Predikin and PredikinDB: a computational framework for the prediction of protein kinase peptide specificity and an associated database of phosphorylation sites

Background: We have previously described an approach to predicting the substrate specificity of serine-threonine protein kinases. The method, named Predikin, identifies key conserved substrate-determining residues in the kinase catalytic domain that contact the substrate in the region of the phosphorylation site and so determine the sequence surrounding the phosphorylation site. Predikin was implemented originally as a web application written in Javascript. Results: Here, we describe a new version of Predikin, completely revised and rewritten as a modular framework that provides multiple enhancements compared with the original. Predikin now consists of two components: (i) PredikinDB, a database of phosphorylation sites that links substrates to kinase sequences and (ii) a Perl module, which provides methods to classify protein kinases, reliably identify substrate-determining residues, generate scoring matrices and score putative phosphorylation sites in query sequences. The performance of Predikin as measured using receiver operator characteristic (ROC) graph analysis equals or surpasses that of existing comparable methods. The Predikin website has been redesigned to incorporate the new features. Conclusion: New features in Predikin include the use of SQL queries to PredikinDB to generate predictions, scoring of predictions, more reliable identification of substrate-determining residues and putative phosphorylation sites, extended options to handle protein kinase and substrate data and an improved web interface. The new features significantly enhance the ability of Predikin to analyse protein kinases and their substrates. Predikin is available at http://predikin.biosci.uq.edu.au.