Dataset Information

Predicting binding sites of hydrolase-inhibitor complexes by combining several methods.

ABSTRACT: Protein-protein interactions play a critical role in protein function. Completion of many genomes is being followed rapidly by major efforts to identify interacting protein pairs experimentally in order to decipher the networks of interacting, coordinated-in-action proteins. Identification of protein-protein interaction sites and detection of specific amino acids that contribute to the specificity and the strength of protein interactions is an important problem with broad applications ranging from rational drug design to the analysis of metabolic and signal transduction networks.In order to increase the power of predictive methods for protein-protein interaction sites, we have developed a consensus methodology for combining four different methods. These approaches include: data mining using Support Vector Machines, threading through protein structures, prediction of conserved residues on the protein surface by analysis of phylogenetic trees, and the Conservatism of Conservatism method of Mirny and Shakhnovich. Results obtained on a dataset of hydrolase-inhibitor complexes demonstrate that the combination of all four methods yield improved predictions over the individual methods.We developed a consensus method for predicting protein-protein interface residues by combining sequence and structure-based methods. The success of our consensus approach suggests that similar methodologies can be developed to improve prediction accuracies for other bioinformatic problems.

SUBMITTER: Sen TZ

PROVIDER: S-EPMC544855 | biostudies-literature | 2004 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Predicting binding sites of hydrolase-inhibitor complexes by combining several methods.

Sen Taner Z TZ Kloczkowski Andrzej A Jernigan Robert L RL Yan Changhui C Honavar Vasant V Ho Kai-Ming KM Wang Cai-Zhuang CZ Ihm Yungok Y Cao Haibo H Gu Xun X Dobbs Drena D

BMC bioinformatics 20041217

<h4>Background</h4>Protein-protein interactions play a critical role in protein function. Completion of many genomes is being followed rapidly by major efforts to identify interacting protein pairs experimentally in order to decipher the networks of interacting, coordinated-in-action proteins. Identification of protein-protein interaction sites and detection of specific amino acids that contribute to the specificity and the strength of protein interactions is an important problem with broad appl ...[more]

PMID: 15606919

Similar Datasets

Project description:Water molecules play an important role in many biological processes in terms of stabilizing protein structures, assisting protein folding, and improving binding affinity. It is well known that, due to the impacts of various environmental factors, it is difficult to identify the conserved water molecules (CWMs) from free water molecules (FWMs) directly as CWMs are normally deeply embedded in proteins and form strong hydrogen bonds with surrounding polar groups. To circumvent this difficulty, in this work, the abundance of spatial structure information and physicochemical properties of water molecules in proteins inspires us to adopt machine learning methods for identifying the CWMs. Therefore, in this study, a machine learning framework to identify the CWMs in the binding sites of the proteins was presented. First, by analyzing water molecules' physicochemical properties and spatial structure information, six features (i.e., atom density, hydrophilicity, hydrophobicity, solvent-accessible surface area, temperature B-factors, and mobility) were extracted. Those features were further analyzed and combined to reach a higher CWM identification rate. As a result, an optimal feature combination was determined. Based on this optimal combination, seven different machine learning models (including support vector machine (SVM), K-nearest neighbor (KNN), decision tree (DT), logistic regression (LR), discriminant analysis (DA), naïve Bayes (NB), and ensemble learning (EL)) were evaluated for their abilities in identifying two categories of water molecules, i.e., CWMs and FWMs. It showed that the EL model was the desired prediction model due to its comprehensive advantages. Furthermore, the presented methodology was validated through a case study of crystal 3skh and extensively compared with Dowser++. The prediction performance showed that the optimal feature combination and the desired EL model in our method could achieve satisfactory prediction accuracy in identifying CWMs from FWMs in the proteins' binding sites.

Project description:Understanding and controlling biological function of proteins at the atomic level is of great importance; allosteric mechanisms provide such an interface. Experimental and computational methods have been developed to search for residue mutations that produce changes in function by altering sites of correlated motion. These methods are often observational in that altered motions are achieved by random sampling without revealing the underlying mechanism(s). We present two deterministic methods founded on structure-function relationships that predict dynamic control sites (i.e., locations that experience correlated motions as a result of altered dynamics). The first method ("static") is based on a single structure conformation (e.g., the wild type (WT)) and utilizes a graph description of atomic connectivity. The local atomic interactions are used to compute the propagation of contact paths. This description of structure connectivity reveals flexible locations that are susceptible to altered dynamics. The second method ("dynamic") is a comparative analysis between the normal modes of a WT structure and a mutant structure. A mapping function is defined that quantifies the significance of the motions in one structure projected onto the motions of the other. Each mode is considered up- or down-regulated according to its change in relative significance. This description of altered dynamics is the basis for a motion correlation analysis, from which the dynamic control sites are readily identified. The methods are theoretically derived and applied using the canonical system dihydrofolate reductase (DHFR). Both methods demonstrate a very high predictive value (p<0.005) in identifying known dynamic control sites. The dynamic method also produces a new hypothesis regarding the mechanism by which the DHFR mutant achieves hyperactivity. These tools are suitable for allosteric investigations and may greatly enhance the speed and effectiveness of other computational and experimental methods.

Project description:BackgroundMany integral membrane proteins, like their non-membrane counterparts, form either transient or permanent multi-subunit complexes in order to carry out their biochemical function. Computational methods that provide structural details of these interactions are needed since, despite their importance, relatively few structures of membrane protein complexes are available.ResultsWe present a method for predicting which residues are in protein-protein binding sites within the transmembrane regions of membrane proteins. The method uses a Random Forest classifier trained on residue type distributions and evolutionary conservation for individual surface residues, followed by spatial averaging of the residue scores. The prediction accuracy achieved for membrane proteins is comparable to that for non-membrane proteins. Also, like previous results for non-membrane proteins, the accuracy is significantly higher for residues distant from the binding site boundary. Furthermore, a predictor trained on non-membrane proteins was found to yield poor accuracy on membrane proteins, as expected from the different distribution of surface residue types between the two classes of proteins. Thus, although the same procedure can be used to predict binding sites in membrane and non-membrane proteins, separate predictors trained on each class of proteins are required. Finally, the contribution of each residue property to the overall prediction accuracy is analyzed and prediction examples are discussed.ConclusionGiven a membrane protein structure and a multiple alignment of related sequences, the presented method gives a prioritized list of which surface residues participate in intramembrane protein-protein interactions. The method has potential applications in guiding the experimental verification of membrane protein interactions, structure-based drug discovery, and also in constraining the search space for computational methods, such as protein docking or threading, that predict membrane protein complex structures.

Dataset Information

Predicting binding sites of hydrolase-inhibitor complexes by combining several methods.

Publications

Predicting binding sites of hydrolase-inhibitor complexes by combining several methods.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets