Enhanced Prediction of Conformational Flexibility and Phosphorylation in Proteins

RF-Phos: A Novel General Phosphorylation Site Prediction Tool Based on Random Forest

Protein phosphorylation is one of the most widespread regulatory mechanisms in eukaryotes. Over the past decade, phosphorylation site prediction has emerged as an important problem in the field of bioinformatics. Here, we report a new method, termed Random Forest-based Phosphosite predictor 2.0 (RF-Phos 2.0), to predict phosphorylation sites given only the primary amino acid sequence of a protein as input. RF-Phos 2.0, which uses random forest with sequence and structural features, is able to identify putative sites of phosphorylation across many protein families. In side-by-side comparisons based on 10-fold cross validation and an independent dataset, RF-Phos 2.0 compares favorably to other popular mammalian phosphosite prediction methods, such as PhosphoSVM, GPS2.1, and Musite.

In silico prediction and characterization of protein post-translational modifications

This review outlines the computational approaches and procedures for predicting post translational modification (PTM)-induced changes in protein conformation and their influence on protein function(s), the latter being assessed as differential affinity in interaction with either low (ligands for receptors or transporters, substrates for enzymes) or high molecular mass molecules (proteins or nucleic acids in supramolecular assemblies). The scope for an in silico approach is discussed against a summary of the in vitro evidence on the structural and functional outcome of protein PTM.

Computational phosphorylation network reconstruction: methods and resources

The succession of protein activation and deactivation mediated by phosphorylation and dephosphorylation events constitutes a key mechanism of molecular information transfer in cellular systems. To deduce the details of those molecular information cascades and networks has been a central goal pursued by both experimental and computational approaches. Many computational network reconstruction methods employing an array of different statistical learning methods have been developed to infer phosphorylation networks based on different types of molecular data sets such as protein sequence, protein structure, or phosphoproteomics data. In this chapter, different computational network inference methods and resources for biological network reconstruction with a particular focus on phosphorylation networks are surveyed.

MutaCYP: Classification of missense mutations in human cytochromes P450

BACKGROUND

Cytochrome P450 monooxygenases (CYPs) represent a large and diverse family of enzymes involved in various biological processes in humans. Individual genome sequencing has revealed multiple mutations in human CYPs, and many missense mutations have been associated with variety of diseases. Since 3D structures are not resolved for most human CYPs, there is a need for a reliable sequence-based prediction that discriminates benign and disease causing mutations.

METHODS

A new prediction method (MutaCYP) has been developed for scoring de novo missense mutations to have a deleterious effect. The method utilizes only five features, all of which are sequence-based: predicted relative solvent accessibility (RSA), variance of predicted RSA among the residues in close sequence proximity, Z-score of Shannon entropy for a given position, difference in similarity scores and weighted difference in size between wild type and new amino acids. The method is based on a single neural network.

RESULTS

MutaCYP achieves MCC = 0.70, Q2 = 88.52%, Recall = 93.40% with Precision = 91.09%, and AUC = 0.909. Comparative evaluation with other existing methods indicates that MutaCYP outperforms SIFT and PolyPhen-2. Predictions by MutaCYP appear to be orthogonal to predictions by the evaluated methods. Potential issues on reliability of annotations of mutations in the existing databases are discussed.

CONCLUSIONS

A new accurate method, MutaCYP, for classification of missense mutations in human CYPs is presented. The prediction model consists of only five sequence-based features, including a real-valued predicted relative solvent accessibility. The method is publicly available at http://research.cchmc.org/MutaSense/.