Improving the performance of an SVM-based method for predicting protein-protein interactions

Predicting the interactions between all the possible pairs of proteins in a given organism (making a protein-protein interaction map) is a crucial subject in bioinformatics. Most of the previous methods based on supervised machine learning use datasets containing approximately the same number of interacting pairs of proteins (positives) and non-interacting pairs of proteins (negatives) for training a classifier and are estimated to yield a large number of false positives. Thinking that the negatives used in previous studies cannot adequately represent all the negatives that need to be taken into account, we have developed a method based on multiple Support Vector Machines (SVMs) that uses more negatives than positives for predicting interactions between pairs of yeast proteins and pairs of human proteins. We show that the performance of a single SVM improved as we increased the number of negatives used for training and that, if more than one CPU is available, an approach using multiple SVMs is useful not only for improving the performance of classifiers but also for reducing the time required for training them. Our approach can also be applied to assessing the reliability of high-throughput interactions.     

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.     

Choosing negative examples for the prediction of protein-protein interactions

The protein-protein interaction networks of even well-studied model organisms are sketchy at best, highlighting the continued need for computational methods to help direct experimentalists in the search for novel interactions. This need has prompted the development of a number of methods for predicting protein-protein interactions based on various sources of data and methodologies. The common method for choosing negative examples for training a predictor of protein-protein interactions is based on annotations of cellular localization, and the observation that pairs of proteins that have different localization patterns are unlikely to interact. While this method leads to high quality sets of non-interacting proteins, we find that this choice can lead to biased estimates of prediction accuracy, because the constraints placed on the distribution of the negative examples makes the task easier. The effects of this bias are demonstrated in the context of both sequence-based and non-sequence based features used for predicting protein-protein interactions.     

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.     

GRIP: A web-based system for constructing Gold Standard datasets for protein-protein interaction prediction

Background  

Information about protein interaction networks is fundamental to understanding protein function and cellular processes. Interaction patterns among proteins can suggest new drug targets and aid in the design of new therapeutic interventions. Efforts have been made to map interactions on a proteomic-wide scale using both experimental and computational techniques. Reference datasets that contain known interacting proteins (positive cases) and non-interacting proteins (negative cases) are essential to support computational prediction and validation of protein-protein interactions. Information on known interacting and non interacting proteins are usually stored within databases. Extraction of these data can be both complex and time consuming. Although, the automatic construction of reference datasets for classification is a useful resource for researchers no public resource currently exists to perform this task.     

Computational and informatics strategies for identification of specific protein interaction partners in affinity purification mass spectrometry experiments

Analysis of protein interaction networks and protein complexes using affinity purification and mass spectrometry (AP/MS) is among most commonly used and successful applications of proteomics technologies. One of the foremost challenges of AP/MS data is a large number of false-positive protein interactions present in unfiltered data sets. Here we review computational and informatics strategies for detecting specific protein interaction partners in AP/MS experiments, with a focus on incomplete (as opposite to genome wide) interactome mapping studies. These strategies range from standard statistical approaches, to empirical scoring schemes optimized for a particular type of data, to advanced computational frameworks. The common denominator among these methods is the use of label-free quantitative information such as spectral counts or integrated peptide intensities that can be extracted from AP/MS data. We also discuss related issues such as combining multiple biological or technical replicates, and dealing with data generated using different tagging strategies. Computational approaches for benchmarking of scoring methods are discussed, and the need for generation of reference AP/MS data sets is highlighted. Finally, we discuss the possibility of more extended modeling of experimental AP/MS data, including integration with external information such as protein interaction predictions based on functional genomics data.     

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.     

Computational Prediction of Protein–Protein Interaction Networks: Algo-rithms and Resources

Protein interactions play an important role in the discovery of protein functions and pathways in biological processes. This is especially true in case of the diseases caused by the loss of specific protein-protein interactions in the organism. The accuracy of experimental results in finding protein-protein interactions, however, is rather dubious and high throughput experimental results have shown both high false positive beside false negative information for protein interaction. Computational methods have attracted tremendous attention among biologists because of the ability to predict protein-protein interactions and validate the obtained experimental results. In this study, we have reviewed several computational methods for protein-protein interaction prediction as well as describing major databases, which store both predicted and detected protein-protein interactions, and the tools used for analyzing protein interaction networks and improving protein-protein interaction reliability.     

A Computationally Guided Protein-Interaction Screen Uncovers Coiled-Coil Interactions Involved in Vesicular Trafficking

Mapping protein-protein interactions at a domain or motif level can provide structural annotation of the interactome. The α-helical coiled coil is among the most common protein-interaction motifs, and proteins predicted to contain coiled coils participate in diverse biological processes. Here, we introduce a combined computational/experimental screening strategy that we used to uncover coiled-coil interactions among proteins involved in vesicular trafficking in Saccharomyces cerevisiae. A number of coiled-coil complexes have already been identified and reported to play important roles in this important biological process. We identify additional examples of coiled coils that can form physical associations. The computational strategy used to prioritize coiled-coil candidates for testing dramatically improved the efficiency of discovery in a large experimental screen. As assessed by comprehensive yeast two-hybrid assays, computational prefiltering retained 90% of positive interacting pairs and eliminated > 60% of negatives from a set of interaction candidates. The coiled-coil-mediated interaction network elucidated using the combined computational/experimental approach comprises 80 coiled-coil associations between 58 protein pairs, among which 21 protein interactions have not been previously reported in interaction databases and 26 interactions were previously known at the protein level but have now been localized to the coiled-coil motif. The coiled-coil-mediated interactions were specific rather than promiscuous, and many interactions could be recapitulated in a green fluorescent protein complementation assay. Our method provides an efficient route to discovering new coiled-coil interactions and uncovers a number of associations that may have functional significance for vesicular trafficking.     

Co-evolutionary analysis of domains in interacting proteins reveals insights into domain-domain interactions mediating protein-protein interactions

Recent advances in functional genomics have helped generate large-scale high-throughput protein interaction data. Such networks, though extremely valuable towards molecular level understanding of cells, do not provide any direct information about the regions (domains) in the proteins that mediate the interaction. Here, we performed co-evolutionary analysis of domains in interacting proteins in order to understand the degree of co-evolution of interacting and non-interacting domains. Using a combination of sequence and structural analysis, we analyzed protein-protein interactions in F1-ATPase, Sec23p/Sec24p, DNA-directed RNA polymerase and nuclear pore complexes, and found that interacting domain pair(s) for a given interaction exhibits higher level of co-evolution than the non-interacting domain pairs. Motivated by this finding, we developed a computational method to test the generality of the observed trend, and to predict large-scale domain-domain interactions. Given a protein-protein interaction, the proposed method predicts the domain pair(s) that is most likely to mediate the protein interaction. We applied this method on the yeast interactome to predict domain-domain interactions, and used known domain-domain interactions found in PDB crystal structures to validate our predictions. Our results show that the prediction accuracy of the proposed method is statistically significant. Comparison of our prediction results with those from two other methods reveals that only a fraction of predictions are shared by all the three methods, indicating that the proposed method can detect known interactions missed by other methods. We believe that the proposed method can be used with other methods to help identify previously unrecognized domain-domain interactions on a genome scale, and could potentially help reduce the search space for identifying interaction sites.     

Construction of reliable protein-protein interaction networks with a new interaction generality measure

MOTIVATION: Recent screening techniques have made large amounts of protein-protein interaction data available, from which biologically important information such as the function of uncharacterized proteins, the existence of novel protein complexes, and novel signal-transduction pathways can be discovered. However, experimental data on protein interactions contain many false positives, making these discoveries difficult. Therefore computational methods of assessing the reliability of each candidate protein-protein interaction are urgently needed. RESULTS: We developed a new 'interaction generality' measure (IG2) to assess the reliability of protein-protein interactions using only the topological properties of their interaction-network structure. Using yeast protein-protein interaction data, we showed that reliable protein-protein interactions had significantly lower IG2 values than less-reliable interactions, suggesting that IG2 values can be used to evaluate and filter interaction data to enable the construction of reliable protein-protein interaction networks.     

Large-scale mapping of human protein interactome using structural complexes

Although the identification of protein interactions by high-throughput (HTP) methods progresses at a fast pace, 'interactome' data sets still suffer from high rates of false positives and low coverage. To map the human protein interactome, we describe a new framework that uses experimental evidence on structural complexes, the atomic details of binding interfaces and evolutionary conservation. The structurally inferred interaction network is highly modular and more functionally coherent compared with experimental interaction networks derived from multiple literature citations. Moreover, structurally inferred and high-confidence HTP networks complement each other well, allowing us to construct a merged network to generate testable hypotheses and provide valuable experimental leads.     

Computational detection of protein complexes in AP-MS experiments

Protein complex identification is an important goal of protein-protein interaction analysis. To date, development of computational methods for detecting protein complexes has been largely motivated by genome-scale interaction data sets from high-throughput assays such as yeast two-hybrid or tandem affinity purification coupled with mass spectrometry (TAP-MS). However, due to the popularity of small to intermediate-scale affinity purification-mass spectrometry (AP-MS) experiments, protein complex detection is increasingly discussed in local network analysis. In such data sets, protein complexes cannot be detected using binary interaction data alone because the data contain interactions with tagged proteins only and, as a result, interactions between all other proteins remain unobserved, limiting the scope of existing algorithms. In this article, we provide a pragmatic review of network graph-based computational algorithms for protein complex analysis in global interactome data, without requiring any computational background. We discuss the practical gap in applying these algorithms to recently surging small to intermediate-scale AP-MS data sets, and review alternative clustering algorithms using quantitative proteomics data and their limitations.     

Development and implementation of an algorithm for detection of protein complexes in large interaction networks

Background  

After complete sequencing of a number of genomes the focus has now turned to proteomics. Advanced proteomics technologies such as two-hybrid assay, mass spectrometry etc. are producing huge data sets of protein-protein interactions which can be portrayed as networks, and one of the burning issues is to find protein complexes in such networks. The enormous size of protein-protein interaction (PPI) networks warrants development of efficient computational methods for extraction of significant complexes.     

An automated framework for NMR resonance assignment through simultaneous slice picking and spin system forming

Despite significant advances in automated nuclear magnetic resonance-based protein structure determination, the high numbers of false positives and false negatives among the peaks selected by fully automated methods remain a problem. These false positives and negatives impair the performance of resonance assignment methods. One of the main reasons for this problem is that the computational research community often considers peak picking and resonance assignment to be two separate problems, whereas spectroscopists use expert knowledge to pick peaks and assign their resonances at the same time. We propose a novel framework that simultaneously conducts slice picking and spin system forming, an essential step in resonance assignment. Our framework then employs a genetic algorithm, directed by both connectivity information and amino acid typing information from the spin systems, to assign the spin systems to residues. The inputs to our framework can be as few as two commonly used spectra, i.e., CBCA(CO)NH and HNCACB. Different from the existing peak picking and resonance assignment methods that treat peaks as the units, our method is based on ‘slices’, which are one-dimensional vectors in three-dimensional spectra that correspond to certain ( \(N, H\) ) values. Experimental results on both benchmark simulated data sets and four real protein data sets demonstrate that our method significantly outperforms the state-of-the-art methods while using a less number of spectra than those methods. Our method is freely available at http://sfb.kaust.edu.sa/Pages/Software.aspx.     

Iteration method for predicting essential proteins based on orthology and protein-protein interaction networks

ABSTRACT: BACKGROUND: Identification of essential proteins plays a significant role in understanding minimal requirements for the cellular survival and development. Many computational methods have been proposed for predicting essential proteins by using the topological features of protein-protein interaction (PPI) networks. However, most of these methods ignored intrinsic biological meaning of proteins. Moreover, PPI data contains many false positives and false negatives. To overcome these limitations, recently many research groups have started to focus on identification of essential proteins by integrating PPI networks with other biological information. However, none of their methods has widely been acknowledged. RESULTS: By considering the facts that essential proteins are more evolutionarily conserved than nonessential proteins and essential proteins frequently bind each other, we propose an iteration method for predicting essential proteins by integrating the orthology with PPI networks, named by ION. Differently from other methods, ION identifies essential proteins depending on not only the connections between proteins but also their orthologous properties and features of their neighbors. ION is implemented to predict essential proteins in S. cerevisiae. Experimental results show that ION can achieve higher identification accuracy than eight other existing centrality methods in terms of area under the curve (AUC). Moreover, ION identifies a large amount of essential proteins which have been ignored by eight other existing centrality methods because of their low-connectivity. Many proteins ranked in top 100 by ION are both essential and belong to the complexes with certain biological functions. Furthermore, no matter how many reference organisms were selected, ION outperforms all eight other existing centrality methods. While using as many as possible reference organisms can improve the performance of ION. Additionally, ION also shows good prediction performance in E.Coli K-12. CONCLUSIONS: The accuracy of predicting essential proteins can be improved by integrating the orthology with PPI networks.