Strategies for the effective identification of remotely related sequences in multiple PSSM search approach

Searches using position specific scoring matrices (PSSMs) have been commonly used in remote homology detection procedures such as PSI-BLAST and RPS-BLAST. A PSSM is generated typically using one of the sequences of a family as the reference sequence. In the case of PSI-BLAST searches the reference sequence is same as the query. Recently we have shown that searches against the database of multiple family-profiles, with each one of the members of the family used as a reference sequence, are more effective than searches against the classical database of single family-profiles. Despite relatively a better overall performance when compared with common sequence-profile matching procedures, searches against the multiple family-profiles database result in a few false positives and false negatives. Here we show that profile length and divergence of sequences used in the construction of a PSSM have major influence on the performance of multiple profile based search approach. We also identify that a simple parameter defined by the number of PSSMs corresponding to a family that is hit, for a query, divided by the total number of PSSMs in the family can distinguish effectively the true positives from the false positives in the multiple profiles search approach.     

Predicting existing targets for new drugs base on strategies for missing interactions

Background

There has been paid more and more attention to supervised classification models in the area of predicting drug-target interactions (DTIs). However, in terms of classification, unavoidable missing DTIs in data would cause three issues which have not yet been addressed appropriately by former approaches. Directly labeled as negatives (non-DTIs), missing DTIs increase the confusion of positives (DTIs) and negatives, aggravate the imbalance between few positives and many negatives, and are usually discriminated as highly-scored false positives, which influence the existing measures sharply.

Results

Under the framework of local classification model (LCM), this work focuses on the scenario of predicting how possibly a new drug interacts with known targets. To address the first two issues, two strategies, Spy and Super-target, are introduced accordingly and further integrated to form a two-layer LCM. In the bottom layer, Spy-based local classifiers for protein targets are built by positives, as well as reliable negatives identified among unlabeled drug-target pairs. In the top layer, regular local classifiers specific to super-targets are built with more positives generated by grouping similar targets and their interactions. Furthermore, to handle the third issue, an additional performance measure, Coverage, is presented for assessing DTI prediction. The experiments based on benchmark datasets are finally performed under five-fold cross validation of drugs to evaluate this approach. The main findings are concluded as follows. (1) Both two individual strategies and their combination are effective to missing DTIs, and the combination wins the best. (2) Having the advantages of less confusing decision boundary at the bottom layer and less biased decision boundary at the top layer, our two-layer LCM outperforms two former approaches. (3) Coverage is more robust to missing interactions than other measures and is able to evaluate how far one needs to go down the list of targets to cover all the proper targets of a drug.

Conclusions

Proposing two strategies and one performance measure, this work has addressed the issues derived from missing interactions, which cause confusing and biased decision boundaries in classifiers, as well as the inappropriate measure of predicting performance, in the scenario of predicting interactions between new drugs and known targets.

     

Constructing support vector machine ensembles for cancer classification based on proteomic profiling

In this study, we present a constructive algorithm for training cooperative support vector machine ensembles （CSVMEs）. CSVME combines ensemble architecture design with cooperative training for individual SVMs in ensembles. Unlike most previous studies on training ensembles, CSVME puts emphasis on both accuracy and collaboration among individual SVMs in an ensemble. A group of SVMs selected on the basis of recursive classifier elimination is used in CSVME, and the number of the individual SVMs selected to construct CSVME is determined by 10-fold cross-validation. This kind of SVME has been tested on two ovarian cancer datasets previously obtained by proteomic mass spectrometry. By combining several individual SVMs, the proposed method achieves better performance than the SVME of all base SVMs.     

Prediction of protein-protein interactions using random decision forest framework

MOTIVATION: Protein interactions are of biological interest because they orchestrate a number of cellular processes such as metabolic pathways and immunological recognition. Domains are the building blocks of proteins; therefore, proteins are assumed to interact as a result of their interacting domains. Many domain-based models for protein interaction prediction have been developed, and preliminary results have demonstrated their feasibility. Most of the existing domain-based methods, however, consider only single-domain pairs (one domain from one protein) and assume independence between domain-domain interactions. RESULTS: In this paper, we introduce a domain-based random forest of decision trees to infer protein interactions. Our proposed method is capable of exploring all possible domain interactions and making predictions based on all the protein domains. Experimental results on Saccharomyces cerevisiae dataset demonstrate that our approach can predict protein-protein interactions with higher sensitivity (79.78%) and specificity (64.38%) compared with that of the maximum likelihood approach. Furthermore, our model can be used to infer interactions not only for single-domain pairs but also for multiple domain pairs.     

Bayesian inference for genomic data integration reduces misclassification rate in predicting protein-protein interactions

Protein-protein interactions (PPIs) are essential to most fundamental cellular processes. There has been increasing interest in reconstructing PPIs networks. However, several critical difficulties exist in obtaining reliable predictions. Noticeably, false positive rates can be as high as >80%. Error correction from each generating source can be both time-consuming and inefficient due to the difficulty of covering the errors from multiple levels of data processing procedures within a single test. We propose a novel Bayesian integration method, deemed nonparametric Bayes ensemble learning (NBEL), to lower the misclassification rate (both false positives and negatives) through automatically up-weighting data sources that are most informative, while down-weighting less informative and biased sources. Extensive studies indicate that NBEL is significantly more robust than the classic naïve Bayes to unreliable, error-prone and contaminated data. On a large human data set our NBEL approach predicts many more PPIs than naïve Bayes. This suggests that previous studies may have large numbers of not only false positives but also false negatives. The validation on two human PPIs datasets having high quality supports our observations. Our experiments demonstrate that it is feasible to predict high-throughput PPIs computationally with substantially reduced false positives and false negatives. The ability of predicting large numbers of PPIs both reliably and automatically may inspire people to use computational approaches to correct data errors in general, and may speed up PPIs prediction with high quality. Such a reliable prediction may provide a solid platform to other studies such as protein functions prediction and roles of PPIs in disease susceptibility.     

Improving the yeast two-hybrid system with permutated fusions proteins: the Varicella Zoster Virus interactome

Background  

Yeast two-hybrid (Y2H) screens have been among the most powerful methods to detect and analyze protein-protein interactions. However, they suffer from a significant degree of false negatives, i.e. true interactions that are not detected, and to a certain degree from false positives, i.e. interactions that appear to take place only in the context of the Y2H assay. While the fraction of false positives remains difficult to estimate, the fraction of false negatives in typical Y2H screens is on the order of 70-90%. Here we present novel Y2H vectors that significantly decrease the number of false negatives and help to mitigate the false positive problem.     

Increasing confidence of protein interactomes using network topological metrics

MOTIVATION: Experimental limitations in high-throughput protein-protein interaction detection methods have resulted in low quality interaction datasets that contained sizable fractions of false positives and false negatives. Small-scale, focused experiments are then needed to complement the high-throughput methods to extract true protein interactions. However, the naturally vast interactomes would require much more scalable approaches. RESULTS: We describe a novel method called IRAP* as a computational complement for repurification of the highly erroneous experimentally derived protein interactomes. Our method involves an iterative process of removing interactions that are confidently identified as false positives and adding interactions detected as false negatives into the interactomes. Identification of both false positives and false negatives are performed in IRAP* using interaction confidence measures based on network topological metrics. Potential false positives are identified amongst the detected interactions as those with very low computed confidence values, while potential false negatives are discovered as the undetected interactions with high computed confidence values. Our results from applying IRAP* on large-scale interaction datasets generated by the popular yeast-two-hybrid assays for yeast, fruit fly and worm showed that the computationally repurified interaction datasets contained potentially lower fractions of false positive and false negative errors based on functional homogeneity. AVAILABILITY: The confidence indices for PPIs in yeast, fruit fly and worm as computed by our method can be found at our website http://www.comp.nus.edu.sg/~chenjin/fpfn.     

A Novel Scoring Approach for Protein Co-Purification Data Reveals High Interaction Specificity

Large-scale protein interaction networks (PINs) have typically been discerned using affinity purification followed by mass spectrometry (AP/MS) and yeast two-hybrid (Y2H) techniques. It is generally recognized that Y2H screens detect direct binary interactions while the AP/MS method captures co-complex associations; however, the latter technique is known to yield prevalent false positives arising from a number of effects, including abundance. We describe a novel approach to compute the propensity for two proteins to co-purify in an AP/MS data set, thereby allowing us to assess the detected level of interaction specificity by analyzing the corresponding distribution of interaction scores. We find that two recent AP/MS data sets of yeast contain enrichments of specific, or high-scoring, associations as compared to commensurate random profiles, and that curated, direct physical interactions in two prominent data bases have consistently high scores. Our scored interaction data sets are generally more comprehensive than those of previous studies when compared against four diverse, high-quality reference sets. Furthermore, we find that our scored data sets are more enriched with curated, direct physical associations than Y2H sets. A high-confidence protein interaction network (PIN) derived from the AP/MS data is revealed to be highly modular, and we show that this topology is not the result of misrepresenting indirect associations as direct interactions. In fact, we propose that the modularity in Y2H data sets may be underrepresented, as they contain indirect associations that are significantly enriched with false negatives. The AP/MS PIN is also found to contain significant assortative mixing; however, in line with a previous study we confirm that Y2H interaction data show weak disassortativeness, thus revealing more clearly the distinctive natures of the interaction detection methods. We expect that our scored yeast data sets are ideal for further biological discovery and that our scoring system will prove useful for other AP/MS data sets.     

Two-hybrid arrays

The two-hybrid system is a genetic method for detecting protein-protein interactions. The assay can be applied to random libraries or arrays of colonies that express defined pairs of proteins. Arrays enable the testing of all possible protein pairs for interactions in a systematic fashion. The array format makes a large number of individual assays comparable and thus greatly simplifies the identification of false positives. Two-hybrid arrays have been used to study interactions among the proteins of yeast, hepatitis C virus, vaccinia virus, Drosophila, Caenorhabditis elegans, mouse and other species, and have already identified thousands of interactions.     

Twilight zone of protein sequence alignments

Sequence alignments unambiguously distinguish between protein pairs of similar and non-similar structure when the pairwise sequence identity is high (>40% for long alignments). The signal gets blurred in the twilight zone of 20-35% sequence identity. Here, more than a million sequence alignments were analysed between protein pairs of known structures to re-define a line distinguishing between true and false positives for low levels of similarity. Four results stood out. (i) The transition from the safe zone of sequence alignment into the twilight zone is described by an explosion of false negatives. More than 95% of all pairs detected in the twilight zone had different structures. More precisely, above a cut-off roughly corresponding to 30% sequence identity, 90% of the pairs were homologous; below 25% less than 10% were. (ii) Whether or not sequence homology implied structural identity depended crucially on the alignment length. For example, if 10 residues were similar in an alignment of length 16 (>60%), structural similarity could not be inferred. (iii) The 'more similar than identical' rule (discarding all pairs for which percentage similarity was lower than percentage identity) reduced false positives significantly. (iv) Using intermediate sequences for finding links between more distant families was almost as successful: pairs were predicted to be homologous when the respective sequence families had proteins in common. All findings are applicable to automatic database searches.     

Effect of pressure on helix‐coil transition of an alanine‐based peptide: An FTIR study

Protein–protein interactions are mediated by complementary amino acids defining complementary surfaces. Typically not all members of a family of related proteins interact equally well with all members of a partner family; thus analysis of the sequence record can reveal the complementary amino acid partners that confer interaction specificity. This article develops methods for learning and using probabilistic graphical models of such residue “cross‐coupling” constraints between interacting protein families, based on multiple sequence alignments and information about which pairs of proteins are known to interact. Our models generalize traditional consensus sequence binding motifs, and provide a probabilistic semantics enabling sound evaluation of the plausibility of new possible interactions. Furthermore, predictions made by the models can be explained in terms of the underlying residue interactions. Our approach supports different levels of prior knowledge regarding interactions, including both one‐to‐one (e.g., pairs of proteins from the same organism) and many‐to‐many (e.g., experimentally identified interactions), and we present a technique to account for possible bias in the represented interactions. We apply our approach in studies of PDZ domains and their ligands, fundamental building blocks in a number of protein assemblies. Our algorithms are able to identify biologically interesting cross‐coupling constraints, to successfully identify known interactions, and to make explainable predictions about novel interactions. Proteins 2009. © 2009 Wiley‐Liss, Inc.     

Prediction of yeast protein-protein interaction network: insights from the Gene Ontology and annotations

A map of protein–protein interactions provides valuable insight into the cellular function and machinery of a proteome. By measuring the similarity between two Gene Ontology (GO) terms with a relative specificity semantic relation, here, we proposed a new method of reconstructing a yeast protein–protein interaction map that is solely based on the GO annotations. The method was validated using high-quality interaction datasets for its effectiveness. Based on a Z-score analysis, a positive dataset and a negative dataset for protein–protein interactions were derived. Moreover, a gold standard positive (GSP) dataset with the highest level of confidence that covered 78% of the high-quality interaction dataset and a gold standard negative (GSN) dataset with the lowest level of confidence were derived. In addition, we assessed four high-throughput experimental interaction datasets using the positives and the negatives as well as GSPs and GSNs. Our predicted network reconstructed from GSPs consists of 40753 interactions among 2259 proteins, and forms 16 connected components. We mapped all of the MIPS complexes except for homodimers onto the predicted network. As a result, ~35% of complexes were identified interconnected. For seven complexes, we also identified some nonmember proteins that may be functionally related to the complexes concerned. This analysis is expected to provide a new approach for predicting the protein–protein interaction maps from other completely sequenced genomes with high-quality GO-based annotations.     

Evaluation of residue-residue contact prediction methods: From retrospective to prospective

Sequence-based residue contact prediction plays a crucial role in protein structure reconstruction. In recent years, the combination of evolutionary coupling analysis (ECA) and deep learning (DL) techniques has made tremendous progress for residue contact prediction, thus a comprehensive assessment of current methods based on a large-scale benchmark data set is very needed. In this study, we evaluate 18 contact predictors on 610 non-redundant proteins and 32 CASP13 targets according to a wide range of perspectives. The results show that different methods have different application scenarios: (1) DL methods based on multi-categories of inputs and large training sets are the best choices for low-contact-density proteins such as the intrinsically disordered ones and proteins with shallow multi-sequence alignments (MSAs). (2) With at least 5L (L is sequence length) effective sequences in the MSA, all the methods show the best performance, and methods that rely only on MSA as input can reach comparable achievements as methods that adopt multi-source inputs. (3) For top L/5 and L/2 predictions, DL methods can predict more hydrophobic interactions while ECA methods predict more salt bridges and disulfide bonds. (4) ECA methods can detect more secondary structure interactions, while DL methods can accurately excavate more contact patterns and prune isolated false positives. In general, multi-input DL methods with large training sets dominate current approaches with the best overall performance. Despite the great success of current DL methods must be stated the fact that there is still much room left for further improvement: (1) With shallow MSAs, the performance will be greatly affected. (2) Current methods show lower precisions for inter-domain compared with intra-domain contact predictions, as well as very high imbalances in precisions between intra-domains. (3) Strong prediction similarities between DL methods indicating more feature types and diversified models need to be developed. (4) The runtime of most methods can be further optimized.     

Predicting protein–protein interactions from sequence using correlation coefficient and high-quality interaction dataset

Identifying protein–protein interactions (PPIs) is critical for understanding the cellular function of the proteins and the machinery of a proteome. Data of PPIs derived from high-throughput technologies are often incomplete and noisy. Therefore, it is important to develop computational methods and high-quality interaction dataset for predicting PPIs. A sequence-based method is proposed by combining correlation coefficient (CC) transformation and support vector machine (SVM). CC transformation not only adequately considers the neighboring effect of protein sequence but describes the level of CC between two protein sequences. A gold standard positives (interacting) dataset MIPS Core and a gold standard negatives (non-interacting) dataset GO-NEG of yeast Saccharomyces cerevisiae were mined to objectively evaluate the above method and attenuate the bias. The SVM model combined with CC transformation yielded the best performance with a high accuracy of 87.94% using gold standard positives and gold standard negatives datasets. The source code of MATLAB and the datasets are available on request under smgsmg@mail.ustc.edu.cn.     

Assessing methods for identifying pair-wise atomic contacts across binding interfaces

An essential step in understanding the molecular basis of protein-protein interactions is the accurate identification of inter-protein contacts. We evaluate a number of common methods used in analyzing protein-protein interfaces: a Voronoi polyhedra-based approach, changes in solvent accessible surface area (DeltaSASA) and various radial cutoffs (closest atom, Cbeta, and centroid). First, we compared the Voronoi polyhedra-based analysis to the DeltaSASA and show that using Voronoi polyhedra finds knob-in-hole contacts. To assess the accuracy between the Voronoi polyhedra-based approach and the various radial cutoff methods, two sets of data were used: a small set of 75 experimental mutants and a larger one of 592 structures of protein-protein interfaces. In an assessment using the small set, the Voronoi polyhedra-based methods, a solvent accessible surface area method, and the closest atom radial method identified 100% of the direct contacts defined by mutagenesis data, but only the Voronoi polyhedra-based method found no false positives. The other radial methods were not able to find all of the direct contacts even using a cutoff of 9A. With the larger set of structures, we compared the overall number contacts using the Voronoi polyhedra-based method as a standard. All the radial methods using a 6-A cutoff identified more interactions, but these putative contacts included many false positives as well as missed many false negatives. While radial cutoffs are quicker to calculate as well as to implement, this result highlights why radial cutoff methods do not have the proper resolution to detail the non-homogeneous packing within protein interfaces, and suggests an inappropriate bias in pair-wise contact potentials. Of the radial cutoff methods, using the closest atom approach exhibits the best approximation to the more intensive Voronoi calculation. Our version of the Voronoi polyhedra-based method QContacts is available at .