Y1 - 2016
N2 - Short linear motifs (SLiMs) play a central role in several biological functions, such as cell regulation, scaffolding, cell signaling, post-translational modification, and cleavage. Identifying SLiMs is an important step for understanding their functions and mechanism. Due to their short length and particular properties, discovery of SLiMs in proteins is a challenge both experimentally and computationally. So far, many existing computational methods adopted many predicted sequence or structures features as input for prediction, there is no report about using position-specific scoring matrix (PSSM) profiles of proteins directly for SLiMs prediction. In this study, we describe a simple method, named as PSSMpred, which only use the evolutionary information generated in form of PSSM profiles of protein sequences for SLiMs prediction. When comparing with other methods tested on the same datasets, PSSMpred achieves the best performances: (1) achieving 0.03–0.1 higher AUC than other methods when tested on HumanTest151; (2) achieving 0.03– 0.05 and 0.03–0.06 higher AUC than other methods when tested on ANCHOR-short and ANCHOR-long respectively.
AB - Short linear motifs (SLiMs) play a central role in several biological functions, such as cell regulation, scaffolding, cell signaling, post-translational modification, and cleavage. Identifying SLiMs is an important step for understanding their functions and mechanism. Due to their short length and particular properties, discovery of SLiMs in proteins is a challenge both experimentally and computationally. So far, many existing computational methods adopted many predicted sequence or structures features as input for prediction, there is no report about using position-specific scoring matrix (PSSM) profiles of proteins directly for SLiMs prediction. In this study, we describe a simple method, named as PSSMpred, which only use the evolutionary information generated in form of PSSM profiles of protein sequences for SLiMs prediction. When comparing with other methods tested on the same datasets, PSSMpred achieves the best performances: (1) achieving 0.03–0.1 higher AUC than other methods when tested on HumanTest151; (2) achieving 0.03– 0.05 and 0.03–0.06 higher AUC than other methods when tested on ANCHOR-short and ANCHOR-long respectively.
KW - Prediction
KW - Protein
KW - Short linear motifs
UR - http://www.scopus.com/inward/record.url?scp=85010204794&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85010204794&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-41009-8_22
DO - 10.1007/978-3-319-41009-8_22
M3 - Chapter
AN - SCOPUS:85010204794
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 206
EP - 214
BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PB - Springer Verlag
ER -