How complete are current yeast and human protein-interaction networks?

Quantitative maps of genetic interactions in yeast - comparative evaluation and integrative analysis

Background

High-throughput genetic screening approaches have enabled systematic means to study how interactions among gene mutations contribute to quantitative fitness phenotypes, with the aim of providing insights into the functional wiring diagrams of genetic interaction networks on a global scale. However, it is poorly known how well these quantitative interaction measurements agree across the screening approaches, which hinders their integrated use toward improving the coverage and quality of the genetic interaction maps in yeast and other organisms.

Results

Using large-scale data matrices from epistatic miniarray profiling (E-MAP), genetic interaction mapping (GIM), and synthetic genetic array (SGA) approaches, we carried out here a systematic comparative evaluation among these quantitative maps of genetic interactions in yeast. The relatively low association between the original interaction measurements or their customized scores could be improved using a matrix-based modelling framework, which enables the use of single- and double-mutant fitness estimates and measurements, respectively, when scoring genetic interactions. Toward an integrative analysis, we show how the detections from the different screening approaches can be combined to suggest novel positive and negative interactions which are complementary to those obtained using any single screening approach alone. The matrix approximation procedure has been made available to support the design and analysis of the future screening studies.

Conclusions

We have shown here that even if the correlation between the currently available quantitative genetic interaction maps in yeast is relatively low, their comparability can be improved by means of our computational matrix approximation procedure, which will enable integrative analysis and detection of a wider spectrum of genetic interactions using data from the complementary screening approaches.

Compartmentalized linkage of genes encoding interacting protein pairs

Emulsion technology has been successfully applied to the fields of next-generation high-throughput sequencing, protein engineering and clinical diagnostics. Here, we extend its scope to proteomics research by developing and characterizing a method, termed iCLIP (in vitro compartmentalized linkage of interacting partners), which enables genes encoding interacting protein pairs to be linked in a single segment of DNA. This will facilitate archiving of the interactomes from library versus library two-hybrid screens as libraries of linked DNAs. We further demonstrate the ability to interrogate a model yeast two-hybrid iCLIP library for interactants by "PCR-pulldown," using a primer specific to a gene of interest along with a universal primer. iCLIP libraries may also be subjected to high-throughput sequencing to generate interactome information. The applicability of the technique is also demonstrated in the related context of the bacterial two-hybrid system.

Large-scale protein interactome mapping: strategies and opportunities

Interactions between proteins are central to any cellular process, and mapping these into a protein network is informative both for the function of individual proteins and the functional organization of the cell as a whole. Many strategies have been developed that are up to this task, and the last 10 years have seen the high-throughput application of a number of those in large-scale, sometimes proteome-wide, interactome mapping efforts. Although initially the quality of the data produced in these screening campaigns has been questioned, quality standards and empirical validation schemes are now in place to ensure high-quality data generation. Through their integration with other 'omics' data, interactomics datasets have proven highly valuable towards applications in different areas of clinical importance.

The evolution of protein interaction networks

The availability of high-throughput methods to detect protein interactions made construction of comprehensive protein interaction networks for several important model organisms possible. Many studies have since focused on uncovering the structural principles of these networks and relating these structures to biological processes. On a global scale, there are striking similarities in the structure of different protein interaction networks, even when distantly related species, such as the yeast Saccharomyces cerevisiae and the fruit fly Drosophila melanogaster, are compared. However, there is also considerable variance in network structures caused by the gain and loss of genes and mutations which alter the interaction behavior of the encoded proteins. Here, we focus on the current state of knowledge on the structure of protein interaction networks and the evolutionary processes that shaped these structures.

Quality control methodology for high-throughput protein-protein interaction screening

Protein-protein interactions are key to many aspects of the cell, including its cytoskeletal structure, the signaling processes in which it is involved, or its metabolism. Failure to form protein complexes or signaling cascades may sometimes translate into pathologic conditions such as cancer or neurodegenerative diseases. The set of all protein interactions between the proteins encoded by an organism constitutes its protein interaction network, representing a scaffold for biological function. Knowing the protein interaction network of an organism, combined with other sources of biological information, can unravel fundamental biological circuits and may help better understand the molecular basics of human diseases. The protein interaction network of an organism can be mapped by combining data obtained from both low-throughput screens, i.e., "one gene at a time" experiments and high-throughput screens, i.e., screens designed to interrogate large sets of proteins at once. In either case, quality controls are required to deal with the inherent imperfect nature of experimental assays. In this chapter, we discuss experimental and statistical methodologies to quantify error rates in high-throughput protein-protein interactions screens.

High-throughput analyses and curation of protein interactions in yeast

The yeast Saccharomyces cerevisiae is the model organism in which protein interactions have been most extensively analyzed. The vast majority of these interactions have been characterized by a variety of sophisticated high-throughput techniques probing different aspects of protein association. This chapter summarizes the major techniques, highlights their complementary nature, discusses the data they produce, and highlights some of the biases from which they suffer. A main focus is the key role played by computational methods for processing, analyzing, and validating the large body of noisy data produced by the experimental procedures. It also describes how computational methods are used to extend the coverage and reliability of protein interaction data by integrating information from heterogeneous sources and reviews the current status of literature-curated data on yeast protein interactions stored in specialized databases.

Understanding protein evolutionary rate by integrating gene co-expression with protein interactions

Background

Among the many factors determining protein evolutionary rate, protein-protein interaction degree (PPID) has been intensively investigated in recent years, but its precise effect on protein evolutionary rate is still heavily debated.

Results

We first confirmed that the correlation between protein evolutionary rate and PPID varies considerably across different protein interaction datasets. Specifically, because of the maximal inconsistency between yeast two-hybrid and other datasets, we reasoned that the difference in experimental methods contributes to our inability to clearly define how PPID affects protein evolutionary rate. To address this, we integrated protein interaction and gene co-expression data to derive a co-expressed protein-protein interaction degree (ePPID) measure, which reflects the number of partners with which a protein can permanently interact. Thus, irrespective of the experimental method employed, we found that (1) ePPID is a better predictor of protein evolutionary rate than PPID, (2) ePPID is a more robust predictor of protein evolutionary rate than PPID, and (3) the contribution of ePPID to protein evolutionary rate is statistically independent of expression level. Analysis of hub proteins in the Structural Interaction Network further supported ePPID as a better predictor of protein evolutionary rate than the number of distinct binding interfaces and clarified the slower evolution of co-expressed multi-interface hub proteins over that of other hub proteins.

Conclusions

Our study firmly established ePPID as a robust predictor of protein evolutionary rate, irrespective of experimental method, and underscored the importance of permanent interactions in shaping the evolutionary outcome.

New insights into protein-protein interaction data lead to increased estimates of the S. cerevisiae interactome size

Background

As protein interactions mediate most cellular mechanisms, protein-protein interaction networks are essential in the study of cellular processes. Consequently, several large-scale interactome mapping projects have been undertaken, and protein-protein interactions are being distilled into databases through literature curation; yet protein-protein interaction data are still far from comprehensive, even in the model organism Saccharomyces cerevisiae. Estimating the interactome size is important for evaluating the completeness of current datasets, in order to measure the remaining efforts that are required.

Results

We examined the yeast interactome from a new perspective, by taking into account how thoroughly proteins have been studied. We discovered that the set of literature-curated protein-protein interactions is qualitatively different when restricted to proteins that have received extensive attention from the scientific community. In particular, these interactions are less often supported by yeast two-hybrid, and more often by more complex experiments such as biochemical activity assays. Our analysis showed that high-throughput and literature-curated interactome datasets are more correlated than commonly assumed, but that this bias can be corrected for by focusing on well-studied proteins. We thus propose a simple and reliable method to estimate the size of an interactome, combining literature-curated data involving well-studied proteins with high-throughput data. It yields an estimate of at least 37, 600 direct physical protein-protein interactions in S. cerevisiae.

Conclusions

Our method leads to higher and more accurate estimates of the interactome size, as it accounts for interactions that are genuine yet difficult to detect with commonly-used experimental assays. This shows that we are even further from completing the yeast interactome map than previously expected.

Combining modularity, conservation, and interactions of proteins significantly increases precision and coverage of protein function prediction

BACKGROUND

While the number of newly sequenced genomes and genes is constantly increasing, elucidation of their function still is a laborious and time-consuming task. This has led to the development of a wide range of methods for predicting protein functions in silico. We report on a new method that predicts function based on a combination of information about protein interactions, orthology, and the conservation of protein networks in different species.

RESULTS

We show that aggregation of these independent sources of evidence leads to a drastic increase in number and quality of predictions when compared to baselines and other methods reported in the literature. For instance, our method generates more than 12,000 novel protein functions for human with an estimated precision of ~76%, among which are 7,500 new functional annotations for 1,973 human proteins that previously had zero or only one function annotated. We also verified our predictions on a set of genes that play an important role in colorectal cancer (MLH1, PMS2, EPHB4 ) and could confirm more than 73% of them based on evidence in the literature.

CONCLUSIONS

The combination of different methods into a single, comprehensive prediction method infers thousands of protein functions for every species included in the analysis at varying, yet always high levels of precision and very good coverage.

DEGAS: de novo discovery of dysregulated pathways in human diseases

Background

Molecular studies of the human disease transcriptome typically involve a search for genes whose expression is significantly dysregulated in sick individuals compared to healthy controls. Recent studies have found that only a small number of the genes in human disease-related pathways show consistent dysregulation in sick individuals. However, those studies found that some pathway genes are affected in most sick individuals, but genes can differ among individuals. While a pathway is usually defined as a set of genes known to share a specific function, pathway boundaries are frequently difficult to assign, and methods that rely on such definition cannot discover novel pathways. Protein interaction networks can potentially be used to overcome these problems.

Methodology/Principal Findings

We present DEGAS (DysrEgulated Gene set Analysis via Subnetworks), a method for identifying connected gene subnetworks significantly enriched for genes that are dysregulated in specimens of a disease. We applied DEGAS to seven human diseases and obtained statistically significant results that appear to home in on compact pathways enriched with hallmarks of the diseases. In Parkinson's disease, we provide novel evidence for involvement of mRNA splicing, cell proliferation, and the 14-3-3 complex in the disease progression. DEGAS is available as part of the MATISSE software package (http://acgt.cs.tau.ac.il/matisse).

Conclusions/Significance

The subnetworks identified by DEGAS can provide a signature of the disease potentially useful for diagnosis, pinpoint possible pathways affected by the disease, and suggest targets for drug intervention.

Integrating diverse biological and computational sources for reliable protein-protein interactions

Background

Protein-protein interactions (PPIs) play important roles in various cellular processes. However, the low quality of current PPI data detected from high-throughput screening techniques has diminished the potential usefulness of the data. We need to develop a method to address the high data noise and incompleteness of PPI data, namely, to filter out inaccurate protein interactions (false positives) and predict putative protein interactions (false negatives).

Results

In this paper, we proposed a novel two-step method to integrate diverse biological and computational sources of supporting evidence for reliable PPIs. The first step, interaction binning or InterBIN, groups PPIs together to more accurately estimate the likelihood (Bin-Confidence score) that the protein pairs interact for each biological or computational evidence source. The second step, interaction classification or InterCLASS, integrates the collected Bin-Confidence scores to build classifiers and identify reliable interactions.

Conclusions

We performed comprehensive experiments on two benchmark yeast PPI datasets. The experimental results showed that our proposed method can effectively eliminate false positives in detected PPIs and identify false negatives by predicting novel yet reliable PPIs. Our proposed method also performed significantly better than merely using each of individual evidence sources, illustrating the importance of integrating various biological and computational sources of data and evidence.

Incomplete and noisy network data as a percolation process

We discuss the ramifications of noisy and incomplete observations of network data on the existence of a giant connected component (GCC). The existence of a GCC in a random graph can be described in terms of a percolation process, and building on general results for classes of random graphs with specified degree distributions we derive percolation thresholds above which GCCs exist. We show that sampling and noise can have a profound effect on the perceived existence of a GCC and find that both processes can destroy it. We also show that the absence of a GCC puts a theoretical upper bound on the false-positive rate and relate our percolation analysis to experimental protein-protein interaction data.

Protein networks reveal detection bias and species consistency when analysed by information-theoretic methods

We apply our recently developed information-theoretic measures for the characterisation and comparison of protein–protein interaction networks. These measures are used to quantify topological network features via macroscopic statistical properties. Network differences are assessed based on these macroscopic properties as opposed to microscopic overlap, homology information or motif occurrences. We present the results of a large–scale analysis of protein–protein interaction networks. Precise null models are used in our analyses, allowing for reliable interpretation of the results. By quantifying the methodological biases of the experimental data, we can define an information threshold above which networks may be deemed to comprise consistent macroscopic topological properties, despite their small microscopic overlaps. Based on this rationale, data from yeast–two–hybrid methods are sufficiently consistent to allow for intra–species comparisons (between different experiments) and inter–species comparisons, while data from affinity–purification mass–spectrometry methods show large differences even within intra–species comparisons.

Manifold learning in protein interactomes

Many studies and applications in the post-genomic era have been devoted to analyze complex biological systems by computational inference methods. We propose to apply manifold learning methods to protein-protein interaction networks (PPIN). Despite their popularity in data-intensive applications, these methods have received limited attention in the context of biological networks. We show that there is both utility and unexplored potential in adopting manifold learning for network inference purposes. In particular, the following advantages are highlighted: (a) fusion with diagnostic statistical tools designed to assign significance to protein interactions based on pre-selected topological features; (b) dissection into components of the interactome in order to elucidate global and local connectivity organization; (c) relevance of embedding the interactome in reduced dimensions for biological validation purposes. We have compared the performances of three well-known techniques--kernel-PCA, RADICAL ICA, and ISOMAP--relatively to their power of mapping the interactome onto new coordinate dimensions where important associations among proteins can be detected, and then back projected such that the corresponding sub-interactomes are reconstructed. This recovery has been done selectively, by using significant information according to a robust statistical procedure, and then standard biological annotation has been provided to validate the results. We expect that a byproduct of using subspace analysis by the proposed techniques is a possible calibration of interactome modularity studies. Supplementary Material is available online at www.libertonlinec.com.

Inferring modules from human protein interactome classes

Background

The integration of protein-protein interaction networks derived from high-throughput screening approaches and complementary sources is a key topic in systems biology. Although integration of protein interaction data is conventionally performed, the effects of this procedure on the result of network analyses has not been examined yet. In particular, in order to optimize the fusion of heterogeneous interaction datasets, it is crucial to consider not only their degree of coverage and accuracy, but also their mutual dependencies and additional salient features.

Results

We examined this issue based on the analysis of modules detected by network clustering methods applied to both integrated and individual (disaggregated) data sources, which we call interactome classes. Due to class diversity, we deal with variable dependencies of data features arising from structural specificities and biases, but also from possible overlaps. Since highly connected regions of the human interactome may point to potential protein complexes, we have focused on the concept of modularity, and elucidated the detection power of module extraction algorithms by independent validations based on GO, MIPS and KEGG. From the combination of protein interactions with gene expressions, a confidence scoring scheme has been proposed before proceeding via GO with further classification in permanent and transient modules.

Conclusions

Disaggregated interactomes are shown to be informative for inferring modularity, thus contributing to perform an effective integrative analysis. Validation of the extracted modules by multiple annotation allows for the assessment of confidence measures assigned to the modules in a protein pathway context. Notably, the proposed multilayer confidence scheme can be used for network calibration by enabling a transition from unweighted to weighted interactomes based on biological evidence.

Databases of protein-protein interactions and complexes

In the current understanding, translation of genomic sequences into proteins is the most important path for realization of genome information. In exercising their intended function, proteins work together through various forms of direct (physical) or indirect interaction mechanisms. For a variety of basic functions, many proteins form a large complex representing a molecular machine or a macromolecular super-structural building block. After several high-throughput techniques for detection of protein-protein interactions had matured, protein interaction data became available in a large scale and curated databases for protein-protein interactions (PPIs) are a new necessity for efficient research. Here, their scope, annotation quality, and retrieval tools are reviewed. In addition, attention is paid to portals that provide unified access to a variety of such databases with added annotation value.

Integrative structure modeling of macromolecular assemblies from proteomics data

Proteomics techniques have been used to generate comprehensive lists of protein interactions in a number of species. However, relatively little is known about how these interactions result in functional multiprotein complexes. This gap can be bridged by combining data from proteomics experiments with data from established structure determination techniques. Correspondingly, integrative computational methods are being developed to provide descriptions of protein complexes at varying levels of accuracy and resolution, ranging from complex compositions to detailed atomic structures.

Functional enrichment analyses and construction of functional similarity networks with high confidence function prediction by PFP

BACKGROUND

A new paradigm of biological investigation takes advantage of technologies that produce large high throughput datasets, including genome sequences, interactions of proteins, and gene expression. The ability of biologists to analyze and interpret such data relies on functional annotation of the included proteins, but even in highly characterized organisms many proteins can lack the functional evidence necessary to infer their biological relevance.

RESULTS

Here we have applied high confidence function predictions from our automated prediction system, PFP, to three genome sequences, Escherichia coli, Saccharomyces cerevisiae, and Plasmodium falciparum (malaria). The number of annotated genes is increased by PFP to over 90% for all of the genomes. Using the large coverage of the function annotation, we introduced the functional similarity networks which represent the functional space of the proteomes. Four different functional similarity networks are constructed for each proteome, one each by considering similarity in a single Gene Ontology (GO) category, i.e. Biological Process, Cellular Component, and Molecular Function, and another one by considering overall similarity with the funSim score. The functional similarity networks are shown to have higher modularity than the protein-protein interaction network. Moreover, the funSim score network is distinct from the single GO-score networks by showing a higher clustering degree exponent value and thus has a higher tendency to be hierarchical. In addition, examining function assignments to the protein-protein interaction network and local regions of genomes has identified numerous cases where subnetworks or local regions have functionally coherent proteins. These results will help interpreting interactions of proteins and gene orders in a genome. Several examples of both analyses are highlighted.

CONCLUSION

The analyses demonstrate that applying high confidence predictions from PFP can have a significant impact on a researchers' ability to interpret the immense biological data that are being generated today. The newly introduced functional similarity networks of the three organisms show different network properties as compared with the protein-protein interaction networks.

Reconstruction and validation of RefRec: a global model for the yeast molecular interaction network

Molecular interaction networks establish all cell biological processes. The networks are under intensive research that is facilitated by new high-throughput measurement techniques for the detection, quantification, and characterization of molecules and their physical interactions. For the common model organism yeast Saccharomyces cerevisiae, public databases store a significant part of the accumulated information and, on the way to better understanding of the cellular processes, there is a need to integrate this information into a consistent reconstruction of the molecular interaction network. This work presents and validates RefRec, the most comprehensive molecular interaction network reconstruction currently available for yeast. The reconstruction integrates protein synthesis pathways, a metabolic network, and a protein-protein interaction network from major biological databases. The core of the reconstruction is based on a reference object approach in which genes, transcripts, and proteins are identified using their primary sequences. This enables their unambiguous identification and non-redundant integration. The obtained total number of different molecular species and their connecting interactions is approximately 67,000. In order to demonstrate the capacity of RefRec for functional predictions, it was used for simulating the gene knockout damage propagation in the molecular interaction network in approximately 590,000 experimentally validated mutant strains. Based on the simulation results, a statistical classifier was subsequently able to correctly predict the viability of most of the strains. The results also showed that the usage of different types of molecular species in the reconstruction is important for accurate phenotype prediction. In general, the findings demonstrate the benefits of global reconstructions of molecular interaction networks. With all the molecular species and their physical interactions explicitly modeled, our reconstruction is able to serve as a valuable resource in additional analyses involving objects from multiple molecular -omes. For that purpose, RefRec is freely available in the Systems Biology Markup Language format.

From Experimental Approaches to Computational Techniques: A Review on the Prediction of Protein-Protein Interactions

A crucial step towards understanding the properties of cellular systems in organisms is to map their network of protein-protein interactions (PPIs) on a proteomic-wide scale completely and as accurately as possible. Uncovering the diverse function of proteins and their interactions within the cell may improve our understanding of disease and provide a basis for the development of novel therapeutic approaches. The development of large-scale high-throughput experiments has resulted in the production of a large volume of data which has aided in the uncovering of PPIs. However, these data are often erroneous and limited in interactome coverage. Therefore, additional experimental and computational methods are required to accelerate the discovery of PPIs. This paper provides a review on the prediction of PPIs addressing key prediction principles and highlighting the common experimental and computational techniques currently employed to infer PPI networks along with relevant studies in the area.

Predicting protein-protein interactions in unbalanced data using the primary structure of proteins

BACKGROUND

Elucidating protein-protein interactions (PPIs) is essential to constructing protein interaction networks and facilitating our understanding of the general principles of biological systems. Previous studies have revealed that interacting protein pairs can be predicted by their primary structure. Most of these approaches have achieved satisfactory performance on datasets comprising equal number of interacting and non-interacting protein pairs. However, this ratio is highly unbalanced in nature, and these techniques have not been comprehensively evaluated with respect to the effect of the large number of non-interacting pairs in realistic datasets. Moreover, since highly unbalanced distributions usually lead to large datasets, more efficient predictors are desired when handling such challenging tasks.

RESULTS

This study presents a method for PPI prediction based only on sequence information, which contributes in three aspects. First, we propose a probability-based mechanism for transforming protein sequences into feature vectors. Second, the proposed predictor is designed with an efficient classification algorithm, where the efficiency is essential for handling highly unbalanced datasets. Third, the proposed PPI predictor is assessed with several unbalanced datasets with different positive-to-negative ratios (from 1:1 to 1:15). This analysis provides solid evidence that the degree of dataset imbalance is important to PPI predictors.

CONCLUSIONS

Dealing with data imbalance is a key issue in PPI prediction since there are far fewer interacting protein pairs than non-interacting ones. This article provides a comprehensive study on this issue and develops a practical tool that achieves both good prediction performance and efficiency using only protein sequence information.

How far can we go with structural mass spectrometry of protein complexes?

Physical interactions between proteins and the formation of stable complexes form the basis of most biological functions. Therefore, a critical step toward understanding the integrated workings of the cell is to determine the structure of protein complexes, and reveal how their structural organization dictates function. Studying the three-dimensional organization of protein assemblies, however, represents a major challenge for structural biologists, due to the large size of the complexes, their heterogeneous composition, their flexibility, and their asymmetric structure. In the last decade, mass spectrometry has proven to be a valuable tool for analyzing such noncovalent complexes. Here, I illustrate the breadth of structural information that can be obtained from this approach, and the steps taken to elucidate the stoichiometry, topology, packing, dynamics, and shape of protein complexes. In addition, I illustrate the challenges that lie ahead, and the future directions toward which the field might be heading.

Building macromolecular assemblies by information-driven docking: introducing the HADDOCK multibody docking server

Over the last years, large scale proteomics studies have generated a wealth of information of biomolecular complexes. Adding the structural dimension to the resulting interactomes represents a major challenge that classical structural experimental methods alone will have difficulties to confront. To meet this challenge, complementary modeling techniques such as docking are thus needed. Among the current docking methods, HADDOCK (High Ambiguity-Driven DOCKing) distinguishes itself from others by the use of experimental and/or bioinformatics data to drive the modeling process and has shown a strong performance in the critical assessment of prediction of interactions (CAPRI), a blind experiment for the prediction of interactions. Although most docking programs are limited to binary complexes, HADDOCK can deal with multiple molecules (up to six), a capability that will be required to build large macromolecular assemblies. We present here a novel web interface of HADDOCK that allows the user to dock up to six biomolecules simultaneously. This interface allows the inclusion of a large variety of both experimental and/or bioinformatics data and supports several types of cyclic and dihedral symmetries in the docking of multibody assemblies. The server was tested on a benchmark of six cases, containing five symmetric homo-oligomeric protein complexes and one symmetric protein-DNA complex. Our results reveal that, in the presence of either bioinformatics and/or experimental data, HADDOCK shows an excellent performance: in all cases, HADDOCK was able to generate good to high quality solutions and ranked them at the top, demonstrating its ability to model symmetric multicomponent assemblies. Docking methods can thus play an important role in adding the structural dimension to interactomes. However, although the current docking methodologies were successful for a vast range of cases, considering the variety and complexity of macromolecular assemblies, inclusion of some kind of experimental information (e.g. from mass spectrometry, nuclear magnetic resonance, cryoelectron microscopy, etc.) will remain highly desirable to obtain reliable results.

Computational approaches for detecting protein complexes from protein interaction networks: a survey

Background

Most proteins form macromolecular complexes to perform their biological functions. However, experimentally determined protein complex data, especially of those involving more than two protein partners, are relatively limited in the current state-of-the-art high-throughput experimental techniques. Nevertheless, many techniques (such as yeast-two-hybrid) have enabled systematic screening of pairwise protein-protein interactions en masse. Thus computational approaches for detecting protein complexes from protein interaction data are useful complements to the limited experimental methods. They can be used together with the experimental methods for mapping the interactions of proteins to understand how different proteins are organized into higher-level substructures to perform various cellular functions.

Results

Given the abundance of pairwise protein interaction data from high-throughput genome-wide experimental screenings, a protein interaction network can be constructed from protein interaction data by considering individual proteins as the nodes, and the existence of a physical interaction between a pair of proteins as a link. This binary protein interaction graph can then be used for detecting protein complexes using graph clustering techniques. In this paper, we review and evaluate the state-of-the-art techniques for computational detection of protein complexes, and discuss some promising research directions in this field.

Conclusions

Experimental results with yeast protein interaction data show that the interaction subgraphs discovered by various computational methods matched well with actual protein complexes. In addition, the computational approaches have also improved in performance over the years. Further improvements could be achieved if the quality of the underlying protein interaction data can be considered adequately to minimize the undesirable effects from the irrelevant and noisy sources, and the various biological evidences can be better incorporated into the detection process to maximize the exploitation of the increasing wealth of biological knowledge available.

Gene prioritization in Type 2 Diabetes using domain interactions and network analysis

Background

Identification of disease genes for Type 2 Diabetes (T2D) by traditional methods has yielded limited success. Based on our previous observation that T2D may result from disturbed protein-protein interactions affected through disrupting modular domain interactions, here we have designed an approach to rank the candidates in the T2D linked genomic regions as plausible disease genes.

Results

Our approach integrates Weight value (Wv) method followed by prioritization using clustering coefficients derived from domain interaction network. Wv for each candidate is calculated based on the assumption that disease genes might be functionally related, mainly facilitated by interactions among domains of the interacting proteins. The benchmarking using a test dataset comprising of both known T2D genes and non-T2D genes revealed that Wv method had a sensitivity and specificity of 0.74 and 0.96 respectively with 9 fold enrichment. The candidate genes having a Wv > 0.5 were called High Weight Elements (HWEs). Further, we ranked HWEs by using the network property-the clustering coefficient (C_i). Each HWE with a C_i < 0.015 was prioritized as plausible disease candidates (HWEc) as previous studies indicate that disease genes tend to avoid dense clustering (with an average C_i of 0.015). This method further prioritized the identified disease genes with a sensitivity of 0.32 and a specificity of 0.98 and enriched the candidate list by 6.8 fold. Thus, from the dataset of 4052 positional candidates the method ranked 435 to be most likely disease candidates. The gene ontology sharing for the candidates showed higher representation of metabolic and signaling processes. The approach also captured genes with unknown functions which were characterized by network motif analysis.

Conclusions

Prioritization of positional candidates is essential for cost-effective and an expedited discovery of disease genes. Here, we demonstrate a novel approach for disease candidate prioritization from numerous loci linked to T2D.

NeMo: Network Module identification in Cytoscape

BACKGROUND

As the size of the known human interactome grows, biologists increasingly rely on computational tools to identify patterns that represent protein complexes and pathways. Previous studies have shown that densely connected network components frequently correspond to community structure and functionally related modules. In this work, we present a novel method to identify densely connected and bipartite network modules based on a log odds score for shared neighbours.

RESULTS

To evaluate the performance of our method (NeMo), we compare it to other widely used tools for community detection including kMetis, MCODE, and spectral clustering. We test these methods on a collection of synthetically constructed networks and the set of MIPS human complexes. We apply our method to the CXC chemokine pathway and find a high scoring functional module of 12 disconnected phospholipase isoforms.

CONCLUSION

We present a novel method that combines a unique neighbour-sharing score with hierarchical agglomerative clustering to identify diverse network communities. The approach is unique in that we identify both dense network and dense bipartite network structures in a single approach. Our results suggest that the performance of NeMo is better than or competitive with leading approaches on both real and synthetic datasets. We minimize model complexity and generalization error in the Bayesian spirit by integrating out nuisance parameters. An implementation of our method is freely available for download as a plugin to Cytoscape through our website and through Cytoscape itself.

Predicting the protein-protein interactions using primary structures with predicted protein surface

BACKGROUND

Many biological functions involve various protein-protein interactions (PPIs). Elucidating such interactions is crucial for understanding general principles of cellular systems. Previous studies have shown the potential of predicting PPIs based on only sequence information. Compared to approaches that require other auxiliary information, these sequence-based approaches can be applied to a broader range of applications.

RESULTS

This study presents a novel sequence-based method based on the assumption that protein-protein interactions are more related to amino acids at the surface than those at the core. The present method considers surface information and maintains the advantage of relying on only sequence data by including an accessible surface area (ASA) predictor recently proposed by the authors. This study also reports the experiments conducted to evaluate a) the performance of PPI prediction achieved by including the predicted surface and b) the quality of the predicted surface in comparison with the surface obtained from structures. The experimental results show that surface information helps to predict interacting protein pairs. Furthermore, the prediction performance achieved by using the surface estimated with the ASA predictor is close to that using the surface obtained from protein structures.

CONCLUSION

This work presents a sequence-based method that takes into account surface information for predicting PPIs. The proposed procedure of surface identification improves the prediction performance with an F-measure of 5.1%. The extracted surfaces are also valuable in other biomedical applications that require similar information.

Computational modeling in systems biology

Interactions among cellular constituents play a crucial role in overall cellular function and organization. These interactions can be viewed as being complementary to the usual "parts list" of genes and proteins and, in conjunction with the expression states of these parts, are key to a systems level understanding of the cell. Here, we review computational approaches to the understanding of the functional roles of cellular networks, ranging from "static" models of network topology to dynamical and stochastic simulations.

Functionally guided alignment of protein interaction networks for module detection

MOTIVATION

Functional module detection within protein interaction networks is a challenging problem due to the sparsity of data and presence of errors. Computational techniques for this task range from purely graph theoretical approaches involving single networks to alignment of multiple networks from several species. Current network alignment methods all rely on protein sequence similarity to map proteins across species.

RESULTS

Here we carry out network alignment using a protein functional similarity measure. We show that using functional similarity to map proteins across species improves network alignment in terms of functional coherence and overlap with experimentally verified protein complexes. Moreover, the results from functional similarity-based network alignment display little overlap (<15%) with sequence similarity-based alignment. Our combined approach integrating sequence and function-based network alignment alongside graph clustering properties offers a 200% increase in coverage of experimental datasets and comparable accuracy to current network alignment methods.

AVAILABILITY

Program binaries and source code is freely available at http://www.stats.ox.ac.uk/research/bioinfo/resources.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Phylogeny-guided interaction mapping in seven eukaryotes

BACKGROUND

The assembly of reliable and complete protein-protein interaction (PPI) maps remains one of the significant challenges in systems biology. Computational methods which integrate and prioritize interaction data can greatly aid in approaching this goal.

RESULTS

We developed a Bayesian inference framework which uses phylogenetic relationships to guide the integration of PPI evidence across multiple datasets and species, providing more accurate predictions. We apply our framework to reconcile seven eukaryotic interactomes: H. sapiens, M. musculus, R. norvegicus, D. melanogaster, C. elegans, S. cerevisiae and A. thaliana. Comprehensive GO-based quality assessment indicates a 5% to 44% score increase in predicted interactomes compared to the input data. Further support is provided by gold-standard MIPS, CYC2008 and HPRD datasets. We demonstrate the ability to recover known PPIs in well-characterized yeast and human complexes (26S proteasome, endosome and exosome) and suggest possible new partners interacting with the putative SWI/SNF chromatin remodeling complex in A. thaliana.

CONCLUSION

Our phylogeny-guided approach compares favorably to two standard methods for mapping PPIs across species. Detailed analysis of predictions in selected functional modules uncovers specific PPI profiles among homologous proteins, establishing interaction-based partitioning of protein families. Provided evidence also suggests that interactions within core complex subunits are in general more conserved and easier to transfer accurately to other organisms, than interactions between these subunits.

POLAR MAPPER: a computational tool for integrated visualization of protein interaction networks and mRNA expression data

Polar Mapper is a computational application for exposing the architecture of protein interaction networks. It facilitates the system-level analysis of mRNA expression data in the context of the underlying protein interaction network. Preliminary analysis of a human protein interaction network and comparison of the yeast oxidative stress and heat shock gene expression responses are addressed as case studies.

Predicting protein-protein interactions in the context of protein evolution

Here we review the methods for the prediction of protein interactions and the ideas in protein evolution that relate to them. The evolutionary assumptions implicit in many of the protein interaction prediction methods are elucidated. We draw attention to the caution needed in deploying certain evolutionary assumptions, in particular cross-organism transfer of interactions by sequence homology, and discuss the known issues in deriving interaction predictions from evidence of co-evolution. We also conject that there is evolutionary knowledge yet to be exploited in the prediction of interactions, in particular the heterogeneity of interactions, the increasing availability of interaction data from multiple species, and the models of protein interaction network growth.

Evolutionarily conserved herpesviral protein interaction networks

Herpesviruses constitute a family of large DNA viruses widely spread in vertebrates and causing a variety of different diseases. They possess dsDNA genomes ranging from 120 to 240 kbp encoding between 70 to 170 open reading frames. We previously reported the protein interaction networks of two herpesviruses, varicella-zoster virus (VZV) and Kaposi's sarcoma-associated herpesvirus (KSHV). In this study, we systematically tested three additional herpesvirus species, herpes simplex virus 1 (HSV-1), murine cytomegalovirus and Epstein-Barr virus, for protein interactions in order to be able to perform a comparative analysis of all three herpesvirus subfamilies. We identified 735 interactions by genome-wide yeast-two-hybrid screens (Y2H), and, together with the interactomes of VZV and KSHV, included a total of 1,007 intraviral protein interactions in the analysis. Whereas a large number of interactions have not been reported previously, we were able to identify a core set of highly conserved protein interactions, like the interaction between HSV-1 UL33 with the nuclear egress proteins UL31/UL34. Interactions were conserved between orthologous proteins despite generally low sequence similarity, suggesting that function may be more conserved than sequence. By combining interactomes of different species we were able to systematically address the low coverage of the Y2H system and to extract biologically relevant interactions which were not evident from single species.

Herpesvirus proteins interact with each other in a complex manner throughout the infectious cycle. This is probably best exemplified in the process where a large number of viral proteins come together to form new viral particles which are subsequently released from the infected cell. A more detailed understanding of how viral proteins interact with each other might assist the development of drugs which may inhibit these interactions and consequently block viral replication. Here we present three genome-wide studies of protein-protein interactions in the herpesviruses herpes simplex virus I, murine cytomegalovirus and Epstein-Barr virus. Altogether we identified 735 interactions in the three viruses, most of which have not previously been reported. By combining these studies with our previously published studies for Kaposi's sarcoma-associated herpesvirus and varicella-zoster virus we were able to perform a comparative analysis of interactions in five related viral species. We observed that a high proportion of interactions were conserved between the different species, despite a low degree of sequence conservation. This implies that by comparing interaction data, we were able to increase the coverage of our viral networks and thus obtain a better and more complete picture of interactions between herpesviral proteins.

Pushing structural information into the yeast interactome by high-throughput protein docking experiments

The last several years have seen the consolidation of high-throughput proteomics initiatives to identify and characterize protein interactions and macromolecular complexes in model organisms. In particular, more that 10,000 high-confidence protein-protein interactions have been described between the roughly 6,000 proteins encoded in the budding yeast genome (Saccharomyces cerevisiae). However, unfortunately, high-resolution three-dimensional structures are only available for less than one hundred of these interacting pairs. Here, we expand this structural information on yeast protein interactions by running the first-ever high-throughput docking experiment with some of the best state-of-the-art methodologies, according to our benchmarks. To increase the coverage of the interaction space, we also explore the possibility of using homology models of varying quality in the docking experiments, instead of experimental structures, and assess how it would affect the global performance of the methods. In total, we have applied the docking procedure to 217 experimental structures and 1,023 homology models, providing putative structural models for over 3,000 protein-protein interactions in the yeast interactome. Finally, we analyze in detail the structural models obtained for the interaction between SAM1-anthranilate synthase complex and the MET30-RNA polymerase III to illustrate how our predictions can be straightforwardly used by the scientific community. The results of our experiment will be integrated into the general 3D-Repertoire pipeline, a European initiative to solve the structures of as many as possible protein complexes in yeast at the best possible resolution. All docking results are available at http://gatealoy.pcb.ub.es/HT_docking/.

Proteins are the main perpetrators of most biological processes. However, they seldom act alone, and most cellular functions are, in fact, carried out by large macromolecular complexes and regulated through intricate protein-protein interaction networks. Consequently, large efforts have been devoted to unveil protein interrelationships in a high-throughput manner, and the last several years have seen the consecution of the first interactome drafts for several model organisms. Unfortunately, these studies only reveal whether two proteins interact, but not the molecular bases of these interactions. A full comprehension of how proteins bind and form complexes can only come from high-resolution, three-dimensional (3D) structures, since they provide the key quasi-atomic details necessary to understand how the individual components in a complex or pathway are assembled and coordinated to function as a molecular unit. Here, we use protein docking experiments, in a high-throughput manner, to predict the 3D structure of over 3,000 interactions in yeast, which will be used to complement the complex structures obtained within the 3D-Repertoire pan-European initiative (http://www.3drepertoire.org).

Assessing the druggability of protein-protein interactions by a supervised machine-learning method

Background

Protein-protein interactions (PPIs) are challenging but attractive targets of small molecule drugs for therapeutic interventions of human diseases. In this era of rapid accumulation of PPI data, there is great need for a methodology that can efficiently select drug target PPIs by holistically assessing the druggability of PPIs. To address this need, we propose here a novel approach based on a supervised machine-learning method, support vector machine (SVM).

Results

To assess the druggability of the PPIs, 69 attributes were selected to cover a wide range of structural, drug and chemical, and functional information on the PPIs. These attributes were used as feature vectors in the SVM-based method. Thirty PPIs known to be druggable were carefully selected from previous studies; these were used as positive instances. Our approach was applied to 1,295 human PPIs with tertiary structures of their protein complexes already solved. The best SVM model constructed discriminated the already-known target PPIs from others at an accuracy of 81% (sensitivity, 82%; specificity, 79%) in cross-validation. Among the attributes, the two with the greatest discriminative power in the best SVM model were the number of interacting proteins and the number of pathways.

Conclusion

Using the model, we predicted several promising candidates for druggable PPIs, such as SMAD4/SKI. As more PPI data are accumulated in the near future, our method will have increased ability to accelerate the discovery of druggable PPIs.

Network-assisted protein identification and data interpretation in shotgun proteomics

Protein assembly and biological interpretation of the assembled protein lists are critical steps in shotgun proteomics data analysis. Although most biological functions arise from interactions among proteins, current protein assembly pipelines treat proteins as independent entities. Usually, only individual proteins with strong experimental evidence, that is, confident proteins, are reported, whereas many possible proteins of biological interest are eliminated. We have developed a clique-enrichment approach (CEA) to rescue eliminated proteins by incorporating the relationship among proteins as embedded in a protein interaction network. In several data sets tested, CEA increased protein identification by 8–23% with an estimated accuracy of 85%. Rescued proteins were supported by existing literature or transcriptome profiling studies at similar levels as confident proteins and at a significantly higher level than abandoned ones. Applying CEA on a breast cancer data set, rescued proteins coded by well-known breast cancer genes. In addition, CEA generated a network view of the proteins and helped show the modular organization of proteins that may underpin the molecular mechanisms of the disease.

Geometric de-noising of protein-protein interaction networks

Understanding complex networks of protein-protein interactions (PPIs) is one of the foremost challenges of the post-genomic era. Due to the recent advances in experimental bio-technology, including yeast-2-hybrid (Y2H), tandem affinity purification (TAP) and other high-throughput methods for protein-protein interaction (PPI) detection, huge amounts of PPI network data are becoming available. Of major concern, however, are the levels of noise and incompleteness. For example, for Y2H screens, it is thought that the false positive rate could be as high as 64%, and the false negative rate may range from 43% to 71%. TAP experiments are believed to have comparable levels of noise.

We present a novel technique to assess the confidence levels of interactions in PPI networks obtained from experimental studies. We use it for predicting new interactions and thus for guiding future biological experiments. This technique is the first to utilize currently the best fitting network model for PPI networks, geometric graphs. Our approach achieves specificity of 85% and sensitivity of 90%. We use it to assign confidence scores to physical protein-protein interactions in the human PPI network downloaded from BioGRID. Using our approach, we predict 251 interactions in the human PPI network, a statistically significant fraction of which correspond to protein pairs sharing common GO terms. Moreover, we validate a statistically significant portion of our predicted interactions in the HPRD database and the newer release of BioGRID. The data and Matlab code implementing the methods are freely available from the web site: http://www.kuchaev.com/Denoising.

Proteins are responsible for much of the biological ‘heavy lifting’ that keeps our cells functioning. However, proteins don't usually work alone; instead they typically bind together to form geometrically and chemically complex structures that are tailored for a specific task. Experimental techniques allow us to detect whether two types of proteins are capable of binding together, or ‘interacting’. This creates a network where two proteins are connected if they have been seen to interact, just as we could regard two people as being connected if they are linked on Facebook. Such protein-protein interaction networks have been developed for several organisms, using a range of methods, all of which are subject to experimental errors. These network data reveal a fascinating and intricate pattern of connections. In particular, it is known that proteins can be arranged into a low-dimensional space, such as a three-dimensional cube, so that interacting proteins are close together. Our work shows that this structure can be exploited to assign confidence levels to recorded protein-protein interactions and predict new interactions that were overlooked experimentally. In tests, we predicted 251 new human protein-protein interactions, and through literature curation we independently validated a statistically significant number of them.

Knowledge based identification of essential signaling from genome-scale siRNA experiments

Background

A systems biology interpretation of genome-scale RNA interference (RNAi) experiments is complicated by scope, experimental variability and network signaling robustness. Over representation approaches (ORA), such as the Hypergeometric or z-score, are an established statistical framework used to associate RNA interference effectors to biologically annotated gene sets or pathways. These methods, however, do not directly take advantage of our growing understanding of the interactome. Furthermore, these methods can miss partial pathway activation and may be biased by protein complexes. Here we present a novel ORA, protein interaction permutation analysis (PIPA), that takes advantage of canonical pathways and established protein interactions to identify pathways enriched for protein interactions connecting RNAi hits.

Results

We use PIPA to analyze genome-scale siRNA cell growth screens performed in HeLa and TOV cell lines. First we show that interacting gene pair siRNA hits are more reproducible than single gene hits. Using protein interactions, PIPA identifies enriched pathways not found using the standard Hypergeometric analysis including the FAK cytoskeletal remodeling pathway. Different branches of the FAK pathway are distinctly essential in HeLa versus TOV cell lines while other portions are uneffected by siRNA perturbations. Enriched hits belong to protein interactions associated with cell cycle regulation, anti-apoptosis, and signal transduction.

Conclusion

PIPA provides an analytical framework to interpret siRNA screen data by merging biologically annotated gene sets with the human interactome. As a result we identify pathways and signaling hypotheses that are statistically enriched to effect cell growth in human cell lines. This method provides a complementary approach to standard gene set enrichment that utilizes the additional knowledge of specific interactions within biological gene sets.

Ontogeny and phylogeny: molecular signatures of selection, constraint, and temporal pleiotropy in the development of Drosophila

Background

Karl Ernst Von Baer noted that species tend to show greater morphological divergence in later stages of development when compared to earlier stages. Darwin originally interpreted these observations via a selectionist framework, suggesting that divergence should be greatest during ontogenic stages in which organisms experienced varying 'conditions of existence' and opportunity for differential selection. Modern hypotheses have focused on the notion that genes and structures involved in early development will be under stronger purifying selection due to the deleterious pleiotropic effects of mutations propagating over the course of ontogeny, also known as the developmental constraint hypothesis.

Results

Using developmental stage-specific expressed sequence tag (EST) libraries, we tested the 2 hypotheses by comparing the rates of evolution of 7,180 genes obtained from 6 species of the Drosophila melanogaster group with respect to ontogeny, and sex and reproduction-related functions in gonadal tissues. Supporting morphological observations, we found evidence of a pattern of increasing mean evolutionary rate in genes that are expressed in subsequent stages of development. Furthermore, supporting expectations that early expressed genes are constrained in divergence, we found that embryo stage genes are involved in a higher mean number of interactions as compared to later stages. We noted that the accelerated divergence of genes in the adult stage is explained by those expressed specifically in the male gonads, whose divergence is driven by positive selection. In addition, accelerated gonadal gene divergence occurs only in the adult stage, suggesting that the effects of selection are observed primarily at the stages during which they are expected occur. Finally, we also found a significant correlation between temporal specificity of gene expression and evolutionary rate, supporting expectations that genes with ubiquitous expression are under stronger constraint.

Conclusion

Taken together, these results support both the developmental constraint hypothesis limiting the divergence of early expressed developmentally important genes, leading to a gradient of divergence rates over ontogeny (embryonic < larval/pupal < adult), as well as Darwin's 'selection opportunity' hypothesis leading to increased divergence in adults, particularly in the case of reproductive tissues. We suggest that a constraint early/opportunity late model best explains divergence over ontogeny.

Interactome and Gene Ontology provide congruent yet subtly different views of a eukaryotic cell

BACKGROUND

The characterization of the global functional structure of a cell is a major goal in bioinformatics and systems biology. Gene Ontology (GO) and the protein-protein interaction network offer alternative views of that structure.

RESULTS

This study presents a comparison of the global structures of the Gene Ontology and the interactome of Saccharomyces cerevisiae. Sensitive, unsupervised methods of clustering applied to a large fraction of the proteome led to establish a GO-interactome correlation value of +0.47 for a general dataset that contains both high and low-confidence interactions and +0.58 for a smaller, high-confidence dataset.

CONCLUSION

The structures of the yeast cell deduced from GO and interactome are substantially congruent. However, some significant differences were also detected, which may contribute to a better understanding of cell function and also to a refinement of the current ontologies.

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 alpha-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.

Predicting protein-protein interactions in Arabidopsis thaliana through integration of orthology, gene ontology and co-expression

Background

Large-scale identification of the interrelationships between different components of the cell, such as the interactions between proteins, has recently gained great interest. However, unraveling large-scale protein-protein interaction maps is laborious and expensive. Moreover, assessing the reliability of the interactions can be cumbersome.

Results

In this study, we have developed a computational method that exploits the existing knowledge on protein-protein interactions in diverse species through orthologous relations on the one hand, and functional association data on the other hand to predict and filter protein-protein interactions in Arabidopsis thaliana. A highly reliable set of protein-protein interactions is predicted through this integrative approach making use of existing protein-protein interaction data from yeast, human, C. elegans and D. melanogaster. Localization, biological process, and co-expression data are used as powerful indicators for protein-protein interactions. The functional repertoire of the identified interactome reveals interactions between proteins functioning in well-conserved as well as plant-specific biological processes. We observe that although common mechanisms (e.g. actin polymerization) and components (e.g. ARPs, actin-related proteins) exist between different lineages, they are active in specific processes such as growth, cancer metastasis and trichome development in yeast, human and Arabidopsis, respectively.

Conclusion

We conclude that the integration of orthology with functional association data is adequate to predict protein-protein interactions. Through this approach, a high number of novel protein-protein interactions with diverse biological roles is discovered. Overall, we have predicted a reliable set of protein-protein interactions suitable for further computational as well as experimental analyses.