BackgroundThe identification of protein-protein interaction sites is a computationally challenging task and important for understanding the biology of protein complexes. There is a rich literature in this field. A broad class of approaches assign to each candidate residue a real-valued score that measures how likely it is that the residue belongs to the interface. The prediction is obtained by thresholding this score.Some probabilistic models classify the residues on the basis of the posterior probabilities. In this paper, we introduce pairwise conditional random fields (pCRFs) in which edges are not restricted to the backbone as in the case of linear-chain CRFs utilized by Li et al. (2007). In fact, any 3D-neighborhood relation can be modeled. On grounds of a generalized Viterbi inference algorithm and a piecewise training process for pCRFs, we demonstrate how to utilize pCRFs to enhance a given residue-wise score-based protein-protein interface predictor on the surface of the protein under study. The features of the pCRF are solely based on the interface predictions scores of the predictor the performance of which shall be improved.ResultsWe performed three sets of experiments with synthetic scores assigned to the surface residues of proteins taken from the data set PlaneDimers compiled by Zellner et al. (2011), from the list published by Keskin et al. (2004) and from the very recent data set due to Cukuroglu et al. (2014). That way we demonstrated that our pCRF-based enhancer is effective given the interface residue score distribution and the non-interface residue score are unimodal.Moreover, the pCRF-based enhancer is also successfully applicable, if the distributions are only unimodal over a certain sub-domain. The improvement is then restricted to that domain. Thus we were able to improve the prediction of the PresCont server devised by Zellner et al. (2011) on PlaneDimers.ConclusionsOur results strongly suggest that pCRFs form a methodological framework to improve residue-wise score-based protein-protein interface predictors given the scores are appropriately distributed. A prototypical implementation of our method is accessible at http://ppicrf.informatik.uni-goettingen.de/index.html.

No abstract

Abstract:The knowledge of protein-DNA interactions is essential to fully understand the molecular activities of life. Many research groups have developed various tools which are either structure-or sequence-based approaches to predict the DNA-binding residues in proteins. The structure-based methods usually achieve good results, but require the knowledge of the 3D structure of protein; while sequence-based methods can be applied to high-throughput of proteins, but require good features. In this study, we present a new information theoretic feature derived from Jensen-Shannon Divergence (JSD) between amino acid distribution of a site and the background distribution of non-binding sites. Our new feature indicates the difference of a certain site from a non-binding site, thus it is informative for detecting binding sites in proteins. We conduct the study with a five-fold cross validation of 263 proteins utilizing the Random Forest classifier. We evaluate the functionality of our new features by combining them with other popular existing features such as position-specific scoring matrix (PSSM), orthogonal binary vector (OBV), and secondary structure (SS). We notice that by adding our features, we can significantly boost the performance of Random Forest classifier, with a clear increment of sensitivity and Matthews correlation coefficient (MCC).

Deep Neural Networks (DNNs) have created a breakthrough in medical image analysis in recent years. Because clinical applications of automated medical analysis are required to be reliable, robust and accurate, it is necessary to devise effective DNNs based models for medical applications. In this paper, we propose an ensemble framework of DNNs for the problem of medical image segmentation with a note that combining multiple models can obtain better results compared to each constituent one. We introduce an effective combining strategy for individual segmentation models based on swarm intelligence, which is a family of optimization algorithms inspired by biological processes. The problem of expensive computational time of the optimizer during the objective function evaluation is relieved by using a surrogate-based method. We train a surrogate on the objective function information of some populations and then use it to predict the objective values of each candidate in the subsequent populations. Experiments run on a number of public datasets indicate that our framework achieves competitive results within reasonable computation time.

In this study, we introduce an ensemble selection method for deep ensemble systems called VEGAS. The deep ensemble models include multiple layers of the ensemble of classifiers (EoC). At each layer, we train the EoC and generates training data for the next layer by concatenating the predictions for training observations and the original training data. The predictions of the classifiers in the last layer are combined by a combining method to obtain the final collaborated prediction. We further improve the prediction accuracy of a deep ensemble model by searching for its optimal configuration, i.e., the optimal set of classifiers in each layer. The optimal configuration is obtained using the Variable-Length Genetic Algorithm (VLGA) to maximize the prediction accuracy of the deep ensemble model on the validation set. We developed three operators of VLGA: roulette wheel selection for breeding, a chunk-based crossover based on the number of classifiers to generate new offsprings, and multiple random points-based mutation on each offspring. The experiments on 20 datasets show that VEGAS outperforms selected benchmark algorithms, including two well-known ensemble methods (Random Forest and XgBoost) and three deep learning methods (Multiple Layer Perceptron, gcForest, and MULES).