Subgraph centrality in complex networks

Feature Selection via Swarm Intelligence for Determining Protein Essentiality

Protein essentiality is fundamental to comprehend the function and evolution of genes. The prediction of protein essentiality is pivotal in identifying disease genes and potential drug targets. Since the experimental methods need many investments in time and funds, it is of great value to predict protein essentiality with high accuracy using computational methods. In this study, we present a novel feature selection named Elite Search mechanism-based Flower Pollination Algorithm (ESFPA) to determine protein essentiality. Unlike other protein essentiality prediction methods, ESFPA uses an improved swarm intelligence⁻based algorithm for feature selection and selects optimal features for protein essentiality prediction. The first step is to collect numerous features with the highly predictive characteristics of essentiality. The second step is to develop a feature selection strategy based on a swarm intelligence algorithm to obtain the optimal feature subset. Furthermore, an elite search mechanism is adopted to further improve the quality of feature subset. Subsequently a hybrid classifier is applied to evaluate the essentiality for each protein. Finally, the experimental results show that our method is competitive to some well-known feature selection methods. The proposed method aims to provide a new perspective for protein essentiality determination.

Relating Vertex and Global Graph Entropy in Randomly Generated Graphs

Combinatoric measures of entropy capture the complexity of a graph but rely upon the calculation of its independent sets, or collections of non-adjacent vertices. This decomposition of the vertex set is a known NP-Complete problem and for most real world graphs is an inaccessible calculation. Recent work by Dehmer et al. and Tee et al. identified a number of vertex level measures that do not suffer from this pathological computational complexity, but that can be shown to be effective at quantifying graph complexity. In this paper, we consider whether these local measures are fundamentally equivalent to global entropy measures. Specifically, we investigate the existence of a correlation between vertex level and global measures of entropy for a narrow subset of random graphs. We use the greedy algorithm approximation for calculating the chromatic information and therefore Körner entropy. We are able to demonstrate strong correlation for this subset of graphs and outline how this may arise theoretically.

A new method for predicting essential proteins based on participation degree in protein complex and subgraph density

Essential proteins are crucial to living cells. Identification of essential proteins from protein-protein interaction (PPI) networks can be applied to pathway analysis and function prediction, furthermore, it can contribute to disease diagnosis and drug design. There have been some experimental and computational methods designed to identify essential proteins, however, the prediction precision remains to be improved. In this paper, we propose a new method for identifying essential proteins based on Participation degree of a protein in protein Complexes and Subgraph Density, named as PCSD. In order to test the performance of PCSD, four PPI datasets (DIP, Krogan, MIPS and Gavin) are used to conduct experiments. The experiment results have demonstrated that PCSD achieves a better performance for predicting essential proteins compared with some competing methods including DC, SC, EC, IC, LAC, NC, WDC, PeC, UDoNC, and compared with the most recent method LBCC, PCSD can correctly predict more essential proteins from certain numbers of top ranked proteins on the DIP dataset, which indicates that PCSD is very effective in discovering essential proteins in most case.

A centrality measure for cycles and subgraphs II

In a recent work we introduced a measure of importance for groups of vertices in a complex network. This centrality for groups is always between 0 and 1 and induces the eigenvector centrality over vertices. Furthermore, its value over any group is the fraction of all network flows intercepted by this group. Here we provide the rigorous mathematical constructions underpinning these results via a semi-commutative extension of a number theoretic sieve. We then established further relations between the eigenvector centrality and the centrality proposed here, showing that the latter is a proper extension of the former to groups of nodes. We finish by comparing the centrality proposed here with the notion of group-centrality introduced by Everett and Borgatti on two real-world networks: the Wolfe's dataset and the protein-protein interaction network of the yeast Saccharomyces cerevisiae. In this latter case, we demonstrate that the centrality is able to distinguish protein complexes.

Efficient disintegration strategies with cost constraint in complex networks: The crucial role of nodes near average degree

The study of network disintegration, including controlling disease spread and destroying terrorist organizations, has wide application scenarios and attracts many researchers. In this paper, we concentrate on the network disintegration problem with heterogeneous disintegration cost, where the disintegration cost to eliminate each node might be non-identical. We first put forward a disintegration cost model and an optimization model for disintegration strategy. Then, we analyze the hub strategy, leaf strategy, and the average degree strategy to investigate the nodes tendency of the optimal disintegration strategy. Numerical experiments in three synthetic networks and real-world networks indicate that the disintegration effect of hub strategy drops gradually when the disintegration cost changes from homogeneity to heterogeneity. For the situation of strong heterogeneity of disintegration cost of each node, average degree strategy achieves the maximum disintegration effect gradually. Also, taking another perspective, average degree strategy might enlighten efficient solutions to protect critical infrastructure through strengthening the nodes which are chosen by the average degree strategy.

Bioinformatics analysis and verification of key genes associated with recurrent respiratory tract infections

We evaluated the key genes related with recurrent respiratory tract infections (RRTIs), and then elucidated the possible molecular mechanisms of RRTIs. Neutrophil was isolated from peripheral bloods of the recurrent lower respiratory tract infection patients and healthy volunteers, respectively. The next generation sequencing information was obtained after RNA extraction, purification, library construction and sequencing. The sequencing information was preprocessed. Bioinformatics analysis including analysis of differentially expre-ssed genes (DEGs), Gene Ontology (GO) and pathway enrichment analysis, protein-protein interaction (PPI) analysis and transcription factors analysis were performed. The key genes were verified by real-time PCR. In total, 17 significant DEGs were obtained in case group compared with the control group by bioinformatics analysis. Then, 6 of 17 genes were detected by real-time PCR. There was statistical significance between case and control groups for peroxisome proliferator-activated receptor-γ (PPARG), prostaglandin-endoperoxide synthase 2 (PTGS2), transferrin (TF) and interleukin-10 (IL-10) (P<0.05), and there was no statistical significance between case and control groups for TIMP metallopeptidase inhibitor 1 (TIMP1) and matrix metallopeptidase 1 (MMP1). PPARG, PTGS2, TF and IL-10 are key genes associated with the progression of RRTIs. We speculate that TIMP1 and MMP1 may also be involved in the progression of RRTIs, but further studies with large number of samples are needed for verification.

Improved flower pollination algorithm for identifying essential proteins

Background

Essential proteins are necessary for the survival and development of cells. The identification of essential proteins can help to understand the minimal requirements for cellular life and it also plays an important role in the disease genes study and drug design. With the development of high-throughput techniques, a large amount of protein-protein interactions data is available to predict essential proteins at the network level. Hitherto, even though a number of essential protein discovery methods have been proposed, the prediction precision still needs to be improved.

Methods

In this paper, we propose a new algorithm, improved Flower Pollination algorithm (FPA) for identifying Essential proteins, named FPE. Different from other existing essential protein discovery methods, we apply FPA which is a new intelligent algorithm imitating pollination behavior of flowering plants in nature to identify essential proteins. Analogous to flower pollination is to find optimal reproduction from the perspective of biological evolution, and the identification of essential proteins is to discover a candidate essential protein set by analyzing the corresponding relationships between FPA algorithm and the prediction of essential proteins, and redefining the positions of flowers and specific pollination process. Moreover, it has been proved that the integration of biological and topological properties can get improved precision for identifying essential proteins. Consequently, we develop a GSC measurement in order to judge the essentiality of proteins, which takes into account not only the Gene expression data, Subcellular localization and protein Complexes information, but also the network topology.

Results

The experimental results show that FPE performs better than the state-of-the-art methods (DC, SC, IC, EC, LAC, NC, PeC, WDC, UDoNC and SON) in terms of the prediction precision, precision-recall curve and jackknife curve for identifying essential proteins and also has high stability.

Conclusions

We confirm that FPE can be used to effectively identify essential proteins by the use of nature-inspired algorithm FPA and the combination of network topology with gene expression data, subcellular localization and protein complexes information. The experimental results have shown the superiority of FPE for the prediction of essential proteins.

Predicting essential proteins by integrating orthology, gene expressions, and PPI networks

Identifying essential proteins is very important for understanding the minimal requirements of cellular life and finding human disease genes as well as potential drug targets. Experimental methods for identifying essential proteins are often costly, time-consuming, and laborious. Many computational methods for such task have been proposed based on the topological properties of protein-protein interaction networks (PINs). However, most of these methods have limited prediction accuracy due to the noisy and incomplete natures of PINs and the fact that protein essentiality may relate to multiple biological factors. In this work, we proposed a new centrality measure, OGN, by integrating orthologous information, gene expressions, and PINs together. OGN determines a protein’s essentiality by capturing its co-clustering and co-expression properties, as well as its conservation in the evolution process. The performance of OGN was tested on the species of Saccharomyces cerevisiae. Compared with several published centrality measures, OGN achieves higher prediction accuracy in both working alone and ensemble.

Where to look for power Laws in urban road networks?

Spatial embeddedness and planarity of urban road networks limit the range of their node degree values. Therefore, pursuing analysis based on the distribution of node degrees e.g. scale free aspect could not be accomplished in urban road networks. We have inspected the distribution of degree, betweenness centrality, weighted degree (based on incident link capacities), and alpha weighted degree for eight urban road networks across the world. These networks are abstracted from Philadelphia (USA), Berlin (Germany), Chicago (USA), Anaheim (USA), Gold Coast (Australia), Birmingham (UK), and Isfahan (Iran). Our results show that although the degree (weighted and unweighted) distributions of these networks are totally different, they all show power law distributions in betweenness centrality. Thus, scale free aspect could be observed in the betweenness centrality distribution. We then analyzed the collapse of network as a result of node removals. The collapse patterns suggest that critical nodes of urban road networks could not be detected solely based on betweenness centrality. Therefore, we conclude that the concept of betweenness centrality in urban road networks is more of functional merit than topological merit. In other words, central nodes play an important role in transmitting the flow but their loss would not harm the connectivity of urban networks. This claim is supported by analyzing the correlation among node flow and node betweenness in Isfahan and Anaheim.

Fronto-limbic dysconnectivity leads to impaired brain network controllability in young people with bipolar disorder and those at high genetic risk

Recent investigations have used diffusion-weighted imaging to reveal disturbances in the neurocircuitry that underlie cognitive-emotional control in bipolar disorder (BD) and in unaffected siblings or children at high genetic risk (HR). It has been difficult to quantify the mechanism by which structural changes disrupt the superimposed brain dynamics, leading to the emotional lability that is characteristic of BD. Average controllability is a concept from network control theory that extends structural connectivity data to estimate the manner in which local neuronal fluctuations spread from a node or subnetwork to alter the state of the rest of the brain. We used this theory to ask whether structural connectivity deficits previously observed in HR individuals (n = 84, mean age 22.4), patients with BD (n = 38, mean age 23.9), and age- and gender-matched controls (n = 96, mean age 22.6) translate to differences in the ability of brain systems to be manipulated between states. Localized impairments in network controllability were seen in the left parahippocampal, left middle occipital, left superior frontal, right inferior frontal, and right precentral gyri in BD and HR groups. Subjects with BD had distributed deficits in a subnetwork containing the left superior and inferior frontal gyri, postcentral gyrus, and insula (p = 0.004). HR participants had controllability deficits in a right-lateralized subnetwork involving connections between the dorsomedial and ventrolateral prefrontal cortex, the superior temporal pole, putamen, and caudate nucleus (p = 0.008). Between-group controllability differences were attenuated after removal of topological factors by network randomization. Some previously reported differences in network connectivity were not associated with controllability-differences, likely reflecting the contribution of more complex brain network properties. These analyses highlight the potential functional consequences of altered brain networks in BD, and may guide future clinical interventions.

•

Control theory estimates how neuronal fluctuations spread from local networks.

•

We compare brain controllability in bipolar disorder and their high-risk relatives.

•

These groups have impaired controllability in networks supporting cognitive and emotional control.

•

Weaker connectivity as well as topological alterations contribute to these changes.

Identifying essential proteins based on sub-network partition and prioritization by integrating subcellular localization information

Essential proteins are important participants in various life activities and play a vital role in the survival and reproduction of living organisms. Identification of essential proteins from protein-protein interaction (PPI) networks has great significance to facilitate the study of human complex diseases, the design of drugs and the development of bioinformatics and computational science. Studies have shown that highly connected proteins in a PPI network tend to be essential. A series of computational methods have been proposed to identify essential proteins by analyzing topological structures of PPI networks. However, the high noise in the PPI data can degrade the accuracy of essential protein prediction. Moreover, proteins must be located in the appropriate subcellular localization to perform their functions, and only when the proteins are located in the same subcellular localization, it is possible that they can interact with each other. In this paper, we propose a new network-based essential protein discovery method based on sub-network partition and prioritization by integrating subcellular localization information, named SPP. The proposed method SPP was tested on two different yeast PPI networks obtained from DIP database and BioGRID database. The experimental results show that SPP can effectively reduce the effect of false positives in PPI networks and predict essential proteins more accurately compared with other existing computational methods DC, BC, CC, SC, EC, IC, NC.

Identification of genes associated with primary open-angle glaucoma by bioinformatics approach

PURPOSE

This study aimed to identify associated genes with primary open-angle glaucoma (POAG) and explore the potentially modular mechanism underlying POAG.

METHODS

We downloaded gene expression profiles data GSE27276 from gene expression omnibus and identified differentially expressed genes between POAG patients and normal controls. Then, gene ontology analysis and kyoto encyclopedia of genes and genomes pathway enrichment were performed to predict the DEGs functions, followed with the construction, centrality analysis, and module mining of protein-protein interaction network.

RESULTS

A total of 552 DEGs including 249 up-regulated and 303 down-regulated genes were identified. The up-regulated DEGs were significantly involved in cell adhesion molecule, while the down-regulated DEGs were significantly involved in complement and coagulation cascades. Centrality analysis screened out 20 genes, among which COL4A4, COL3A1, COL1A2, ITGB5, COL5A2, and COL5A1 were shared in ECM-receptor interaction and focal adhesion pathways. In the sub-network, COL5A2, COL8A2, and COL5A1 were significantly enriched in biological function of eye morphogenesis and eye development, while LAMA5, COL3A1, COL1A2, and COL5A1 were significantly enriched in vasculature development and blood vessel development.

CONCLUSIONS

Six genes, including COL4A4, COL3A1, COL1A2, ITGB5, COL5A2, and COL5A1, ECM-receptor interaction and focal adhesion pathway, are potentially involved in the pathogenesis of POAG via participating in pathways of ECM-receptor interaction and focal adhesion.

Differentially categorized structural brain hubs are involved in different microstructural, functional, and cognitive characteristics and contribute to individual identification

Very little is known regarding whether structural hubs of human brain networks that enable efficient information communication may be classified into different categories. Using three multimodal neuroimaging data sets, we construct individual structural brain networks and further identify hub regions based on eight widely used graph-nodal metrics, followed by comprehensive characteristics and reproducibility analyses. We show the three categories of structural hubs in the brain network, namely, aggregated, distributed, and connector hubs. Spatially, these distinct categories of hubs are primarily located in the default-mode system and additionally in the visual and limbic systems for aggregated hubs, in the frontoparietal system for distributed hubs, and in the sensorimotor and ventral attention systems for connector hubs. These categorized hubs exhibit various distinct characteristics to support their differentiated roles, involving microstructural organization, wiring costs, topological vulnerability, functional modular integration, and cognitive flexibility; moreover, these characteristics are better in the hubs than nonhubs. Finally, all three categories of hubs display high across-session spatial similarities and act as structural fingerprints with high predictive rates (100%, 100%, and 84.2%) for individual identification. Collectively, we highlight three categories of brain hubs with differential microstructural, functional and, cognitive associations, which shed light on topological mechanisms of the human connectome.

DiffSLC: A graph centrality method to detect essential proteins of a protein-protein interaction network

Identification of central genes and proteins in biomolecular networks provides credible candidates for pathway analysis, functional analysis, and essentiality prediction. The DiffSLC centrality measure predicts central and essential genes and proteins using a protein-protein interaction network. Network centrality measures prioritize nodes and edges based on their importance to the network topology. These measures helped identify critical genes and proteins in biomolecular networks. The proposed centrality measure, DiffSLC, combines the number of interactions of a protein and the gene coexpression values of genes from which those proteins were translated, as a weighting factor to bias the identification of essential proteins in a protein interaction network. Potentially essential proteins with low node degree are promoted through eigenvector centrality. Thus, the gene coexpression values are used in conjunction with the eigenvector of the network's adjacency matrix and edge clustering coefficient to improve essentiality prediction. The outcome of this prediction is shown using three variations: (1) inclusion or exclusion of gene co-expression data, (2) impact of different coexpression measures, and (3) impact of different gene expression data sets. For a total of seven networks, DiffSLC is compared to other centrality measures using Saccharomyces cerevisiae protein interaction networks and gene expression data. Comparisons are also performed for the top ranked proteins against the known essential genes from the Saccharomyces Gene Deletion Project, which show that DiffSLC detects more essential proteins and has a higher area under the ROC curve than other compared methods. This makes DiffSLC a stronger alternative to other centrality methods for detecting essential genes using a protein-protein interaction network that obeys centrality-lethality principle. DiffSLC is implemented using the igraph package in R, and networkx package in Python. The python package can be obtained from git.io/diffslcpy. The R implementation and code to reproduce the analysis is available via git.io/diffslc.

Centralities in simplicial complexes. Applications to protein interaction networks

Complex networks can be used to represent complex systems which originate in the real world. Here we study a transformation of these complex networks into simplicial complexes, where cliques represent the simplices of the complex. We extend the concept of node centrality to that of simplicial centrality and study several mathematical properties of degree, closeness, betweenness, eigenvector, Katz, and subgraph centrality for simplicial complexes. We study the degree distributions of these centralities at the different levels. We also compare and describe the differences between the centralities at the different levels. Using these centralities we study a method for detecting essential proteins in PPI networks of cells and explain the varying abilities of the centrality measures at the different levels in identifying these essential proteins.

RNA Sequencing Uncovers Molecular Mechanisms Underlying Pathological Complete Response to Chemotherapy in Patients with Operable Breast Cancer

Background

This study aimed to identify key genes contributing to pathological complete response (pCR) to chemotherapy by mRNA sequencing (RNA-seq).

Material/Methods

RNA was extracted from the frozen biopsy tissue of patients with pathological complete response and patients with non-pathological complete response. Sequencing was performed on the HiSeq2000 platform. Differentially expressed genes (DEGs) were identified between the pCR group and non-pCR (NpCR) group. Pathway enrichment analysis of DEGs was performed. A protein-protein interaction network was constructed, then module analysis was performed to identify a subnetwork. Finally, transcription factors were predicted.

Results

A total of 673 DEGs were identified, including 419 upregulated ones and 254 downregulated ones. The PPI network constructed consisted of 276 proteins forming 471 PPI pairs, and a subnetwork containing 18 protein nodes was obtained. Pathway enrichment analysis revealed that PLCB4 and ADCY6 were enriched in pathways renin secretion, gastric acid secretion, gap junction, inflammatory mediator regulation of TRP channels, retrograde endocannabinoid signaling, melanogenesis, cGMP-PKG signaling pathway, calcium signaling pathway, chemokine signaling pathway, cAMP signaling pathway, and rap1 signaling pathway. CNR1 was enriched in the neuroactive ligand-receptor interaction pathway, retrograde endocannabinoid signaling pathway, and rap1 signaling pathway. The transcription factor-gene network consists of 15 transcription factors and 16 targeted genes, of which 5 were downregulated and 10 were upregulated.

Conclusions

We found key genes that may contribute to pCR to chemotherapy, such as PLCB4, ADCY6, and CNR1, as well as some transcription factors.

Initial gut microbiota structure affects sensitivity to DSS-induced colitis in a mouse model

The dextran sulfate sodium (DSS)-induced colitis model is a widely applied mouse model, but controversial results have been obtained from experiments using the same mouse strain under the same conditions. Because the gut microbiota play an important role in DSS-induced colitis, it is essential to evaluate the influence of the initial gut microbiota in this model. Here, we identified significant variations in the initial gut microbiota of different batches of mice and found that the initial intestinal microbiota had a profound influence on DSS-induced colitis. We performed three independent trials using the same C57BL/6J mouse model with DSS treatment and used high-throughput 16S rRNA gene sequencing to analyze the gut microbiota. We found that the structure and composition of the gut microbiota in mice with severe colitis, as compared with mice with milder colon damage, had unique features, such as an increase in Akkermansia bacteria and a decrease in Barnesiella spp. Moreover, these varied gut bacteria in the different trials also showed different responses to DSS treatment. Our work suggests that, in studies using mouse models, the gut microbiota must be considered when examining mechanisms of diseases, to ensure that comparable results are obtained.

Financial fluctuations anchored to economic fundamentals: A mesoscopic network approach

We demonstrate the existence of an empirical linkage between nominal financial networks and the underlying economic fundamentals, across countries. We construct the nominal return correlation networks from daily data to encapsulate sector-level dynamics and infer the relative importance of the sectors in the nominal network through measures of centrality and clustering algorithms. Eigenvector centrality robustly identifies the backbone of the minimum spanning tree defined on the return networks as well as the primary cluster in the multidimensional scaling map. We show that the sectors that are relatively large in size, defined with three metrics, viz., market capitalization, revenue and number of employees, constitute the core of the return networks, whereas the periphery is mostly populated by relatively smaller sectors. Therefore, sector-level nominal return dynamics are anchored to the real size effect, which ultimately shapes the optimal portfolios for risk management. Our results are reasonably robust across 27 countries of varying degrees of prosperity and across periods of market turbulence (2008-09) as well as periods of relative calmness (2012-13 and 2015-16).

Network-based analysis of the molecular mechanisms of multiple myeloma and monoclonal gammopathy of undetermined significance

The present study aimed to reveal the molecular mechanisms of multiple myeloma (MM) and monoclonal gammopathy of undetermined significance (MGUS). This was a secondary study on microarray dataset GSE80608, downloaded from the Gene Expression Omnibus database, which included 10 control samples, 10 MGUS samples and 10 MM samples. Differentially expressed genes (DEGs) were identified between control and MGUS samples, and between control and MM samples. A protein-protein interaction (PPI) network was built for studying the interactions between the DEGs. Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis was performed for the genes in a gene co-expression network. A microRNA (miRNA/miR)-gene network was built to the evaluate possible the miRNAs and genes involved in the diseases. The present study identified 136 common upregulated DEGs and 165 common downregulated DEGs between MM and MGUS. Pathway enrichment analysis of the genes in the gene co-expression network revealed that the complement and coagulation cascades pathway was significantly enriched for certain complement and coagulation-associated genes. Endothelin-1 (EDN1) was significantly enriched in the hypoxia inducible factor-1 (HIF-1) and tumor necrosis factor signaling pathways. EDN1 was an important node in the PPI network, and a target gene of let-7e, let-7b and miR-19a in the miRNA-gene network. The results of the present study indicate that complement and coagulation-associated genes, the complement and coagulation cascades pathway, EDN1, let-7e, let-7b-5p, miR-19a, and the tumor necrosis factor and HIF-1 signaling pathways may all be implicated in MM and MGUS.

Task modulates functional connectivity networks in free viewing behavior

In free visual exploration, eye-movement is immediately followed by dynamic reconfiguration of brain functional connectivity. We studied the task-dependency of this process in a combined visual search-change detection experiment. Participants viewed two (nearly) same displays in succession. First time they had to find and remember multiple targets among distractors, so the ongoing task involved memory encoding. Second time they had to determine if a target had changed in orientation, so the ongoing task involved memory retrieval. From multichannel EEG recorded during 200 ms intervals time-locked to fixation onsets, we estimated the functional connectivity using a weighted phase lag index at the frequencies of theta, alpha, and beta bands, and derived global and local measures of the functional connectivity graphs. We found differences between both memory task conditions for several network measures, such as mean path length, radius, diameter, closeness and eccentricity, mainly in the alpha band. Both the local and the global measures indicated that encoding involved a more segregated mode of operation than retrieval. These differences arose immediately after fixation onset and persisted for the entire duration of the lambda complex, an evoked potential commonly associated with early visual perception. We concluded that encoding and retrieval differentially shape network configurations involved in early visual perception, affecting the way the visual input is processed at each fixation. These findings demonstrate that task requirements dynamically control the functional connectivity networks involved in early visual perception.

Link Prediction in Evolving Networks Based on Popularity of Nodes

Link prediction aims to uncover the underlying relationship behind networks, which could be utilized to predict missing edges or identify the spurious edges. The key issue of link prediction is to estimate the likelihood of potential links in networks. Most classical static-structure based methods ignore the temporal aspects of networks, limited by the time-varying features, such approaches perform poorly in evolving networks. In this paper, we propose a hypothesis that the ability of each node to attract links depends not only on its structural importance, but also on its current popularity (activeness), since active nodes have much more probability to attract future links. Then a novel approach named popularity based structural perturbation method (PBSPM) and its fast algorithm are proposed to characterize the likelihood of an edge from both existing connectivity structure and current popularity of its two endpoints. Experiments on six evolving networks show that the proposed methods outperform state-of-the-art methods in accuracy and robustness. Besides, visual results and statistical analysis reveal that the proposed methods are inclined to predict future edges between active nodes, rather than edges between inactive nodes.

A new computational strategy for identifying essential proteins based on network topological properties and biological information

Essential proteins are the proteins that are indispensable to the survival and development of an organism. Deleting a single essential protein will cause lethality or infertility. Identifying and analysing essential proteins are key to understanding the molecular mechanisms of living cells. There are two types of methods for predicting essential proteins: experimental methods, which require considerable time and resources, and computational methods, which overcome the shortcomings of experimental methods. However, the prediction accuracy of computational methods for essential proteins requires further improvement. In this paper, we propose a new computational strategy named CoTB for identifying essential proteins based on a combination of topological properties, subcellular localization information and orthologous protein information. First, we introduce several topological properties of the protein-protein interaction (PPI) network. Second, we propose new methods for measuring orthologous information and subcellular localization and a new computational strategy that uses a random forest prediction model to obtain a probability score for the proteins being essential. Finally, we conduct experiments on four different Saccharomyces cerevisiae datasets. The experimental results demonstrate that our strategy for identifying essential proteins outperforms traditional computational methods and the most recently developed method, SON. In particular, our strategy improves the prediction accuracy to 89, 78, 79, and 85 percent on the YDIP, YMIPS, YMBD and YHQ datasets at the top 100 level, respectively.

The Significant Pathways and Genes Underlying the Colon Cancer Treatment by the Traditional Chinese Medicine PHY906

BACKGROUND

We attempted to explore the molecular mechanism underlying PHY906 intervention of colon cancer.

METHODS

The microarray data of tumors treated by PHY906 and PBS alone were downloaded from the public Gene Expression Omnibus database. The dataset was further analyzed for the differentially expressed genes (DEGs) and their related biological functions were analyzed, followed by function and pathways. Protein-protein interaction (PPI) network was constructed and the significant nodes were screened by network centralities and then the significant modules analysis. Besides, they were clustered and transcriptional factors (TFs) were predicted.

RESULTS

The gene expression patterns changed induced by PHY906 treatment, including 414 upregulated and 337 downregulated DEGs. The biological process of response to steroid hormone stimulus and regulation of interferon-gamma production were significantly enriched by DEGs. Ezh2 (enhancer of zeste 2) was found to be the key node in PPI network. There are 12 significant TFs predicted for module 1 genes and 3 TFs for module 2 genes.

CONCLUSIONS

PHY906 treatment may function in protecting the epithelial barrier against tumor cell invasion by modulating IFN-γ level and mediating cancer cell death by activating the response to steroid hormone stimulus and activating the response to steroid hormone stimulus. E2f1, Hsfy2, and Nfyb may be therapeutic targets for colon cancer. PHY906 showed treatment efficacy in modulating cell apoptosis by intervening interferon-gamma production and response to steroid hormone stimulus. Ezh2 and its TFs such as E2f1, Hsfy2, and Nfyb may be the potential therapeutic targets for anticancer agents development.

Identification of novel genes associated with fracture healing in osteoporosis induced by Krm2 overexpression or Lrp5 deficiency

The aim of the present study was to screen potential key genes associated with osteoporotic fracture healing. The microarray data from the Gene Expression Omnibus database accession number GSE51686, were downloaded and used to identify differentially expressed genes (DEGs) in fracture callus tissue samples obtained from the femora of type I collagen (Col1a1)-kringle containing transmembrane protein 2 (Krm2) mice and low density lipoprotein receptor-related protein 5^−/− (Lrp5^−/−) transgenic mice of osteoporosis compared with those in wild-type (WT) mice. Enrichment analysis was performed to reveal the DEG function. In addition, protein-protein interactions (PPIs) of DEGs were analyzed using the Search Tool for the Retrieval of Interacting Genes database. The coexpression associations between hub genes in the PPI network were investigated, and a coexpression network was constructed. A total of 841 DEGs (335 upregulated and 506 downregulated) were identified in the Col1a1-Krm2 vs. the WT group, and 50 DEGs (16 upregulated and 34 downregulated) were identified in the Lrp5^−/− vs. the WT group. The DEGs in Col1a1-Krm2 mice were primarily associated with immunity and cell adhesion (GO: 0007155) functions. By contrast, the DEGs in Lrp5^−/− mice were significantly associated with muscle system process (GO: 0003012) and regulation of transcription (GO: 0006355). In addition, a series of DEGs demonstrated a higher score in the PPI network, and were observed to be coexpressed in the coexpression network, and included thrombospondin 2 (Thbs2), syndecan 2 (Sdc2), FK506 binding protein 10 (Fkbp10), 2–5-oligoadneylate synthase-like protein 2 (Oasl2), interferon induced protein with tetratricopeptide repeats (Ifit) 1 and Ifit2. Thbs2 and Sdc2 were significantly correlated with extracellular matrix-receptor interactions. The results suggest that Thbs2, Sdc2, Fkbp10, Oasl2, Ifit1 and Ifit2 may serve important roles during the fracture healing process in osteoporosis. In addition, this is the first study to demonstrate that Sdc2, Fkbp10, Oasl2, Ifit1 and Ifit2 may be associated with osteoporotic fracture healing.

Graph Theory at the Service of Electroencephalograms

The brain is one of the largest and most complex organs in the human body and EEG is a noninvasive electrophysiological monitoring method that is used to record the electrical activity of the brain. Lately, the functional connectivity in human brain has been regarded and studied as a complex network using EEG signals. This means that the brain is studied as a connected system where nodes, or units, represent different specialized brain regions and links, or connections, represent communication pathways between the nodes. Graph theory and theory of complex networks provide a variety of measures, methods, and tools that can be useful to efficiently model, analyze, and study EEG networks. This article is addressed to computer scientists who wish to be acquainted and deal with the study of EEG data and also to neuroscientists who would like to become familiar with graph theoretic approaches and tools to analyze EEG data.

Identification of key genes and pathways for peri-implantitis through the analysis of gene expression data

The present study attempted to identify potential key genes and pathways of peri-implantitis, and to investigate the possible mechanisms associated with it. An array data of GSE57631 was downloaded, including six samples of peri-implantitis tissue and two samples of normal tissue from the Gene Expression Omnibus (GEO) database. The differentially expressed genes (DEGs) in the peri-implantitis samples compared with normal ones were analyzed with the limma package. Moreover, Gene Ontology annotation and the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses for DEGs were performed by DAVID. A protein-protein interaction (PPI) network was established using Cytoscape software, and significant modules were analyzed using Molecular Complex Detection. A total of 819 DEGs (759 upregulated and 60 downregulated) were identified in the peri-implantitis samples compared with normal ones. Moreover, the PPI network was constructed with 413 nodes and 1,114 protein pairs. Heat shock protein HSP90AA1 (90 kDa α, member 1), a hub node with higher node degrees in module 4, was significantly enriched in antigen processing, in the presentation pathway and nucleotide-binding oligomerization domain (NOD)-like receptor-signaling pathway. In addition, nuclear factor-κ-B1 (NFKB1) was enriched in the NOD-like receptor-signaling pathway in KEGG pathway enrichment analysis for upregulated genes. The proteasome is the most significant pathway in module 1 with the highest P-value. Therefore, the results of the present study suggested that HSP90AA1 and NFKB1 may be potential key genes, and the NOD-like receptor signaling pathway and proteasome may be potential pathways associated with peri-implantitis development.

United Complex Centrality for Identification of Essential Proteins from PPI Networks

Essential proteins are indispensable for the survival or reproduction of an organism. Identification of essential proteins is not only necessary for the understanding of the minimal requirements for cellular life, but also important for the disease study and drug design. With the development of high-throughput techniques, a large number of protein-protein interaction data are available, which promotes the studies of essential proteins from the network level. Up to now, though a series of computational methods have been proposed, the prediction precision still needs to be improved. In this paper, we propose a new method, United complex Centrality (UC), to identify essential proteins by integrating the protein complexes with the topological features of protein-protein interaction (PPI) networks. By analyzing the relationship between the essential proteins and the known protein complexes of S. cerevisiae and human, we find that the proteins in complexes are more likely to be essential compared with the proteins not included in any complexes and the proteins appeared in multiple complexes are more inclined to be essential compared to those only appeared in a single complex. Considering that some protein complexes generated by computational methods are inaccurate, we also provide a modified version of UC with parameter alpha, named UC-P. The experimental results show that protein complex information can help identify the essential proteins more accurate both for the PPI network of S. cerevisiae and that of human. The proposed method UC performs obviously better than the eight previously proposed methods (DC, IC, EC, SC, BC, CC, NC, and LAC) for identifying essential proteins.

A morphometric analysis of vegetation patterns in dryland ecosystems

Vegetation in dryland ecosystems often forms remarkable spatial patterns. These range from regular bands of vegetation alternating with bare ground, to vegetated spots and labyrinths, to regular gaps of bare ground within an otherwise continuous expanse of vegetation. It has been suggested that spotted vegetation patterns could indicate that collapse into a bare ground state is imminent, and the morphology of spatial vegetation patterns, therefore, represents a potentially valuable source of information on the proximity of regime shifts in dryland ecosystems. In this paper, we have developed quantitative methods to characterize the morphology of spatial patterns in dryland vegetation. Our approach is based on algorithmic techniques that have been used to classify pollen grains on the basis of textural patterning, and involves constructing feature vectors to quantify the shapes formed by vegetation patterns. We have analysed images of patterned vegetation produced by a computational model and a small set of satellite images from South Kordofan (South Sudan), which illustrates that our methods are applicable to both simulated and real-world data. Our approach provides a means of quantifying patterns that are frequently described using qualitative terminology, and could be used to classify vegetation patterns in large-scale satellite surveys of dryland ecosystems.

Exploring the "Middle Earth" of network spectra via a Gaussian matrix function

We study a Gaussian matrix function of the adjacency matrix of artificial and real-world networks. We motivate the use of this function on the basis of a dynamical process modeled by the time-dependent Schrödinger equation with a squared Hamiltonian. In particular, we study the Gaussian Estrada index-an index characterizing the importance of eigenvalues close to zero. This index accounts for the information contained in the eigenvalues close to zero in the spectra of networks. Such a method is a generalization of the so-called "Folded Spectrum Method" used in quantum molecular sciences. Here, we obtain bounds for this index in simple graphs, proving that it reaches its maximum for star graphs followed by complete bipartite graphs. We also obtain formulas for the Estrada Gaussian index of Erdős-Rényi random graphs and for the Barabási-Albert graphs. We also show that in real-world networks, this index is related to the existence of important structural patterns, such as complete bipartite subgraphs (bicliques). Such bicliques appear naturally in many real-world networks as a consequence of the evolutionary processes giving rise to them. In general, the Gaussian matrix function of the adjacency matrix of networks characterizes important structural information not described in previously used matrix functions of graphs.

Graph-theoretical model of global human interactome reveals enhanced long-range communicability in cancer networks

Malignant transformation is known to involve substantial rearrangement of the molecular genetic landscape of the cell. A common approach to analysis of these alterations is a reductionist one and consists of finding a compact set of differentially expressed genes or associated signaling pathways. However, due to intrinsic tumor heterogeneity and tissue specificity, biomarkers defined by a small number of genes/pathways exhibit substantial variability. As an alternative to compact differential signatures, global features of genetic cell machinery are conceivable. Global network descriptors suggested in previous works are, however, known to potentially be biased by overrepresentation of interactions between frequently studied genes-proteins. Here, we construct a cellular network of 74538 directional and differential gene expression weighted protein-protein and gene regulatory interactions, and perform graph-theoretical analysis of global human interactome using a novel, degree-independent feature—the normalized total communicability (NTC). We apply this framework to assess differences in total information flow between different cancer (BRCA/COAD/GBM) and non-cancer interactomes. Our experimental results reveal that different cancer interactomes are characterized by significant enhancement of long-range NTC, which arises from circulation of information flow within robustly organized gene subnetworks. Although enhancement of NTC emerges in different cancer types from different genomic profiles, we identified a subset of 90 common genes that are related to elevated NTC in all studied tumors. Our ontological analysis shows that these genes are associated with enhanced cell division, DNA replication, stress response, and other cellular functions and processes typically upregulated in cancer. We conclude that enhancement of long-range NTC manifested in the correlated activity of genes whose tight coordination is required for survival and proliferation of all tumor cells, and, thus, can be seen as a graph-theoretical equivalent to some hallmarks of cancer. The computational framework for differential network analysis presented herein is of potential interest for a wide range of network perturbation problems given by single or multiple gene-protein activation-inhibition.

Analysis of gene expression profiles of non-small cell lung cancer at different stages reveals significantly altered biological functions and candidate genes

We attempt to dissect the pathology of non-small cell lung cancer (NSCLC) patients at different stages and discover the novel candidate genes. Microarray data (GSE21933) were retrieved from the Gene Expression Omnibus database. The differential expression profiles of lung tumor tissues during different stages were analyzed. The significantly altered functions and pathways were assessed and the key nodes in a protein-protein interaction (PPI) network were screened out. Then, the coexpression gene pairs and tumor-related genes were assessed. RT-PCR analysis was performed to validate the candidate gene, natural killer-tumor recognition sequence (NKTR). The number of differentially expressed genes (DEGs) for stage IB, IIB, IIIA and IV tumors were 499, 602, 592 and 457, respectively. Most of the DEGs were NSCLC-related genes identified through literature research. A few genes were commonly downregulated in all the 4 stages of tumors, such as CNTN6 and LBX2. The DEGs in early‑stage tumors were closely related with the negative regulation of signal transduction, the apoptosis pathway and the p53 signaling pathway. DEGs in late-stage tumors were significantly enriched in transcription, response to organic substances and synapse regulation-related biological processes. A total of 16 genes (including NKTR) made up the significant coexpression network. NKTR was a key node in the PPI network and was significantly upregulated in lung cancer cells. The mechanism of NSCLC progression in different tumor stages may be different. NKTR may be the target candidate for NSCLC prevention and treatment.

Using LTI Dynamics to Identify the Influential Nodes in a Network

Networks are used for modeling numerous technical, social or biological systems. In order to better understand the system dynamics, it is a matter of great interest to identify the most important nodes within the network. For a large set of problems, whether it is the optimal use of available resources, spreading information efficiently or even protection from malicious attacks, the most important node is the most influential spreader, the one that is capable of propagating information in the shortest time to a large portion of the network. Here we propose the Node Imposed Response (NiR), a measure which accurately evaluates node spreading power. It outperforms betweenness, degree, k-shell and h-index centrality in many cases and shows the similar accuracy to dynamics-sensitive centrality. We utilize the system-theoretic approach considering the network as a Linear Time-Invariant system. By observing the system response we can quantify the importance of each node. In addition, our study provides a robust tool set for various protective strategies.

MORO: a Cytoscape app for relationship analysis between modularity and robustness in large-scale biological networks

BACKGROUND

Although there have been many studies revealing that dynamic robustness of a biological network is related to its modularity characteristics, no proper tool exists to investigate the relation between network dynamics and modularity.

RESULTS

Accordingly, we developed a novel Cytoscape app called MORO, which can conveniently analyze the relationship between network modularity and robustness. We employed an existing algorithm to analyze the modularity of directed graphs and a Boolean network model for robustness calculation. In particular, to ensure the robustness algorithm's applicability to large-scale networks, we implemented it as a parallel algorithm by using the OpenCL library. A batch-mode simulation function was also developed to verify whether an observed relationship between modularity and robustness is conserved in a large set of randomly structured networks. The app provides various visualization modes to better elucidate topological relations between modules, and tabular results of centrality and gene ontology enrichment analyses of modules. We tested the proposed app to analyze large signaling networks and showed an interesting relationship between network modularity and robustness.

CONCLUSIONS

Our app can be a promising tool which efficiently analyzes the relationship between modularity and robustness in large signaling networks.

Computational Discovery of Putative Leads for Drug Repositioning through Drug-Target Interaction Prediction

De novo experimental drug discovery is an expensive and time-consuming task. It requires the identification of drug-target interactions (DTIs) towards targets of biological interest, either to inhibit or enhance a specific molecular function. Dedicated computational models for protein simulation and DTI prediction are crucial for speed and to reduce the costs associated with DTI identification. In this paper we present a computational pipeline that enables the discovery of putative leads for drug repositioning that can be applied to any microbial proteome, as long as the interactome of interest is at least partially known. Network metrics calculated for the interactome of the bacterial organism of interest were used to identify putative drug-targets. Then, a random forest classification model for DTI prediction was constructed using known DTI data from publicly available databases, resulting in an area under the ROC curve of 0.91 for classification of out-of-sampling data. A drug-target network was created by combining 3,081 unique ligands and the expected ten best drug targets. This network was used to predict new DTIs and to calculate the probability of the positive class, allowing the scoring of the predicted instances. Molecular docking experiments were performed on the best scoring DTI pairs and the results were compared with those of the same ligands with their original targets. The results obtained suggest that the proposed pipeline can be used in the identification of new leads for drug repositioning. The proposed classification model is available at http://bioinformatics.ua.pt/software/dtipred/.

The emergence of multi-resistant bacterial strains and the existing void in the discovery and development of new classes of antibiotics is a growing concern. Indeed, some bacterial strains are now resistant to last-line antibiotics and considered untreatable. Drug repositioning has been suggested as a strategy to minimize time and cost expenses until the drug reaches the market, compared to traditional drug design. Drug-target interactions (DTIs) are the basis of rational drug design and thus, we proposed a computational approach to predict DTIs solely based on the primary sequence of the protein and the simplified molecular-input line-entry system of the ligand. In addition, network metrics are used to identify vital putative drug-targets in bacteria. Molecular docking experiments were performed to compare the binding affinities between a given ligand and a putative drug-target, as well as with their original targets. According to the docking results, the predicted DTIs have better or similar binding activities than the ligand and their real target, indicating the validity of the proposed model.

Graph theory and cognition: A complementary avenue for examining neuropsychological status in epilepsy

The recent revision of the classification of the epilepsies released by the ILAE Commission on Classification and Terminology (2005-2009) has been a major development in the field. Papers in this section of the special issue explore the relevance of other techniques to examine, categorize, and classify cognitive and behavioral comorbidities in epilepsy. In this review, we investigate the applicability of graph theory to understand the impact of epilepsy on cognition compared with controls and, then, the patterns of cognitive development in normally developing children which would set the stage for prospective comparisons of children with epilepsy and controls. The overall goal is to examine the potential utility of this analytic tool and approach to conceptualize the cognitive comorbidities in epilepsy. Given that the major cognitive domains representing cognitive function are interdependent, the associations between neuropsychological abilities underlying these domains can be referred to as a cognitive network. Therefore, the architecture of this cognitive network can be quantified and assessed using graph theory methods, rendering a novel approach to the characterization of cognitive status. We first provide fundamental information about graph theory procedures, followed by application of these techniques to cross-sectional analysis of neuropsychological data in children with epilepsy compared with that of controls, concluding with prospective analysis of neuropsychological development in younger and older healthy controls. This article is part of a Special Issue entitled "The new approach to classification: Rethinking cognition and behavior in epilepsy".

Prediction of Essential Proteins Based on Local Interaction Density

Prediction of essential proteins which is aided by computer science and supported from high throughput data is a more efficient method compared with time consuming and expensive experimental approaches. There are many computational approaches reported, however they are usually sensitive to various network structures so that their robustness are generally poor. In this paper, a novel topological centrality measure for predicting essential proteins based on local interaction density, named as LID, is proposed. It is different from previous measures that LID takes the essentiality of a node from interaction densities among its neighbors through topological analyses of real proteins in a protein complex set first time at the viewpoint of biological modules. LID is applied to four different yeast protein interaction networks, which are obtained, respectively, from the DIP database and the BioGRID database. The experimental results show that the number of essential proteins detected by LID universally exceeds or approximates the best performance of other 10 topological centrality measures in all 24 comparisons of four networks: DC, BC, ClusterC, CloseC, MNC, SoECC(NC), LAC, SC, EigC, and InfoC. The better robustness of LID for multiple data sets will make it to be a new core topological centrality measure to improve the performance of prediction for more species protein interaction networks.

Disruptions in cortico-subcortical covariance networks associated with anxiety in new-onset childhood epilepsy

Anxiety disorders represent a prevalent psychiatric comorbidity in both adults and children with epilepsy for which the etiology remains controversial. Neurobiological contributions have been suggested, but only limited evidence suggests abnormal brain volumes particularly in children with epilepsy and anxiety. Since the brain develops in an organized fashion, covariance analyses between different brain regions can be investigated as a network and analyzed using graph theory methods. We examined 46 healthy children (HC) and youth with recent onset idiopathic epilepsies with (n = 24) and without (n = 62) anxiety disorders. Graph theory (GT) analyses based on the covariance between the volumes of 85 cortical/subcortical regions were investigated. Both groups with epilepsy demonstrated less inter-modular relationships in the synchronization of cortical/subcortical volumes compared to controls, with the epilepsy and anxiety group presenting the strongest modular organization. Frontal and occipital regions in non-anxious epilepsy, and areas throughout the brain in children with epilepsy and anxiety, showed the highest centrality compared to controls. Furthermore, most of the nodes correlating to amygdala volumes were subcortical structures, with the exception of the left insula and the right frontal pole, which presented high betweenness centrality (BC); therefore, their influence in the network is not necessarily local but potentially influencing other more distant regions. In conclusion, children with recent onset epilepsy and anxiety demonstrate large scale disruptions in cortical and subcortical brain regions. Network science may not only provide insight into the possible neurobiological correlates of important comorbidities of epilepsy, but also the ways that cortical and subcortical disruption occurs.

A novel topological centrality measure capturing biologically important proteins

Topological centrality in protein interaction networks and its biological implications have widely been investigated in the past. In the present study, a novel metric of centrality-weighted sum of loads eigenvector centrality (WSL-EC)-based on graph spectra is defined and its performance in identifying topologically and biologically important nodes is comparatively investigated with common metrics of centrality in a human protein-protein interaction network. The metric can capture nodes from peripherals of the network differently from conventional eigenvector centrality. Different metrics were found to selectively identify hub sets that are significantly associated with different biological processes. The widely accepted metrics degree centrality, betweenness centrality, subgraph centrality and eigenvector centrality are subject to a bias towards super-hubs, whereas WSL-EC is not affected by the presence of super-hubs. WSL-EC outperforms other metrics of centrality in detecting biologically central nodes such as pathogen-interacting, cancer, ageing, HIV-1 or disease-related proteins and proteins involved in immune system processes and autoimmune diseases in the human interactome.

Predicting essential proteins based on subcellular localization, orthology and PPI networks

Background

Essential proteins play an indispensable role in the cellular survival and development. There have been a series of biological experimental methods for finding essential proteins; however they are time-consuming, expensive and inefficient. In order to overcome the shortcomings of biological experimental methods, many computational methods have been proposed to predict essential proteins. The computational methods can be roughly divided into two categories, the topology-based methods and the sequence-based ones. The former use the topological features of protein-protein interaction (PPI) networks while the latter use the sequence features of proteins to predict essential proteins. Nevertheless, it is still challenging to improve the prediction accuracy of the computational methods.

Results

Comparing with nonessential proteins, essential proteins appear more frequently in certain subcellular locations and their evolution more conservative. By integrating the information of subcellular localization, orthologous proteins and PPI networks, we propose a novel essential protein prediction method, named SON, in this study. The experimental results on S.cerevisiae data show that the prediction accuracy of SON clearly exceeds that of nine competing methods: DC, BC, IC, CC, SC, EC, NC, PeC and ION.

Conclusions

We demonstrate that, by integrating the information of subcellular localization, orthologous proteins with PPI networks, the accuracy of predicting essential proteins can be improved. Our proposed method SON is effective for predicting essential proteins.

An ensemble framework for identifying essential proteins

Background

Many centrality measures have been proposed to mine and characterize the correlations between network topological properties and protein essentiality. However, most of them show limited prediction accuracy, and the number of common predicted essential proteins by different methods is very small.

Results

In this paper, an ensemble framework is proposed which integrates gene expression data and protein-protein interaction networks (PINs). It aims to improve the prediction accuracy of basic centrality measures. The idea behind this ensemble framework is that different protein-protein interactions (PPIs) may show different contributions to protein essentiality. Five standard centrality measures (degree centrality, betweenness centrality, closeness centrality, eigenvector centrality, and subgraph centrality) are integrated into the ensemble framework respectively. We evaluated the performance of the proposed ensemble framework using yeast PINs and gene expression data. The results show that it can considerably improve the prediction accuracy of the five centrality measures individually. It can also remarkably increase the number of common predicted essential proteins among those predicted by each centrality measure individually and enable each centrality measure to find more low-degree essential proteins.

Conclusions

This paper demonstrates that it is valuable to differentiate the contributions of different PPIs for identifying essential proteins based on network topological characteristics. The proposed ensemble framework is a successful paradigm to this end.

Electronic supplementary material

The online version of this article (doi:10.1186/s12859-016-1166-7) contains supplementary material, which is available to authorized users.