Predicting protein structure classes from function predictions

MOTIVATION: We introduce a new approach to using the information contained in sequence-to-function prediction data in order to recognize protein template classes, a critical step in predicting protein structure. The data on which our method is based comprise probabilities of functional categories; for given query sequences these probabilities are obtained by a neural net that has previously been trained on a variety of functionally important features. On a training set of sequences we assess the relevance of individual functional categories for identifying a given structural family. Using a combination of the most relevant categories, the likelihood of a query sequence to belong to a specific family can be estimated. RESULTS: The performance of the method is evaluated using cross-validation. For a fixed structural family and for every sequence, a score is calculated that measures the evidence for family membership. Even for structural families of small size, family members receive significantly higher scores. For some examples, we show that the relevant functional features identified by this method are biologically meaningful. The proposed approach can be used to improve existing sequence-to-structure prediction methods. AVAILABILITY: Matlab code is available on request from the authors. The data are available at http://www.mpisb.mpg.de/~sommer/Fun2Struc/     

Enzyme family classification by support vector machines

One approach for facilitating protein function prediction is to classify proteins into functional families. Recent studies on the classification of G-protein coupled receptors and other proteins suggest that a statistical learning method, Support vector machines (SVM), may be potentially useful for protein classification into functional families. In this work, SVM is applied and tested on the classification of enzymes into functional families defined by the Enzyme Nomenclature Committee of IUBMB. SVM classification system for each family is trained from representative enzymes of that family and seed proteins of Pfam curated protein families. The classification accuracy for enzymes from 46 families and for non-enzymes is in the range of 50.0% to 95.7% and 79.0% to 100% respectively. The corresponding Matthews correlation coefficient is in the range of 54.1% to 96.1%. Moreover, 80.3% of the 8,291 correctly classified enzymes are uniquely classified into a specific enzyme family by using a scoring function, indicating that SVM may have certain level of unique prediction capability. Testing results also suggest that SVM in some cases is capable of classification of distantly related enzymes and homologous enzymes of different functions. Effort is being made to use a more comprehensive set of enzymes as training sets and to incorporate multi-class SVM classification systems to further enhance the unique prediction accuracy. Our results suggest the potential of SVM for enzyme family classification and for facilitating protein function prediction. Our software is accessible at http://jing.cz3.nus.edu.sg/cgi-bin/svmprot.cgi.     

Optimal breeding-value prediction using a sparse selection index

Genomic prediction uses DNA sequences and phenotypes to predict genetic values. In homogeneous populations, theory indicates that the accuracy of genomic prediction increases with sample size. However, differences in allele frequencies and linkage disequilibrium patterns can lead to heterogeneity in SNP effects. In this context, calibrating genomic predictions using a large, potentially heterogeneous, training data set may not lead to optimal prediction accuracy. Some studies tried to address this sample size/homogeneity trade-off using training set optimization algorithms; however, this approach assumes that a single training data set is optimum for all individuals in the prediction set. Here, we propose an approach that identifies, for each individual in the prediction set, a subset from the training data (i.e., a set of support points) from which predictions are derived. The methodology that we propose is a sparse selection index (SSI) that integrates selection index methodology with sparsity-inducing techniques commonly used for high-dimensional regression. The sparsity of the resulting index is controlled by a regularization parameter (λ); the G-Best Linear Unbiased Predictor (G-BLUP) (the prediction method most commonly used in plant and animal breeding) appears as a special case which happens when λ = 0. In this study, we present the methodology and demonstrate (using two wheat data sets with phenotypes collected in 10 different environments) that the SSI can achieve significant (anywhere between 5 and 10%) gains in prediction accuracy relative to the G-BLUP.     

PreSPI: a domain combination based prediction system for protein-protein interaction

With the accumulation of protein and its related data on the Internet, many domain-based computational techniques to predict protein interactions have been developed. However, most techniques still have many limitations when used in real fields. They usually suffer from low accuracy in prediction and do not provide any interaction possibility ranking method for multiple protein pairs. In this paper, we propose a probabilistic framework to predict the interaction probability of proteins and develop an interaction possibility ranking method for multiple protein pairs. Using the ranking method, one can discern the protein pairs that are more likely to interact with each other in multiple protein pairs. The validity of the prediction model was evaluated using an interacting set of protein pairs in yeast and an artificially generated non-interacting set of protein pairs. When 80% of the set of interacting protein pairs in the DIP (Database of Interacting Proteins) was used as a learning set of interacting protein pairs, high sensitivity (77%) and specificity (95%) were achieved for the test groups containing common domains with the learning set of proteins within our framework. The stability of the prediction model was also evident when tested over DIP CORE, HMS-PCI and TAP data. In the validation of the ranking method, we reveal that some correlations exist between the interacting probability and the accuracy of the prediction.     

Bioinformatics methods to predict protein structure and function. A practical approach

Protein structure prediction by using bioinformatics can involve sequence similarity searches, multiple sequence alignments, identification and characterization of domains, secondary structure prediction, solvent accessibility prediction, automatic protein fold recognition, constructing three-dimensional models to atomic detail, and model validation. Not all protein structure prediction projects involve the use of all these techniques. A central part of a typical protein structure prediction is the identification of a suitable structural target from which to extrapolate three-dimensional information for a query sequence. The way in which this is done defines three types of projects. The first involves the use of standard and well-understood techniques. If a structural template remains elusive, a second approach using nontrivial methods is required. If a target fold cannot be reliably identified because inconsistent results have been obtained from nontrivial data analyses, the project falls into the third type of project and will be virtually impossible to complete with any degree of reliability. In this article, a set of protocols to predict protein structure from sequence is presented and distinctions among the three types of project are given. These methods, if used appropriately, can provide valuable indicators of protein structure and function.     

Feature selection and transduction for prediction of molecular bioactivity for drug design

MOTIVATION: In drug discovery a key task is to identify characteristics that separate active (binding) compounds from inactive (non-binding) ones. An automated prediction system can help reduce resources necessary to carry out this task. RESULTS: Two methods for prediction of molecular bioactivity for drug design are introduced and shown to perform well in a data set previously studied as part of the KDD (Knowledge Discovery and Data Mining) Cup 2001. The data is characterized by very few positive examples, a very large number of features (describing three-dimensional properties of the molecules) and rather different distributions between training and test data. Two techniques are introduced specifically to tackle these problems: a feature selection method for unbalanced data and a classifier which adapts to the distribution of the the unlabeled test data (a so-called transductive method). We show both techniques improve identification performance and in conjunction provide an improvement over using only one of the techniques. Our results suggest the importance of taking into account the characteristics in this data which may also be relevant in other problems of a similar type.     

Parallel post-source decay for increasing protein identification confidence levels from 2-D gels

Peptide mass fingerprinting (PMF) has over the years become one of the most commonly used tools for high-throughput analysis and identification of proteins. This method is applicable when relatively simple samples have to be analysed and it is commonly used for analysing proteins previously separated by 2-DE. The most common type of instrument used for this approach is the MALDI-TOF that has proved to be particularly suitable for the PMF analysis because of its characteristics of speed, robustness, sensitivity and automation. We have used a MALDI-TOF equipped with a novel parallel PSD capability (MALDI micro MX), to perform the analysis of two sets of different biological samples isolated by 2-DE. By using a method that integrates the data obtained by PMF analysis with the PSD data obtained in the same experiment, we show that the new multiplexed PSD solution increases the protein identification rate compared to the normal PMF approach. We also investigated the use of a charge-directed fragmentation modification reagent to improve the identification rate and confidence levels.     

Prediction of pi-turns in proteins using PSI-BLAST profiles and secondary structure information

Due to the structural and functional importance of tight turns, some methods have been proposed to predict gamma-turns, beta-turns, and alpha-turns in proteins. In the past, studies of pi-turns were made, but not a single prediction approach has been developed so far. It will be useful to develop a method for identifying pi-turns in a protein sequence. In this paper, the support vector machine (SVM) method has been introduced to predict pi-turns from the amino acid sequence. The training and testing of this approach is performed with a newly collected data set of 640 non-homologous protein chains containing 1931 pi-turns. Different sequence encoding schemes have been explored in order to investigate their effects on the prediction performance. With multiple sequence alignment and predicted secondary structure, the final SVM model yields a Matthews correlation coefficient (MCC) of 0.556 by a 7-fold cross-validation. A web server implementing the prediction method is available at the following URL: http://210.42.106.80/piturn/.     

Ligand-Target Prediction by Structural Network Biology Using nAnnoLyze

Target identification is essential for drug design, drug-drug interaction prediction, dosage adjustment and side effect anticipation. Specifically, the knowledge of structural details is essential for understanding the mode of action of a compound on a target protein. Here, we present nAnnoLyze, a method for target identification that relies on the hypothesis that structurally similar binding sites bind similar ligands. nAnnoLyze integrates structural information into a bipartite network of interactions and similarities to predict structurally detailed compound-protein interactions at proteome scale. The method was benchmarked on a dataset of 6,282 pairs of known interacting ligand-target pairs reaching a 0.96 of area under the Receiver Operating Characteristic curve (AUC) when using the drug names as an input feature for the classifier, and a 0.70 of AUC for “anonymous” compounds or compounds not present in the training set. nAnnoLyze resulted in higher accuracies than its predecessor, AnnoLyze. We applied the method to predict interactions for all the compounds in the DrugBank database with each human protein structure and provide examples of target identification for known drugs against human diseases. The accuracy and applicability of our method to any compound indicate that a comparative docking approach such as nAnnoLyze enables large-scale annotation and analysis of compound–protein interactions and thus may benefit drug development.     

Using Chou's amphiphilic pseudo-amino acid composition and support vector machine for prediction of enzyme subfamily classes

With the rapid increment of protein sequence data, it is indispensable to develop automated and reliable predictive methods for protein function annotation. One approach for facilitating protein function prediction is to classify proteins into functional families from primary sequence. Being the most important group of all proteins, the accurate prediction for enzyme family classes and subfamily classes is closely related to their biological functions. In this paper, for the prediction of enzyme subfamily classes, the Chou's amphiphilic pseudo-amino acid composition [Chou, K.C., 2005. Using amphiphilic pseudo amino acid composition to predict enzyme subfamily classes. Bioinformatics 21, 10-19] has been adopted to represent the protein samples for training the 'one-versus-rest' support vector machine. As a demonstration, the jackknife test was performed on the dataset that contains 2640 oxidoreductase sequences classified into 16 subfamily classes [Chou, K.C., Elrod, D.W., 2003. Prediction of enzyme family classes. J. Proteome Res. 2, 183-190]. The overall accuracy thus obtained was 80.87%. The significant enhancement in the accuracy indicates that the current method might play a complementary role to the exiting methods.     

Local combinational variables: an approach used in DNA-binding helix-turn-helix motif prediction with sequence information

Sequence-based approach for motif prediction is of great interest and remains a challenge. In this work, we develop a local combinational variable approach for sequence-based helix-turn-helix (HTH) motif prediction. First we choose a sequence data set for 88 proteins of 22 amino acids in length to launch an optimized traversal for extracting local combinational segments (LCS) from the data set. Then after LCS refinement, local combinational variables (LCV) are generated to construct prediction models for HTH motifs. Prediction ability of LCV sets at different thresholds is calculated to settle a moderate threshold. The large data set we used comprises 13 HTH families, with 17 455 sequences in total. Our approach predicts HTH motifs more precisely using only primary protein sequence information, with 93.29% accuracy, 93.93% sensitivity and 92.66% specificity. Prediction results of newly reported HTH-containing proteins compared with other prediction web service presents a good prediction model derived from the LCV approach. Comparisons with profile-HMM models from the Pfam protein families database show that the LCV approach maintains a good balance while dealing with HTH-containing proteins and non-HTH proteins at the same time. The LCV approach is to some extent a complementary to the profile-HMM models for its better identification of false-positive data. Furthermore, genome-wide predictions detect new HTH proteins in both Homo sapiens and Escherichia coli organisms, which enlarge applications of the LCV approach. Software for mining LCVs from sequence data set can be obtained from anonymous ftp site ftp://cheminfo.tongji.edu.cn/LCV/freely.     

Mining Proteins with Non-Experimental Annotations Based on an Active Sample Selection Strategy for Predicting Protein Subcellular Localization

Subcellular localization of a protein is important to understand proteins’ functions and interactions. There are many techniques based on computational methods to predict protein subcellular locations, but it has been shown that many prediction tasks have a training data shortage problem. This paper introduces a new method to mine proteins with non-experimental annotations, which are labeled by non-experimental evidences of protein databases to overcome the training data shortage problem. A novel active sample selection strategy is designed, taking advantage of active learning technology, to actively find useful samples from the entire data pool of candidate proteins with non-experimental annotations. This approach can adequately estimate the “value” of each sample, automatically select the most valuable samples and add them into the original training set, to help to retrain the classifiers. Numerical experiments with for four popular multi-label classifiers on three benchmark datasets show that the proposed method can effectively select the valuable samples to supplement the original training set and significantly improve the performances of predicting classifiers.     

Fuzzy ARTMAP prediction of biological activities for potential HIV-1 protease inhibitors using a small molecular data set

Obtaining satisfactory results with neural networks depends on the availability of large data samples. The use of small training sets generally reduces performance. Most classical Quantitative Structure-Activity Relationship (QSAR) studies for a specific enzyme system have been performed on small data sets. We focus on the neuro-fuzzy prediction of biological activities of HIV-1 protease inhibitory compounds when inferring from small training sets. We propose two computational intelligence prediction techniques which are suitable for small training sets, at the expense of some computational overhead. Both techniques are based on the FAMR model. The FAMR is a Fuzzy ARTMAP (FAM) incremental learning system used for classification and probability estimation. During the learning phase, each sample pair is assigned a relevance factor proportional to the importance of that pair. The two proposed algorithms in this paper are: 1) The GA-FAMR algorithm, which is new, consists of two stages: a) During the first stage, we use a genetic algorithm (GA) to optimize the relevances assigned to the training data. This improves the generalization capability of the FAMR. b) In the second stage, we use the optimized relevances to train the FAMR. 2) The Ordered FAMR is derived from a known algorithm. Instead of optimizing relevances, it optimizes the order of data presentation using the algorithm of Dagher et al. In our experiments, we compare these two algorithms with an algorithm not based on the FAM, the FS-GA-FNN introduced in [4], [5]. We conclude that when inferring from small training sets, both techniques are efficient, in terms of generalization capability and execution time. The computational overhead introduced is compensated by better accuracy. Finally, the proposed techniques are used to predict the biological activities of newly designed potential HIV-1 protease inhibitors.     

All-atom folding of the three-helix HIV accessory protein with an adaptive parallel tempering method

All-atom protein structure prediction from the amino acid sequence alone remains an important goal of biophysical chemistry. Recent progress in force field development and validation suggests that the PFF01 free-energy force field correctly predicts the native conformation of various helical proteins as the global optimum of its free-energy surface. Reproducible protein structure prediction requires the availability of efficient optimization methods to locate the global minima of such complex potentials. Here we investigate an adapted version of the parallel tempering method as an efficient parallel stochastic optimization method for protein structure prediction. Using this approach we report the reproducible all-atom folding of the three-helix 40 amino acid HIV accessory protein from random conformations to within 2.4 A backbone RMS deviation from the experimental structure with modest computational resources.     

Multi-generation genomic prediction of maize yield using parametric and non-parametric sparse selection indices

Genomic prediction models are often calibrated using multi-generation data. Over time, as data accumulates, training data sets become increasingly heterogeneous. Differences in allele frequency and linkage disequilibrium patterns between the training and prediction genotypes may limit prediction accuracy. This leads to the question of whether all available data or a subset of it should be used to calibrate genomic prediction models. Previous research on training set optimization has focused on identifying a subset of the available data that is optimal for a given prediction set. However, this approach does not contemplate the possibility that different training sets may be optimal for different prediction genotypes. To address this problem, we recently introduced a sparse selection index (SSI) that identifies an optimal training set for each individual in a prediction set. Using additive genomic relationships, the SSI can provide increased accuracy relative to genomic-BLUP (GBLUP). Non-parametric genomic models using Gaussian kernels (KBLUP) have, in some cases, yielded higher prediction accuracies than standard additive models. Therefore, here we studied whether combining SSIs and kernel methods could further improve prediction accuracy when training genomic models using multi-generation data. Using four years of doubled haploid maize data from the International Maize and Wheat Improvement Center (CIMMYT), we found that when predicting grain yield the KBLUP outperformed the GBLUP, and that using SSI with additive relationships (GSSI) lead to 5–17% increases in accuracy, relative to the GBLUP. However, differences in prediction accuracy between the KBLUP and the kernel-based SSI were smaller and not always significant.Subject terms: Quantitative trait, Genetic models     

EMG-force modeling using parallel cascade identification

Measuring force production in muscles is important for many applications such as gait analysis, medical rehabilitation, and human-machine interaction. Substantial research has focused on finding signal processing and modeling techniques which give accurate estimates of muscle force from the surface-recorded electromyogram (EMG). The proposed methods often do not capture both the nonlinearities and dynamic components of the EMG-force relation. In this study, parallel cascade identification (PCI) is used as a dynamic estimation tool to map surface EMG recordings from upper-arm muscles to the induced force at the wrist. PCI mapping involves generating a parallel connection of a series of linear dynamic and nonlinear static blocks. The PCI model parameters were initialized to obtain the best force prediction. A comparison between PCI and a previously published Hill-based orthogonalization scheme, that captures physiological behaviour of the muscles, has shown 44% improvement in force prediction by PCI (averaged over all subjects in relative-mean-square sense). The improved performance is attributed to the structural capability of PCI to capture nonlinear dynamic effects in the generated force.     

Hybridized distance- and contact-based hierarchical structure modeling for folding soluble and membrane proteins

Crystallography and NMR system (CNS) is currently a widely used method for fragment-free ab initio protein folding from inter-residue distance or contact maps. Despite its widespread use in protein structure prediction, CNS is a decade-old macromolecular structure determination system that was originally developed for solving macromolecular geometry from experimental restraints as opposed to predictive modeling driven by interaction map data. As such, the adaptation of the CNS experimental structure determination protocol for ab initio protein folding is intrinsically anomalous that may undermine the folding accuracy of computational protein structure prediction. In this paper, we propose a new CNS-free hierarchical structure modeling method called DConStruct for folding both soluble and membrane proteins driven by distance and contact information. Rigorous experimental validation shows that DConStruct attains much better reconstruction accuracy than CNS when tested with the same input contact map at varying contact thresholds. The hierarchical modeling with iterative self-correction employed in DConStruct scales at a much higher degree of folding accuracy than CNS with the increase in contact thresholds, ultimately approaching near-optimal reconstruction accuracy at higher-thresholded contact maps. The folding accuracy of DConStruct can be further improved by exploiting distance-based hybrid interaction maps at tri-level thresholding, as demonstrated by the better performance of our method in folding free modeling targets from the 12th and 13th rounds of the Critical Assessment of techniques for protein Structure Prediction (CASP) experiments compared to popular CNS- and fragment-based approaches and energy-minimization protocols, some of which even using much finer-grained distance maps than ours. Additional large-scale benchmarking shows that DConStruct can significantly improve the folding accuracy of membrane proteins compared to a CNS-based approach. These results collectively demonstrate the feasibility of greatly improving the accuracy of ab initio protein folding by optimally exploiting the information encoded in inter-residue interaction maps beyond what is possible by CNS.     

Quantitative analysis of prion-protein degradation by constitutive and immuno-20S proteasomes indicates differences correlated with disease susceptibility

The main part of cytosolic protein degradation depends on the ubiquitin-proteasome system. Proteasomes degrade their substrates into small peptide fragments, some of which are translocated into the endoplasmatic reticulum and loaded onto MHC class I molecules, which are then transported to the cell surface for inspection by CTL. A reliable prediction of proteasomal cleavages in a given protein for the identification of CTL epitopes would benefit immensely from additional cleavage data for the training of prediction algorithms. To increase the knowledge about proteasomal specificity and to gain more insight into the relation of proteasomal activity and susceptibility to prion disease, we digested sheep prion protein with human constitutive and immuno-20S proteasomes. All fragments generated in the digest were quantified. Our results underline the different cleavage specificities of constitutive and immunoproteasomes and provide data for the training of prediction programs for proteasomal cleavages. Furthermore, the kinetic analysis of proteasomal digestion of two different alleles of prion protein shows that even small changes in a protein sequence can affect the overall efficiency of proteasomal processing and thus provides more insight into the possible molecular background of allelic variations and the pathogenicity of prion proteins.     

SVM-Prot: Web-based support vector machine software for functional classification of a protein from its primary sequence

Prediction of protein function is of significance in studying biological processes. One approach for function prediction is to classify a protein into functional family. Support vector machine (SVM) is a useful method for such classification, which may involve proteins with diverse sequence distribution. We have developed a web-based software, SVMProt, for SVM classification of a protein into functional family from its primary sequence. SVMProt classification system is trained from representative proteins of a number of functional families and seed proteins of Pfam curated protein families. It currently covers 54 functional families and additional families will be added in the near future. The computed accuracy for protein family classification is found to be in the range of 69.1-99.6%. SVMProt shows a certain degree of capability for the classification of distantly related proteins and homologous proteins of different function and thus may be used as a protein function prediction tool that complements sequence alignment methods. SVMProt can be accessed at http://jing.cz3.nus.edu.sg/cgi-bin/svmprot.cgi.     

基于SVM 的药物靶点预测方法及其应用

目的:基于已知药物靶点和潜在药物靶点蛋白的一级结构相似性,结合SVM技术研究新的有效的药物靶点预测方法。方法:构造训练样本集,提取蛋白质序列的一级结构特征,进行数据预处理,选择最优核函数,优化参数并进行特征选择,训练最优预测模型,检验模型的预测效果。以G蛋白偶联受体家族的蛋白质为预测集,应用建立的最优分类模型对其进行潜在药物靶点挖掘。结果:基于SVM所建立的最优分类模型预测的平均准确率为81.03%。应用最优分类器对构造的G蛋白预测集进行预测,结果发现预测排位在前20的蛋白质中有多个与疾病相关。特别的,其中有两个G蛋白在治疗靶点数据库(TTD)中显示已作为临床试验的药物靶点。结论:基于SVM和蛋白质序列特征的药物靶点预测方法是有效的,应用该方法预测出的潜在药物靶点能够为发现新的药靶提供参考。