Prediction of protein-protein interactions using random decision forest framework

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

Mapping protein family interactions: intramolecular and intermolecular protein family interaction repertoires in the PDB and yeast

In the postgenomic era, one of the most interesting and important challenges is to understand protein interactions on a large scale. The physical interactions between protein domains are fundamental to the workings of a cell: in multi-domain polypeptide chains, in multi-subunit proteins and in transient complexes between proteins that also exist independently. To study the large-scale patterns and evolution of interactions between protein domains, we view interactions between protein domains in terms of the interactions between structural families of evolutionarily related domains. This allows us to classify 8151 interactions between individual domains in the Protein Data Bank and the yeast Saccharomyces cerevisiae in terms of 664 types of interactions, between protein families. At least 51 interactions do not occur in the Protein Data Bank and can only be derived from the yeast data. The map of interactions between protein families has the form of a scale-free network, meaning that most protein families only interact with one or two other families, while a few families are extremely versatile in their interactions and are connected to many families. We observe that almost half of all known families engage in interactions with domains from their own family. We also see that the repertoires of interactions of domains within and between polypeptide chains overlap mostly for two specific types of protein families: enzymes and same-family interactions. This suggests that different types of protein interaction repertoires exist for structural, functional and regulatory reasons. Copyright 12001 Academic Press.     

Functional aspects of polyisoprenoid protein substituents: roles in protein-protein interaction and trafficking

There are now numerous examples of post-translational modification with geranylgeranyl or farnesyl substituents. Once thought of as solely a mechanism for association of proteins with membranes, other functional aspects of protein prenylation have come to be appreciated. Although, in almost all instances, such proteins are membrane associated, they are often found to also engage in protein-protein interactions. In some instances, such interactions are critical aspects of prenylated protein trafficking. In this review, the role of prenylation in mediating protein-protein interactions will be considered. The hypothesis will be developed that such interactions occur through recognition of the prenyl group and a second domain, on the prenylated protein, by a heterodimeric protein partner.     

PPI_SVM: Prediction of protein-protein interactions using machine learning,domain-domain affinities and frequency tables

Protein-protein interactions (PPI) control most of the biological processes in a living cell. In order to fully understand protein functions, a knowledge of protein-protein interactions is necessary. Prediction of PPI is challenging, especially when the three-dimensional structure of interacting partners is not known. Recently, a novel prediction method was proposed by exploiting physical interactions of constituent domains. We propose here a novel knowledge-based prediction method, namely PPI_SVM, which predicts interactions between two protein sequences by exploiting their domain information. We trained a two-class support vector machine on the benchmarking set of pairs of interacting proteins extracted from the Database of Interacting Proteins (DIP). The method considers all possible combinations of constituent domains between two protein sequences, unlike most of the existing approaches. Moreover, it deals with both single-domain proteins and multi domain proteins; therefore it can be applied to the whole proteome in high-throughput studies. Our machine learning classifier, following a brainstorming approach, achieves accuracy of 86%, with specificity of 95%, and sensitivity of 75%, which are better results than most previous methods that sacrifice recall values in order to boost the overall precision. Our method has on average better sensitivity combined with good selectivity on the benchmarking dataset. The PPI_SVM source code, train/test datasets and supplementary files are available freely in the public domain at: .     

The susceptibility of disulfide bonds to modification in keratin fibers undergoing tensile stress

Cysteine residues perform a dual role in mammalian hairs. The majority help stabilize the overall assembly of keratins and their associated proteins, but a proportion of inter-molecular disulfide bonds are assumed to be associated with hair mechanical flexibility. Hair cortical microstructure is hierarchical, with a complex macro-molecular organization resulting in arrays of intermediate filaments at a scale of micrometres. Intermolecular disulfide bonds occur within filaments and between them and the surrounding matrix. Wool fibers provide a good model for studying various contributions of differently situated disulfide bonds to fiber mechanics. Within this context, it is not known if all intermolecular disulfide bonds contribute equally, and, if not, then do the disproportionally involved cysteine residues occur at common locations on proteins? In this study, fibers from Romney sheep were subjected to stretching or to their breaking point under wet or dry conditions to detect, through labeling, disulfide bonds that were broken more often than randomly. We found that some cysteines were labeled more often than randomly and that these vary with fiber water content (water disrupts protein-protein hydrogen bonds). Many of the identified cysteine residues were located close to the terminal ends of keratins (head or tail domains) and keratin-associated proteins. Some cysteines in the head and tail domains of type II keratin K85 were labeled in all experimental conditions. When inter-protein hydrogen bonds were disrupted under wet conditions, disulfide labeling occurred in the head domains of type II keratins, likely affecting keratin-keratin-associated protein interactions, and tail domains of the type I keratins, likely affecting keratin-keratin interactions. In contrast, in dry fibers (containing more protein-protein hydrogen bonding), disulfide labeling was also observed in the central domains of affected keratins. This central “rod” region is associated with keratin-keratin interactions between anti-parallel heterodimers in the tetramer of the intermediate filament.     

Inferring protein-protein interactions through high-throughput interaction data from diverse organisms

MOTIVATION: Identifying protein-protein interactions is critical for understanding cellular processes. Because protein domains represent binding modules and are responsible for the interactions between proteins, computational approaches have been proposed to predict protein interactions at the domain level. The fact that protein domains are likely evolutionarily conserved allows us to pool information from data across multiple organisms for the inference of domain-domain and protein-protein interaction probabilities. RESULTS: We use a likelihood approach to estimating domain-domain interaction probabilities by integrating large-scale protein interaction data from three organisms, Saccharomyces cerevisiae, Caenorhabditis elegans and Drosophila melanogaster. The estimated domain-domain interaction probabilities are then used to predict protein-protein interactions in S.cerevisiae. Based on a thorough comparison of sensitivity and specificity, Gene Ontology term enrichment and gene expression profiles, we have demonstrated that it may be far more informative to predict protein-protein interactions from diverse organisms than from a single organism. AVAILABILITY: The program for computing the protein-protein interaction probabilities and supplementary material are available at http://bioinformatics.med.yale.edu/interaction.     

Transient protein-protein interactions: structural, functional, and network properties

Transient interactions, which involve protein interactions that are formed and broken easily, are important in many aspects of cellular function. Here we describe structural and functional properties of transient interactions between globular domains and between globular domains, short peptides, and disordered regions. The importance of posttranslational modifications in transient interactions is also considered. We review techniques used in the detection of the different types of transient protein-protein interactions. We also look at the role of transient interactions within protein-protein interaction networks and consider their contribution to different aspects of these networks.     

A lock-and-key model for protein-protein interactions

MOTIVATION: Protein-protein interaction networks are one of the major post-genomic data sources available to molecular biologists. They provide a comprehensive view of the global interaction structure of an organism's proteome, as well as detailed information on specific interactions. Here we suggest a physical model of protein interactions that can be used to extract additional information at an intermediate level: It enables us to identify proteins which share biological interaction motifs, and also to identify potentially missing or spurious interactions. RESULTS: Our new graph model explains observed interactions between proteins by an underlying interaction of complementary binding domains (lock-and-key model). This leads to a novel graph-theoretical algorithm to identify bipartite subgraphs within protein-protein interaction networks where the underlying data are taken from yeast two-hybrid experimental results. By testing on synthetic data, we demonstrate that under certain modelling assumptions, the algorithm will return correct domain information about each protein in the network. Tests on data from various model organisms show that the local and global patterns predicted by the model are indeed found in experimental data. Using functional and protein structure annotations, we show that bipartite subnetworks can be identified that correspond to biologically relevant interaction motifs. Some of these are novel and we discuss an example involving SH3 domains from the Saccharomyces cerevisiae interactome. AVAILABILITY: The algorithm (in Matlab format) is available (see http://www.maths.strath.ac.uk/~aas96106/lock_key.html).     

Reduced polymorphism in domains involved in protein-protein interactions

Genome sequencing of various individuals or isolates of the same species allows studying the polymorphism level of specific proteins and protein domains. Here we ask whether domains that are known to be involved in mediating protein-protein interactions show lower polymorphism than other domains. To this end we take advantage of a recent genome sequence dataset of 39 Saccahromyces cerevisiae strains and the experimentally determined protein interaction network of the laboratory strain. We analyze the polymorphism in domain residues involved in interactions at various levels of resolution, depending on their likelihood to be interaction mediators. We find that domains involved in interactions are less polymorphic than other domains. Furthermore, as the likelihood of a residue to be involved in interaction increases, its polymorphism decreases. Our results suggest that purifying selection operates on domains capable of mediating protein interactions to maintain their function.     

Gene Fusion Analysis in the Battle against the African Endemic Sleeping Sickness

The protozoan Trypanosoma brucei causes African Trypanosomiasis or sleeping sickness in humans, which can be lethal if untreated. Most available pharmacological treatments for the disease have severe side-effects. The purpose of this analysis was to detect novel protein-protein interactions (PPIs), vital for the parasite, which could lead to the development of drugs against this disease to block the specific interactions. In this work, the Domain Fusion Analysis (Rosetta Stone method) was used to identify novel PPIs, by comparing T. brucei to 19 organisms covering all major lineages of the tree of life. Overall, 49 possible protein-protein interactions were detected, and classified based on (a) statistical significance (BLAST e-value, domain length etc.), (b) their involvement in crucial metabolic pathways, and (c) their evolutionary history, particularly focusing on whether a protein pair is split in T. brucei and fused in the human host. We also evaluated fusion events including hypothetical proteins, and suggest a possible molecular function or involvement in a certain biological process. This work has produced valuable results which could be further studied through structural biology or other experimental approaches so as to validate the protein-protein interactions proposed here. The evolutionary analysis of the proteins involved showed that, gene fusion or gene fission events can happen in all organisms, while some protein domains are more prone to fusion and fission events and present complex evolutionary patterns.     

Global topological features of cancer proteins in the human interactome

MOTIVATION: The study of interactomes, or networks of protein-protein interactions, is increasingly providing valuable information on biological systems. Here we report a study of cancer proteins in an extensive human protein-protein interaction network constructed by computational methods. RESULTS: We show that human proteins translated from known cancer genes exhibit a network topology that is different from that of proteins not documented as being mutated in cancer. In particular, cancer proteins show an increase in the number of proteins they interact with. They also appear to participate in central hubs rather than peripheral ones, mirroring their greater centrality and participation in networks that form the backbone of the proteome. Moreover, we show that cancer proteins contain a high ratio of highly promiscuous structural domains, i.e., domains with a high propensity for mediating protein interactions. These observations indicate an underlying evolutionary distinction between the two groups of proteins, reflecting the central roles of proteins, whose mutations lead to cancer. CONTACT: paul.bates@cancer.org.uk SUPPLEMENTARY INFORMATION: The interactome data are available though the PIP (Potential Interactions of Proteins) web server at http://bmm.cancerresearchuk.org/servers/pip. Further additional material is available at http://bmm.cancerresearchuk.org/servers/pip/bioinformatics/     

Proteome scanning to predict PDZ domain interactions using support vector machines

Background  

PDZ domains mediate protein-protein interactions involved in important biological processes through the recognition of short linear motifs in their target proteins. Two recent independent studies have used protein microarray or phage display technology to detect PDZ domain interactions with peptide ligands on a large scale. Several computational predictors of PDZ domain interactions have been developed, however they are trained using only protein microarray data and focus on limited subsets of PDZ domains. An accurate predictor of genomic PDZ domain interactions would allow the proteomes of organisms to be scanned for potential binders. Such an application would require an accurate and precise predictor to avoid generating too many false positive hits given the large amount of possible interactors in a given proteome. Once validated these predictions will help to increase the coverage of current PDZ domain interaction networks and further our understanding of the roles that PDZ domains play in a variety of biological processes.     

High-throughput analysis of peptide-binding modules

Modular protein interaction domains (PIDs) that recognize linear peptide motifs are found in hundreds of proteins within the human genome. Some PIDs such as SH2, 14-3-3, Chromo, and Bromo domains serve to recognize posttranslational modification (PTM) of amino acids (such as phosphorylation, acetylation, methylation, etc.) and translate these into discrete cellular responses. Other modules such as SH3 and PSD-95/Discs-large/ZO-1 (PDZ) domains recognize linear peptide epitopes and serve to organize protein complexes based on localization and regions of elevated concentration. In both cases, the ability to nucleate-specific signaling complexes is in large part dependent on the selectivity of a given protein module for its cognate peptide ligand. High-throughput (HTP) analysis of peptide-binding domains by peptide or protein arrays, phage display, mass spectrometry, or other HTP techniques provides new insight into the potential protein-protein interactions prescribed by individual or even whole families of modules. Systems level analyses have also promoted a deeper understanding of the underlying principles that govern selective protein-protein interactions and how selectivity evolves. Lastly, there is a growing appreciation for the limitations and potential pitfalls associated with HTP analysis of protein-peptide interactomes. This review will examine some of the common approaches utilized for large-scale studies of PIDs and suggest a set of standards for the analysis and validation of datasets from large-scale studies of peptide-binding modules. We will also highlight how data from large-scale studies of modular interaction domain families can provide insight into systems level properties such as the linguistics of selective interactions.     

PDZ domains and their binding partners: structure, specificity, and modification

PDZ domains are abundant protein interaction modules that often recognize short amino acid motifs at the C-termini of target proteins. They regulate multiple biological processes such as transport, ion channel signaling, and other signal transduction systems. This review discusses the structural characterization of PDZ domains and the use of recently emerging technologies such as proteomic arrays and peptide libraries to study the binding properties of PDZ-mediated interactions. Regulatory mechanisms responsible for PDZ-mediated interactions, such as phosphorylation in the PDZ ligands or PDZ domains, are also discussed. A better understanding of PDZ protein-protein interaction networks and regulatory mechanisms will improve our knowledge of many cellular and biological processes.     

Predicting protein-protein interactions from protein domains using a set cover approach

One goal of contemporary proteome research is the elucidation of cellular protein interactions. Based on currently available protein-protein interaction and domain data, we introduce a novel method, maximum specificity set cover (MSSC), for the prediction of protein-protein interactions. In our approach, we map the relationship between interactions of proteins and their corresponding domain architectures to a generalized weighted set cover problem. The application of a greedy algorithm provides sets of domain interactions which explain the presence of protein interactions to the largest degree of specificity. Utilizing domain and protein interaction data of S. cerevisiae, MSSC enables prediction of previously unknown protein interactions, links that are well supported by a high tendency of coexpression and functional homogeneity of the corresponding proteins. Focusing on concrete examples, we show that MSSC reliably predicts protein interactions in well-studied molecular systems, such as the 26S proteasome and RNA polymerase II of S. cerevisiae. We also show that the quality of the predictions is comparable to the maximum likelihood estimation while MSSC is faster. This new algorithm and all data sets used are accessible through a Web portal at http://ppi-cse.nd.edu     

Unraveling lipid/protein interaction in model lipid bilayers by Atomic Force Microscopy

The current view of the biological membrane is that in which lipids and proteins mutually interact to accomplish membrane functions. The lateral heterogeneity of the lipid bilayer can induce partitioning of membrane-associated proteins, favoring protein-protein interaction and influence signaling and trafficking. The Atomic Force Microscope allows to study the localization of membrane-associated proteins with respect to the lipid organization at the single molecule level and without the need for fluorescence staining. These features make AFM a technique of choice to study lipid/protein interactions in model systems or native membranes. Here we will review the technical aspects inherent to and the main results obtained by AFM in the study of protein partitioning in lipid domains concentrating in particular on GPI-anchored proteins, lipidated proteins, and transmembrane proteins. Whenever possible, we will also discuss the functional consequences of what has been imaged by Atomic Force Microscopy.