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.
CITATION STYLE
Fang, C., Noguchi, T., Yamana, H., & Sun, F. (2016). Identifying protein short linear motifs by position-specific scoring matrix. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9713 LNCS, pp. 206–214). Springer Verlag. https://doi.org/10.1007/978-3-319-41009-8_22
Mendeley helps you to discover research relevant for your work.