Protein-protein interaction networks and subnetworks in the biology of disease

Gene expression networks regulated by human personality

Genome-wide association studies of human personality have been carried out, but transcription of the whole genome has not been studied in relation to personality in humans. We collected genome-wide expression profiles of adults to characterize the regulation of expression and function in genes related to human personality. We devised an innovative multi-omic approach to network analysis to identify the key control elements and interactions in multi-modular networks. We identified sets of transcribed genes that were co-expressed in specific brain regions with genes known to be associated with personality. Then we identified the minimum networks for the co-localized genes using bioinformatic resources. Subjects were 459 adults from the Young Finns Study who completed the Temperament and Character Inventory and provided peripheral blood for genomic and transcriptomic analysis. We identified an extrinsic network of 45 regulatory genes from seed genes in brain regions involved in self-regulation of emotional reactivity to extracellular stimuli (e.g., self-regulation of anxiety) and an intrinsic network of 43 regulatory genes from seed genes in brain regions involved in self-regulation of interpretations of meaning (e.g., production of concepts and language). We discovered that interactions between the two networks were coordinated by a control hub of 3 miRNAs and 3 protein-coding genes shared by both. Interactions of the control hub with proteins and ncRNAs identified more than 100 genes that overlap directly with known personality-related genes and more than another 4000 genes that interact indirectly. We conclude that the six-gene hub is the crux of an integrative network that orchestrates information-transfer throughout a multi-modular system of over 4000 genes enriched in liquid-liquid-phase-separation (LLPS)-related RNAs, diverse transcription factors, and hominid-specific miRNAs and lncRNAs. Gene expression networks associated with human personality regulate neuronal plasticity, epigenesis, and adaptive functioning by the interactions of salience and meaning in self-awareness.

Gene expression profile analysis unravelled the systems level association of renal cell carcinoma with diabetic nephropathy and Matrix-metalloproteinase-9 as a potential therapeutic target

Type 2 diabetes (T2D) and cancer share many common risk factors. However, the potential biological link that connects the two at the molecular level is still unclear. The experimental evidence suggests that several genes and their pathways may be involved in developing cancerous conditions associated with diabetes. In this study, we identified the protein-protein interaction (PPI) networks and the hub protein(s) that interlink T2D and cancer using genome-scale differential gene expression profiles. Further, the PPI network of AMP-activated protein kinase (AMPK) in cancer was analyzed to explore novel insights into the molecular association between the two conditions. The densely connected regions were analyzed by constructing the backbone and subnetworks with key nodes and shortest pathways, respectively. The PPI network studies identified Matrix-metalloproteinase-9 (MMP-9) as a hub protein playing a vital role in glomerulonephritis tubular diseases and some genetic kidney diseases. MMP-9 was also associated with different growth factors, like tumor necrosis factor (TNF-α), transforming growth factor 1 (TGF-1), and pathways like chemokine signaling, NOD-like receptor signaling, etc. Further, the molecular docking and molecular dynamic simulation studies supported the druggability of MMP-9, suggesting it as a potential therapeutic target in treating renal cell carcinoma linked with diabetic kidney disease.Communicated by Ramaswamy H. Sarma.

Early diagnostic and prognostic biomarkers for gastric cancer: systems-level molecular basis of subsequent alterations in gastric mucosa from chronic atrophic gastritis to gastric cancer

PURPOSE

It is important to comprehend how the molecular mechanisms shift when gastric cancer in its early stages (GC). We employed integrative bioinformatics approaches to locate various biological signalling pathways and molecular fingerprints to comprehend the pathophysiology of the GC. To facilitate the discovery of their possible biomarkers, a rapid diagnostic may be made, which leads to an improved diagnosis and improves the patient's prognosis.

METHODS

Through protein-protein interaction networks, functional differentially expressed genes (DEGs), and pathway enrichment studies, we examined the gene expression profiles of individuals with chronic atrophic gastritis and GC.

RESULTS

A total of 17 DEGs comprising 8 upregulated and 9 down-regulated genes were identified from the microarray dataset from biopsies with chronic atrophic gastritis and GC. These DEGs were primarily enriched for CDK regulation of DNA replication and mitotic M-M/G1 phase pathways, according to KEGG analysis (p > 0.05). We discovered two hub genes, MCM7 and CDC6, in the protein-protein interaction network we obtained for the 17 DEGs (expanded with increased maximum interaction with 110 nodes and 2103 edges). MCM7 was discovered to be up-regulated in GC tissues following confirmation using the GEPIA and Human Protein Atlas databases.

CONCLUSION

The elevated expression of MCM7 in both chronic atrophic gastritis and GC, as shown by our comprehensive investigation, suggests that this protein may serve as a promising biomarker for the early detection of GC.

EGRET: edge aggregated graph attention networks and transfer learning improve protein-protein interaction site prediction

MOTIVATION

Protein-protein interactions (PPIs) are central to most biological processes. However, reliable identification of PPI sites using conventional experimental methods is slow and expensive. Therefore, great efforts are being put into computational methods to identify PPI sites.

RESULTS

We present Edge Aggregated GRaph Attention NETwork (EGRET), a highly accurate deep learning-based method for PPI site prediction, where we have used an edge aggregated graph attention network to effectively leverage the structural information. We, for the first time, have used transfer learning in PPI site prediction. Our proposed edge aggregated network, together with transfer learning, has achieved notable improvement over the best alternate methods. Furthermore, we systematically investigated EGRET's network behavior to provide insights about the causes of its decisions.

AVAILABILITY

EGRET is freely available as an open source project at https://github.com/Sazan-Mahbub/EGRET.

CONTACT

shams_bayzid@cse.buet.ac.bd.

Comprehensive analysis of the LncRNAs, MiRNAs, and MRNAs acting within the competing endogenous RNA network of LGG

Messenger RNA (mRNA) and long noncoding RNA (lncRNA) targets interact via competitive microRNA (miRNA) binding. However, the roles of cancer-specific lncRNAs in the competing endogenous RNA (ceRNA) networks of low-grade glioma (LGG) remain unclear. This study obtained RNA sequencing data for normal solid tissue and LGG primary tumour tissue from The Cancer Genome Atlas database. We used a computational method to analyse the relationships among the mRNAs, lncRNAs, and miRNAs in these samples. Gene ontology (GO) function and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was used to predict the biological processes (BPs) and pathways associated with these genes. Kaplan-Meier survival analysis was used to evaluate the association between the expression levels of specific mRNAs, lncRNAs, and miRNAs and overall survival. Finally, we created a ceRNA network describing the relationships among these mRNAs, lncRNAs, and miRNAs using Cytoscape 3.5.1. A total of 2555 differentially expressed (DE) mRNAs, 218 DElncRNAs, and 192 DEmiRNAs were identified using R. In addition, GO and KEGG pathway analysis of the mRNAs and lncRNAs in the ceRNA network identified 10 BPs, 10 cell components, 10 molecular functions, and 48 KEGG pathways as selectively enriched. A total of 55 lncRNAs, 50 miRNAs, and 10 mRNAs from this network were shown to be closely associated with overall survival in LGG. Finally, 59 miRNAs, 235 mRNAs, and 17 lncRNAs were used to develop a ceRNA network comprising 313 nodes and 1046 edges. This study helps expand our understanding of ceRNA networks and serves to clarify the underlying pathogenesis mechanism of LGG.

Smell Detection Agent Optimisation Framework and Systems Biology Approach to Detect Dys-Regulated Subnetwork in Cancer Data

Network biology has become a key tool in unravelling the mechanisms of complex diseases. Detecting dys-regulated subnetworks from molecular networks is a task that needs efficient computational methods. In this work, we constructed an integrated network using gene interaction data as well as protein–protein interaction data of differentially expressed genes derived from the microarray gene expression data. We considered the level of differential expression as well as the topological weight of proteins in interaction network to quantify dys-regulation. Then, a nature-inspired Smell Detection Agent (SDA) optimisation algorithm is designed with multiple agents traversing through various paths in the network. Finally, the algorithm provides a maximum weighted module as the optimum dys-regulated subnetwork. The analysis is performed for samples of triple-negative breast cancer as well as colorectal cancer. Biological significance analysis of module genes is also done to validate the results. The breast cancer subnetwork is found to contain (i) valid biomarkers including PIK3CA, PTEN, BRCA1, AR and EGFR; (ii) validated drug targets TOP2A, CDK4, HDAC1, IL6, BRCA1, HSP90AA1 and AR; (iii) synergistic drug targets EGFR and BIRC5. Moreover, based on the weight values assigned to nodes in the subnetwork, PLK1, CTNNB1, IGF1, AURKA, PCNA, HSPA4 and GAPDH are proposed as drug targets for further studies. For colorectal cancer module, the analysis revealed the occurrence of approved drug targets TYMS, TOP1, BRAF and EGFR. Considering the higher weight values, HSP90AA1, CCNB1, AKT1 and CXCL8 are proposed as drug targets for experimentation. The derived subnetworks possess cancer-related pathways as well. The SDA-derived breast cancer subnetwork is compared with that of tools such as MCODE and Minimum Spanning Tree, and observed a higher enrichment (75%) of significant elements. Thus, the proposed nature-inspired algorithm is a novel approach to derive the optimum dys-regulated subnetwork from huge molecular network.

Genome-scale protein interaction network construction and topology analysis of functional hypothetical proteins in Helicobacter pylori divulges novel therapeutic targets

The emergence and spread of multi-drug resistance among Helicobacter pylori (H. pylori) strain raise more stakes for genetic research for discovering new drugs. The quantity of uncharacterized hypothetical proteins in the genome may provide an opportunity to explore their property and promulgation could act as a platform for designing the drugs, making them an intriguing genetic target. In this context, the present study aims to identify the key hypothetical proteins (HPs) and their biological regulatory processes in H. pylori. This investigation could provide a foundation to establish the molecular connectivity among the pathways using topological analysis of the protein interaction networks (PINs). The giant network derived from the extended network has 374 nodes connected via 925 edges. A total of 43 proteins with high betweenness centrality (BC), 54 proteins with a large degree, and 23 proteins with high BC and large degrees have been identified. HP 1479, HP 0056, HP 1481, HP 1021, HP 0043, HP 1019, gmd, flgA, HP 0472, HP 1486, HP 1478, and HP 1473 are categorized as hub nodes because they have a higher number of direct connections and are potentially more important in understanding HP's molecular interactions. The pathway enrichment analysis of the network clusters revealed significant involvement of HPs in pathways such as flagellar assembly, bacterial chemotaxis and lipopolysaccharide biosynthesis. This comprehensive computational study revealed HP's functional role and its druggability characteristics, which could be useful in the development of drugs to combat H. pylori infections.

Quantitative FRET (qFRET) Technology for the Determination of Protein-Protein Interaction Affinity in Solution

Protein-protein interactions play pivotal roles in life, and the protein interaction affinity confers specific protein interaction events in physiology or pathology. Förster resonance energy transfer (FRET) has been widely used in biological and biomedical research to detect molecular interactions in vitro and in vivo. The FRET assay provides very high sensitivity and efficiency. Several attempts have been made to develop the FRET assay into a quantitative measurement for protein-protein interaction affinity in the past. However, the progress has been slow due to complicated procedures or because of challenges in differentiating the FRET signal from other direct emission signals from donor and receptor. This review focuses on recent developments of the quantitative FRET analysis and its application in the determination of protein-protein interaction affinity (K_D), either through FRET acceptor emission or donor quenching methods. This paper mainly reviews novel theatrical developments and experimental procedures rather than specific experimental results. The FRET-based approach for protein interaction affinity determination provides several advantages, including high sensitivity, high accuracy, low cost, and high-throughput assay. The FRET-based methodology holds excellent potential for those difficult-to-be expressed proteins and for protein interactions in living cells.

Network topology analysis of essential genes interactome of Helicobacter pylori to explore novel therapeutic targets

The Helicobacter pylori chronic colonization produces a wide range of gastric diseases in the gastric mucosa by abetting inflammation. Amidst coevolution and reorganization of its metabolism with humans, it has become difficult still imperative to understand and prevent its growth. This study focus to explore functional insights into identification of hub proteins/genes by aggregating the behavior of genes connected in a protein-protein interaction (PPI) network. We have constructed a PPI network of 123 essential genes along with 1213 interactions in H. pylori 26695. The degree and other centrality measures analysis assist in identifying the important hub nodes, which are top-ranked proteins. A total of nine proteins (recA, guaA, dnaK, rpsB, rplQ, rpmA, rpmC, rpmF, and rpsE) were obtained with high degree (k), betweenness centrality (BC) value. Gene ontology analysis reveals 8, 5 and 3 GO terms correspond to biological processes, cellular components and molecular function respectively. Gene complexes of hypothetical proteins (HPs) were related to aminoacyl-tRNA biosynthesis, biosynthesis of secondary metabolites, bacterial secretion system and protein export. The MCODE analysis revealed that protein from module M1, M3 and M6 include the proteins which have highest degree and BC values. It is noteworthy to mention that the bifunctional GMP synthase/glutamine amidotransferase protein (guaA), molecular chaperon (dnaK), recombinase A (recA) constitute as hub proteins. As a result, these genes are considered as network hub nodes that might be used as therapeutic targets. Our analysis affords a detailed understanding of the molecular process and pathways regulated by the essential genes in H. pylori 26695.

SAAMBE-SEQ: a sequence-based method for predicting mutation effect on protein-protein binding affinity

MOTIVATION

Vast majority of human genetic disorders are associated with mutations that affect protein-protein interactions by altering wild-type binding affinity. Therefore, it is extremely important to assess the effect of mutations on protein-protein binding free energy to assist the development of therapeutic solutions. Currently, the most popular approaches use structural information to deliver the predictions, which precludes them to be applicable on genome-scale investigations. Indeed, with the progress of genomic sequencing, researchers are frequently dealing with assessing effect of mutations for which there is no structure available.

RESULTS

Here, we report a Gradient Boosting Decision Tree machine learning algorithm, the SAAMBE-SEQ, which is completely sequence-based and does not require structural information at all. SAAMBE-SEQ utilizes 80 features representing evolutionary information, sequence-based features and change of physical properties upon mutation at the mutation site. The approach is shown to achieve Pearson correlation coefficient (PCC) of 0.83 in 5-fold cross validation in a benchmarking test against experimentally determined binding free energy change (ΔΔG). Further, a blind test (no-STRUC) is compiled collecting experimental ΔΔG upon mutation for protein complexes for which structure is not available and used to benchmark SAAMBE-SEQ resulting in PCC in the range of 0.37-0.46. The accuracy of SAAMBE-SEQ method is found to be either better or comparable to most advanced structure-based methods. SAAMBE-SEQ is very fast, available as webserver and stand-alone code, and indeed utilizes only sequence information, and thus it is applicable for genome-scale investigations to study the effect of mutations on protein-protein interactions.

AVAILABILITY AND IMPLEMENTATION

SAAMBE-SEQ is available at http://compbio.clemson.edu/saambe_webserver/indexSEQ.php#started.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Prediction of protein-binding residues: dichotomy of sequence-based methods developed using structured complexes versus disordered proteins

MOTIVATION

There are over 30 sequence-based predictors of the protein-binding residues (PBRs). They use either structure-annotated or disorder-annotated training datasets, potentially creating a dichotomy where the structure-/disorder-specific models may not be able to cross-over to accurately predict the other type. Moreover, the structure-trained predictors were shown to substantially cross-predict PBRs among residues that interact with non-protein partners (nucleic acids and small ligands). We address these issues by performing first-of-its-kind comparative study of a representative collection of disorder- and structure-trained predictors using a comprehensive benchmark set with the structure- and disorder-derived annotations of PBRs (to analyze the cross-over) and the protein-, nucleic acid- and small ligand-binding proteins (to study the cross-predictions).

RESULTS

Three predictors provide accurate results: SCRIBER, ANCHOR and disoRDPbind. Some of the structure-trained methods make accurate predictions on the structure-annotated proteins. Similarly, the disorder-trained predictors predict well on the disorder-annotated proteins. However, the considered predictors generally fail to cross-over, with the exception of SCRIBER. Our study also reveals that virtually all methods substantially cross-predict PBRs, except for SCRIBER for the structure-annotated proteins and disoRDPbind for the disorder-annotated proteins. We formulate a novel hybrid predictor, hybridPBRpred, that combines results produced by disoRDPbind and SCRIBER to accurately predict disorder- and structure-annotated PBRs. HybridPBRpred generates accurate results that cross-over structure- and disorder-annotated proteins and produces relatively low amount of cross-predictions, offering an accurate alternative to predict PBRs.

AVAILABILITY AND IMPLEMENTATION

HybridPBRpred webserver, benchmark dataset and supplementary information are available at http://biomine.cs.vcu.edu/servers/hybridPBRpred/.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Comprehensive Analysis of Survival-Related lncRNAs, miRNAs, and mRNAs Forming a Competing Endogenous RNA Network in Gastric Cancer

To analyze and construct a survival-related endogenous RNA (ceRNA) network in gastric cancer (GC) with lymph node metastasis, we obtained expression profiles of long non-coding RNAs (lncRNAs), mRNAs, and microRNAs (miRNAs) in GC from The Cancer Genome Atlas database. The edgeR package was used to screen differentially expressed lncRNAs, mRNAs, and miRNAs between GC patients with lymphatic metastasis and those without lymphatic metastasis. Then, we used univariate Cox regression analysis to identify survival-related differentially expressed RNAs. In addition, we used multivariate Cox regression analysis to screen lncRNAs, miRNAs, and mRNAs for use in the prognostic prediction models. The results showed that 2,247 lncRNAs, 155 miRNAs, and 1,253 mRNAs were differentially expressed between the two patient groups. Using univariate Cox regression analysis, we found that 395 lncRNAs, eight miRNAs, and 180 mRNAs were significantly related to the survival time of GC patients. We next created a survival-related network consisting of 59 lncRNAs, seven miRNAs, and 36 mRNAs. In addition, we identified eight RNAs associated with prognosis by multivariate Cox regression analysis, comprising three lncRNAs (AC094104.2, AC010457.1, and AC091832.1), two miRNAs (miR-653-5p and miR-3923), and three mRNAs (C5orf46, EPHA8, and HPR); these were used to construct the prognostic prediction models, and their risk scores could be used to assess GC patients' prognosis. In conclusion, this study provides new insights into ceRNA networks in GC and the screening of prognostic biomarkers for GC.

PROBselect: accurate prediction of protein-binding residues from proteins sequences via dynamic predictor selection

Motivation

Knowledge of protein-binding residues (PBRs) improves our understanding of protein−protein interactions, contributes to the prediction of protein functions and facilitates protein−protein docking calculations. While many sequence-based predictors of PBRs were published, they offer modest levels of predictive performance and most of them cross-predict residues that interact with other partners. One unexplored option to improve the predictive quality is to design consensus predictors that combine results produced by multiple methods.

Results

We empirically investigate predictive performance of a representative set of nine predictors of PBRs. We report substantial differences in predictive quality when these methods are used to predict individual proteins, which contrast with the dataset-level benchmarks that are currently used to assess and compare these methods. Our analysis provides new insights for the cross-prediction concern, dissects complementarity between predictors and demonstrates that predictive performance of the top methods depends on unique characteristics of the input protein sequence. Using these insights, we developed PROBselect, first-of-its-kind consensus predictor of PBRs. Our design is based on the dynamic predictor selection at the protein level, where the selection relies on regression-based models that accurately estimate predictive performance of selected predictors directly from the sequence. Empirical assessment using a low-similarity test dataset shows that PROBselect provides significantly improved predictive quality when compared with the current predictors and conventional consensuses that combine residue-level predictions. Moreover, PROBselect informs the users about the expected predictive quality for the prediction generated from a given input protein.

Availability and implementation

PROBselect is available at http://bioinformatics.csu.edu.cn/PROBselect/home/index.

Supplementary information

Supplementary data are available at Bioinformatics online.

Comprehensive analysis of prognostic biomarkers in lung adenocarcinoma based on aberrant lncRNA-miRNA-mRNA networks and Cox regression models

Lung adenocarcinoma (LUAD) is the leading cause of cancer-related death worldwide, and its underlying mechanism remains unclear. Accumulating evidence has highlighted that long non-coding RNA (lncRNA) acts as competitive endogenous RNA (ceRNA) and plays an important role in the occurrence and development of LUAD. Here, we comprehensively analyzed and provided an overview of the lncRNAs, miRNAs, and mRNAs associated with LUAD from The Cancer Genome Atlas (TCGA) database. Then, differentially expressed lncRNAs (DElncRNA), miRNAs (DEmiRNA), and mRNAs (DEmRNA) were used to construct a lncRNA–miRNA–mRNA regulatory network according to interaction information from miRcode, TargetScan, miRTarBase, and miRDB. Finally, the RNAs of the network were analyzed for survival and submitted for Cox regression analysis to construct prognostic indicators. A total of 1123 DElncRNAs, 95 DEmiRNAs, and 2296 DEmRNAs were identified (|log₂FoldChange| (FC) > 2 and false discovery rate (FDR) or adjusted P value < 0.01). The ceRNA network was established based on this and included 102 lncRNAs, 19 miRNAs, and 33 mRNAs. The DEmRNAs in the ceRNA network were found to be enriched in various cancer-related biological processes and pathways. We detected 22 lncRNAs, 12 mRNAs, and 1 miRNA in the ceRNA network that were significantly associated with the overall survival of patients with LUAD (P < 0.05). We established three prognostic prediction models and calculated the area under the 1,3,5-year curve (AUC) values of lncRNA, mRNA, and miRNA, respectively. Among them, the prognostic index (PI) of lncRNA showed good predictive ability which was 0.737, 0.702 and 0.671 respectively, and eight lncRNAs can be used as candidate prognostic biomarkers for LUAD. In conclusion, our study provides a new perspective on the prognosis and diagnosis of LUAD on a genome-wide basis, and develops independent prognostic biomarkers for LUAD.

SCRIBER: accurate and partner type-specific prediction of protein-binding residues from proteins sequences

Motivation

Accurate predictions of protein-binding residues (PBRs) enhances understanding of molecular-level rules governing protein–protein interactions, helps protein–protein docking and facilitates annotation of protein functions. Recent studies show that current sequence-based predictors of PBRs severely cross-predict residues that interact with other types of protein partners (e.g. RNA and DNA) as PBRs. Moreover, these methods are relatively slow, prohibiting genome-scale use.

Results

We propose a novel, accurate and fast sequence-based predictor of PBRs that minimizes the cross-predictions. Our SCRIBER (SeleCtive pRoteIn-Binding rEsidue pRedictor) method takes advantage of three innovations: comprehensive dataset that covers multiple types of binding residues, novel types of inputs that are relevant to the prediction of PBRs, and an architecture that is tailored to reduce the cross-predictions. The dataset includes complete protein chains and offers improved coverage of binding annotations that are transferred from multiple protein–protein complexes. We utilize innovative two-layer architecture where the first layer generates a prediction of protein-binding, RNA-binding, DNA-binding and small ligand-binding residues. The second layer re-predicts PBRs by reducing overlap between PBRs and the other types of binding residues produced in the first layer. Empirical tests on an independent test dataset reveal that SCRIBER significantly outperforms current predictors and that all three innovations contribute to its high predictive performance. SCRIBER reduces cross-predictions by between 41% and 69% and our conservative estimates show that it is at least 3 times faster. We provide putative PBRs produced by SCRIBER for the entire human proteome and use these results to hypothesize that about 14% of currently known human protein domains bind proteins.

Availability and implementation

SCRIBER webserver is available at http://biomine.cs.vcu.edu/servers/SCRIBER/.

Supplementary information

Supplementary data are available at Bioinformatics online.

Genome-wide analysis of prognostic-related lncRNAs, miRNAs and mRNAs forming a competing endogenous RNA network in lung squamous cell carcinoma

PURPOSE

As a type of cancer with the highest morbidity and mortality, lung squamous cell carcinoma (LUSC) has a very poor prognosis. Long-non-coding RNA (lncRNA) has recently attracted attentions because it can play the role of competing endogenous RNA (ceRNA) to inhibit microRNA (miRNA) functions. In this study, we aimed to find prognosis-related lncRNAs, miRNAs and mRNAs and construct a prognosis-related ceRNA network.

METHODS

The original LUSC RNA-sequencing data and miRNA profiles data were downloaded from the cancer genome atlas (TCGA) database. Differentially expressed lncRNAs, miRNAs and mRNAs were then identified between patients with lymph node metastasis and no lymph node metastasis. Univariate Cox regression analysis was performed to find the survival-associated lncRNAs, miRNAs and mRNAs. Subsequently, prognostic-related ceRNA network was established. By multivariate Cox regression analysis, three lncRNA signatures and three mRNA signatures were developed and used for predicting LUSC patients' survival.

RESULTS

A total of 224 lncRNAs, 160 miRNAs, 913 mRNAs were identified between samples with lymph node metastasis and no lymph node metastasis. Univariate Cox regression analysis showed that, among them, 28 lncRNAs, 8 miRNAs, 105 mRNAs were significantly associated with patients' overall survival time. Further pathway and enrichment analysis suggested that these mRNAs were associated with the regulation of transmembrane transport, regulation of blood circulation, plasma lipoprotein particle organization. Then we constructed a survival-related ceRNA network including 9 lncRNAs, 8 miRNAs and 23 mRNAs. Additionally, a multivariate Cox regression analysis demonstrated that three lncRNAs (AL161431.1, LINC02389, APCDD1L.DT) and three mRNAs (KLK6, SLITRK5, CCDC177) had a significant prognostic value. Risk score indicated that lncRNA signature and mRNA signature could independently predict overall survival in LUSC patients.

CONCLUSION

The current study provided a better understanding of the ceRNA network in the progression of LUSC and laid a theoretical foundation for LUSC prognosis.

Genome-wide screening diagnostic biomarkers and the construction of prognostic model of hepatocellular carcinoma

Although methods in diagnosis and therapy of hepatocellular carcinoma (HCC) have made significant progress in decades, the overall survival (OS) of HCC remains dissatisfactory, so it is particularly important to find better diagnostic and prognostic biomarkers. In this study, we found a more reliable potential diagnostic biomarkers and constructed a more accurate prognostic evaluation model based on integrated transcriptome sequencing analysis of multiple independent data sets. First, we performed quality evaluation and differential analysis on seven Gene Expression Omnibus (GEO) data sets, and then comprehensively analyzed the differentially expressed genes with a robust rank aggregation algorithm. Next, Least absolute shrinkage and selection operator (LASSO) regression was used to establish an 8-gene prognostic risk score (RS) model. Finally, the prognostic model was further validated in the GEO data set. Also, RS has independence on other clinicopathological characteristics but has similarities in prognostic assessment compared with the T stage. Moreover, the combination of T stage and prognostic RS model based on the 8-gene had a better prognostic evaluation effect. In brief, our research suggest that the prognostic risk model of 8 genes has important clinical significance in HCC patients, and can further enrich the prognostic guidance value of the traditional T stage.

Mapping of Protein-Protein Interactions: Web-Based Resources for Revealing Interactomes

BACKGROUND

The significant number of protein-protein interactions (PPIs) discovered by harnessing concomitant advances in the fields of sequencing, crystallography, spectrometry and two-hybrid screening suggests astonishing prospects for remodelling drug discovery. The PPI space which includes up to 650 000 entities is a remarkable reservoir of potential therapeutic targets for every human disease. In order to allow modern drug discovery programs to leverage this, we should be able to discern complete PPI maps associated with a specific disorder and corresponding normal physiology.

OBJECTIVE

Here, we will review community available computational programs for predicting PPIs and web-based resources for storing experimentally annotated interactions.

METHODS

We compared the capacities of prediction tools: iLoops, Struck2Net, HOMCOS, COTH, PrePPI, InterPreTS and PRISM to predict recently discovered protein interactions.

RESULTS

We described sequence-based and structure-based PPI prediction tools and addressed their peculiarities. Additionally, since the usefulness of prediction algorithms critically depends on the quality and quantity of the experimental data they are built on; we extensively discussed community resources for protein interactions. We focused on the active and recently updated primary and secondary PPI databases, repositories specialized to the subject or species, as well as databases that include both experimental and predicted PPIs.

CONCLUSION

PPI complexes are the basis of important physiological processes and therefore, possible targets for cell-penetrating ligands. Reliable computational PPI predictions can speed up new target discoveries through prioritization of therapeutically relevant protein-protein complexes for experimental studies.

Analysis of competing endogenous RNA network identifies a poorly differentiated cancer-specific RNA signature for hepatocellular carcinoma

Plenty of evidence has suggested that long noncoding RNAs (lncRNAs) play a vital role in competing endogenous RNA (ceRNA) networks. Poorly differentiated hepatocellular carcinoma (PDHCC) is a malignant phenotype. This paper aimed to explore the effect and the underlying regulatory mechanism of lncRNAs on PDHCC as a kind of ceRNA. Additionally, prognosis prediction was assessed. A total of 943 messenger RNAs (mRNAs), 86 miRNAs, and 468 lncRNAs that were differentially expressed between 137 PDHCCs and 235 well-differentiated HCCs were identified. Thereafter, a ceRNA network related to the dysregulated lncRNAs was established according to bioinformatic analysis and included 29 lncRNAs, 9 miRNAs, and 96 mRNAs. RNA-related overall survival (OS) curves were determined using the Kaplan-Meier method. The lncRNA ARHGEF7-AS2 was markedly correlated with OS in HCC (P = .041). Moreover, Cox regression analysis revealed that patients with low ARHGEF7-AS2 expression were associated with notably shorter survival time (P = .038). In addition, the area under the curve values of the lncRNA signature for 1-, 3-, and 5-year survival were 0.806, 0.741, and 0.701, respectively. Furthermore, a lncRNA nomogram was established, and the C-index of the internal validation was 0.717. In vitro experiments were performed to demonstrate that silencing ARHGEF7-AS2 expression significantly promoted HCC cell proliferation and migration. Taken together, our findings shed more light on the ceRNA network related to lncRNAs in PDHCC, and ARHGEF7-AS2 may be used as an independent biomarker to predict the prognosis of HCC.

Prognostic lncRNAs, miRNAs, and mRNAs Form a Competing Endogenous RNA Network in Colon Cancer

Purpose: To develop a multi-RNA-based model to provide survival risk prediction for colon cancer by constructing a competing endogenous RNAs (ceRNAs) network.

Methods: The prognostic information and expression of the lncRNAs, miRNAs, and mRNAs in colon cancer specimens from The Cancer Genome Atlas (TCGA) were assessed. Constructing prognostic models used the differentially expressed RNAs. Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses and Gene Ontology were used to identify the functional role of the ceRNA network in the prognosis of colon cancer.

Results: Five lncRNAs (AC007384.1, AC002511.1, AC012640.1, C17orf82, and AP001619.1), 8 miRNAs (hsa-mir-141, hsa-mir-150, hsa-mir-375, hsa-mir-96, hsa-mir-107, hsa-mir-106a, hsa-mir-200a, and hsa-mir-1271), and 5 mRNAs (BDNF, KLF4, SESN2, SMOC1, and TRIB3) were highly correlated with tumor status and tumor stage. Three prognostic models based on the 5 lncRNAs, 8 miRNAs, and 5 mRNAs were constructed. The prognostic ability was 0.850 for the lncRNA-based model, 0.811 for the miRNA-based model, and 0.770 for the mRNA-based model. Patients with high-risk scores revealed worse overall survival. The KEGG pathways were significantly enriched in the “neuroactive ligand-receptor interaction.”

Conclusion: This study identified several potential prognostic biomarkers to construct a multi-RNA-based prognostic model for colon cancer.

An integrative module analysis of DNA methylation landscape in aging

To investigate the molecular mechanism of aging, the combination of module analysis and DNA methylation data was used to detect dynamically controlled modules for aging. Multiple differential expression networks (DENs) were constructed based on the microarray profiles across different aging groups (<70 years, 70–80 years, and >80 years). Next, a module-based approach was utilized to extract the common candidate modules across all age groups. We used Module Connectivity Dynamic Score (MCDS) to quantify the connectivity change of the common modules among the different age groups. Functional analyses were implemented for the genes in the common modules to further identify the significant biological processes. A total of two DENs were constructed. Overall 657 informative genes were screened out. When false discovery rate (FDR) was set as 0.05, we found that 148 modules were significant. Only 1 significant 2-differential modules (DMs) (module 493) with dynamic changes was discovered. Significantly, the genes in the module 493 participated in 7 significant pathways, including pentose phosphate pathway, carbon metabolism, and citrate cycle (TCA cycle). In conclusion, pathway functions [pentose phosphate pathway, carbon metabolism, citrate cycle (TCA cycle), chromosomal instability, ateroid biosynthesis, PPAR signaling pathway, and immune response] may serve as potential therapeutic targets in aging.

Molecular and disease association of gestational diabetes mellitus affected mother and placental datasets reveal a strong link between insulin growth factor (IGF) genes in amino acid transport pathway: A network biology approach

Discerning the relationship between molecules involved in diseases based on their underlying biological mechanisms is one of the greatest challenges in therapeutic development today. Gestational diabetes mellitus (GDM) is one of the most common complications during pregnancy, which adversely affects both mothers and offspring during and after pregnancy. We have constructed two datasets of (GDM associated genes from affected mother and placenta to systematically analyze and evaluate their interactions like gene-gene, gene-protein, gene-microRNA (miRNA), gene-transcription factors, and gene-associated diseases to enhance our current knowledge, which may lead to further advancements in disease diagnosis, prognosis, and treatment. The results identify the key genes with respect to maternal dataset as insulin receptor, insulin (INS), leptin (LEP), glucokinase, and hepatocyte nuclear factor 1 alpha, whereas from placenta include insulin-like growth factor 1, growth hormone receptor, and breast cancer anti-estrogen resistance protein 1, which are found to be highly enriched in pancreas, ovary, adipocyte, heart, and placental tissues. The key transcription factors include Sp1 transcription factor, pancreatic and duodenal homeobox 1, and hepatocyte nuclear factor 4 alpha, whereas miRNA includes has-miR-5699-5p and has-miR-3158-3p. The study also reveals that GDM has associations with diseases like type I and II diabetes mellitus, obesity, and preeclampsia. More significantly, we could trace out a significant connection between the key molecules like LEP and placental growth hormone from mother and placental dataset, which plays a critical role in INS secretion, INS signaling, and β-cell dysfunction pathways.

Identification of differential modules in ankylosing spondylitis using systemic module inference and the attract method

The objective of the present study was to identify differential modules in ankylosing spondylitis (AS) by integrating network analysis, module inference and the attract method. To achieve this objective, four steps were conducted. The first step was disease objective network (DON) for AS, and healthy objective network (HON) inference dependent on gene expression data, protein-protein interaction networks and Spearman's correlation coefficient. In the second step, module detection was performed by utilizing a clique-merging algorithm, which comprised of exploring maximal cliques by clique algorithm and refining or merging maximal cliques with a high overlap. The third part was seed module evaluation through module pair matches by Jaccard score and module correlation density (MCD) calculation. Finally, in the fourth step, differential modules between the AS and healthy groups were identified based on a gene set enrichment analysis-analysis of variance model in the attract method. There were 5,301 nodes and 28,176 interactions both in DON and HON. A total of 20 and 21 modules were detected for the AS and healthy group, respectively. Notably, six seed modules across two groups were identified with Jaccard score ≥0.5, and these were ranked in descending order of differential MCD (ΔC). Seed module 1 had the highest ΔC of 0.077 and Jaccard score of 1.000. By accessing the attract method, one differential module between the AS group and healthy group was identified. In conclusion, the present study successfully identified one differential module for AS that may be a potential marker for AS target therapy and provide insights for future research on this disease.

Models and Data Sources Used in Systems Medicine

Background: Systems medicine is a new approach for the development and selection of treatment strategies for patients with complex diseases. It is often referred to as the application of systems biology methods for decision making in patient care. For systems medicine computer applications, many different data sources have to be integrated and included into models. This is a challenging task for Medical Informatics since the approach exceeds traditional systems like Electronic Health Records. To prioritize research activities for systems medicine applications, it is necessary to get an overview over modelling methods and data sources already used in this field.

Objectives: We performed a systematic literature review with the objective to capture current use of 1) modelling methods and 2) data sources in systems medicine related research projects.

Methods: We queried the MEDLINE and ScienceDirect databases for papers associated with the search term systems medicine and related terms. Papers were screened and assessed in full text in a two-step process according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement guidelines.

Results: The queries returned 698 articles of which 34 papers were finally included into the study. A multitude of modelling approaches such as machine learning and network analysis was identified and classified. Since these approaches are also used in other domains, no methods specific for systems medicine could be identified. Omics data are the most widely used data types followed by clinical data. Most studies only include a rather limited number of data sources.

Conclusions: Currently, many different modelling approaches are used in systems medicine. Thus, highly flexible modular solutions are necessary for systems medicine clinical applications. However, the number of data sources included into the models is limited and most projects currently focus on prognosis. To leverage the potential of systems medicine further, it will be necessary to focus on treatment strategies for patients and consider a broader range of data.

Screening feature modules and pathways in glioma using EgoNet

Background

To investigate differential egonetwork modules and pathways in glioma using EgoNet algorithm.

Methodology

Based on microarray data, EgoNet algorithm mainly comprised three stages: construction of differential co-expression network (DCN); EgoNet algorithm used to identify candidate ego-network modules based on the increased classification accuracy; statistical significance for candidate modules using random permutation testing. After that, pathway enrichment analysis for differential ego-network modules was implemented to illuminate the biological processes.

Results

We obtained 109 ego genes. From every ego gene, we progressively grew the ego-networks by levels; we extracted 109 ego-networks and the mean node size in an ego-network was 6. By setting the classification accuracy threshold at 0.90 and the count of nodes in an ego-network module at 10, we extracted 8 candidate ego-network modules. After random permutation test with 1000 times, 5 modules including module 59, 72, 78, 86, and 90 were identified to be significant. Of note, the genes of module 90 and 86 were enriched in the pathway of resolution of sister chromatid cohesion and mitotic prometaphase, respectively.

Conclusion

The identified modules and their corresponding ego genes might be beneficial in revealing the pathology underlying glioma and give insight for future research of glioma.

EgoNet identifies differential ego-modules and pathways related to prednisolone resistance in childhood acute lymphoblastic leukemia

PURPOSE

To extract feature ego-modules and pathways in childhood acute lymphoblastic leukemia (ALL) resistant to prednisolone treatment, and further to explore the mechanisms behind prednisolone resistance.

MATERIALS AND METHODS

EgoNet algorithm was used to identify candidate ego-network modules, mainly via constructing differential co-expression network (DCN); selecting ego genes; collecting ego-network modules; refining candidate modules. Afterwards, statistical significance was calculated for these candidate modules. Biological functions of differential ego-network modules were identified using Reactome database. To verify this proposed method can lead to truly positive findings in clinical settings, support vector machine (SVM) was utilized to compute the AUC values for each significant pathway using 3-fold cross-validation method. To predict the reliability of our findings, another established method (attract) was used to identify the differential attractor modules using the same microarray profile.

RESULTS

After eliminating the modules with classification accuracy < 0.9 and node number < 15, only ego-network module 30 was eligible. After significance calculation, module 30 was significant. Module 30 was enriched in APC/C-mediated degradation of cell cycle proteins. The AUC for the significant pathway of APC/C-mediated degradation of cell cycle proteins was 0.915. Although the attract method obtained more modules, the module identified by our proposed method owned more gene nodes, and had more classification ability (AUC = 0.915).

CONCLUSION

One differential ego-network module identified in childhood ALL resistance to prednisolone based on DCN and EgoNet, might be helpful to reveal the mechanisms underlying prednisolone resistance in childhood ALL.

Identifying dysregulated pathways in postmenopausal osteoporosis through investigation of crosstalk between pathways

The present study aimed to identify potential dysregulated pathways to further reveal the molecular mechanisms of postmenopausal osteoporosis (PMOP) based on pathway‑interaction network (PIN) analysis, which considers crosstalk between pathways. Protein‑protein interaction (PPI) data and pathway information were derived from STRING and Reactome Pathway databases, respectively. According to the gene expression profiles, pathway data and PPI information, a PIN was constructed with each node representing a biological pathway. Principal component analysis was used to compute the pathway activity for each pathway, and the seed pathway was selected. Subsequently, dysregulated pathways were extracted from the PIN based on the seed pathway and the increased classification accuracy, which was measured using the area under the curve (AUC) index according to 5‑fold cross validation. A PIN comprising 2,725 interactions was constructed, which was used to detect dysregulated pathways. Notably, the 'mitotic prometaphase' pathway was selected and defined as a seed pathway. Starting with the seed pathway, network‑based analysis successfully identified one pathway set for PMOP comprising eight dysregulated pathways (such as mitotic prometaphase, resolution of sister chromatid cohesion, mRNA splicing and mRNA splicing‑major) with an AUC score of 0.85, which may provide potential biomarkers for targeted therapy for PMOP.

A novel method to identify differential pathways in uterine leiomyomata based on network strategy

The aim of the present study was to identify differential pathways in uterine leiomyomata (UL) using a novel method based on protein-protein interaction networks and pathway analysis. The pathway networks were constructed by examining the intersections of the Reactome database and the Search Tool for the Retrieval of Interacting Genes/proteins (STRING) protein-protein interaction (PPI) networks. The Objective network was defined as the differential expressed genes (DEGs) associated with the interactions identified by STRING. Topological centrality (degree) analysis was performed for the Objective network to explore the hub genes and hub networks. Subsequent to isolating the intersections between the Pathway and Objective networks, randomization tests were conducted to identify the differential pathways. There were 559,598 interactions in the Pathway networks. A total of 657 genes with 3,835 interactions were mapped in the Objective network, which included 20 hub genes. It was identified that 358 pathways demonstrated interaction with the Objective network, such as Signal Transduction, Immune System and Signaling by G-protein-coupled receptor (GPCR). By accessing the randomization tests, P-values of these pathways were close to 0, which indicated that they were significantly different. The present study successfully identified differential pathways (such as signal transduction, immune system and signaling by GPCR) in UL, which may be potential biomarkers in the detection and treatment of UL.

An integrated lncRNA, microRNA and mRNA signature to improve prognosis prediction of colorectal cancer

Although the outcome of patients with colorectal cancer (CRC) has improved significantly, prognosis evaluation still presents challenges due to the disease heterogeneity. Increasing evidences revealed the close correlation between aberrant expression of certain RNAs and the prognosis. We envisioned that combined multiple types of RNAs into a single classifier could improve postoperative risk classification and add prognostic value to the current stage system. Firstly, differentially expressed RNAs including mRNAs, miRNAs and lncRNAs were identified by two different algorithms. Then survival and LASSO analysis was conducted to screen survival-related DERs and build a multi-RNA-based classifier for CRC patient stratification. The prognostic value of the classifier was self-validated in the TCGA CRC cohort and further validated in an external independent set. Finally, survival receiver operating characteristic analysis was used to assess the performance of prognostic prediction. We found that the multi-RNA-based classifier consisted by 12 mRNAs, 1miRNA and 1 lncRNA, which could divide the patients into high and low risk groups with significantly different overall survival (training set: HR 2.54, 95%CI 1.67-3.87, p<0.0001; internal testing set: HR 2.54, 95%CI 1.67-3.87, p<0.0001; validation set: HR 5.02, 95% CI 2.2–11.6; p=0·0002). In addition, the classifier is not only independent of clinical features but also with a similar prognostic ability to the well-established TNM stage (AUC of ROC 0.83 versus 0.74, 95% CI = 0.608-0.824, P =0.0878). Furthermore, combination of the multi-RNA-based classifier with clinical features was a more powerful predictor of prognosis than either of the two parameters alone. In conclusion, the multi-RNA-based classifier may have important clinical implications in the selection of patients with CRC who are at high risk of mortality and add prognostic value to the current stage system.

A novel method to identify hub pathways of rheumatoid arthritis based on differential pathway networks

The aim of the current study was to identify hub pathways of rheumatoid arthritis (RA) using a novel method based on differential pathway network (DPN) analysis. The present study proposed a DPN where protein-protein interaction (PPI) network was integrated with pathway-pathway interactions. Pathway data was obtained from background PPI network and the Reactome pathway database. Subsequently, pathway interactions were extracted from the pathway data by building randomized gene-gene interactions and a weight value was assigned to each pathway interaction using Spearman correlation coefficient (SCC) to identify differential pathway interactions. Differential pathway interactions were visualized using Cytoscape to construct a DPN. Topological analysis was conducted to identify hub pathways that possessed the top 5% degree distribution of DPN. Modules of DPN were mined according to ClusterONE. A total of 855 pathways were selected to build pathway interactions. By filtrating pathway interactions of weight values >0.7, a DPN with 312 nodes and 791 edges was obtained. Topological degree analysis revealed 15 hub pathways, such as heparan sulfate/heparin-glycosaminoglycan (HS-GAG) degradation, HS-GAG metabolism and keratan sulfate degradation for RA based on DPN. Furthermore, hub pathways were also important in modules, which validated the significance of hub pathways. In conclusion, the proposed method is a computationally efficient way to identify hub pathways of RA, which identified 15 hub pathways that may be potential biomarkers and provide insight to future investigation and treatment of RA.

Revealing radiotherapy- and chemoradiation-induced pathway dynamics in glioblastoma by analyzing multiple differential networks

The progression of glioblastoma (GBM) is driven by dynamic alterations in the activity and connectivity of gene pathways. Revealing these dynamic events is necessary in order to understand the pathological mechanisms of, and develop effective treatments for, GBM. The present study aimed to investigate dynamic alterations in pathway activity and connectivity across radiotherapy and chemoradiation conditions in GBM, and to give system-level insights into molecular mechanisms for GBM therapy. A total of two differential co-expression networks (DCNs) were constructed using Pearson correlation coefficient analysis and one sided t-tests, based on gene expression profiles and protein-protein interaction networks, one for each condition. Subsequently, shared differential modules across DCNs were detected via significance analysis for candidate modules, which were obtained according to seed selection, module search by seed expansion and refinement of searched modules. As condition-specific differential modules mediate differential biological processes, the module connectivity dynamic score (MCDS) was implemented to explore dynamic alterations among them. Based on DCNs with 287 nodes and 1,052 edges, a total of 28 seed genes and seven candidate modules were identified. Following significance analysis, five shared differential modules were identified in total. Dynamic alterations among these differential modules were identified using the MCDS, and one module with significant dynamic alterations was identified, termed the dynamic module. The present study revealed the dynamic alterations of shared differential modules, identified one dynamic module between the radiotherapy and chemoradiation conditions, and demonstrated that pathway dynamics may applied to the study of the pathogenesis and therapy of GBM.

Benchmarking pathway interaction network for colorectal cancer to identify dysregulated pathways

Different pathways act synergistically to participate in many biological processes. Thus, the purpose of our study was to extract dysregulated pathways to investigate the pathogenesis of colorectal cancer (CRC) based on the functional dependency among pathways. Protein-protein interaction (PPI) information and pathway data were retrieved from STRING and Reactome databases, respectively. After genes were aligned to the pathways, each pathway activity was calculated using the principal component analysis (PCA) method, and the seed pathway was discovered. Subsequently, we constructed the pathway interaction network (PIN), where each node represented a biological pathway based on gene expression profile, PPI data, as well as pathways. Dysregulated pathways were then selected from the PIN according to classification performance and seed pathway. A PIN including 11,960 interactions was constructed to identify dysregulated pathways. Interestingly, the interaction of mRNA splicing and mRNA splicing-major pathway had the highest score of 719.8167. Maximum change of the activity score between CRC and normal samples appeared in the pathway of DNA replication, which was selected as the seed pathway. Starting with this seed pathway, a pathway set containing 30 dysregulated pathways was obtained with an area under the curve score of 0.8598. The pathway of mRNA splicing, mRNA splicing-major pathway, and RNA polymerase I had the maximum genes of 107. Moreover, we found that these 30 pathways had crosstalks with each other. The results suggest that these dysregulated pathways might be used as biomarkers to diagnose CRC.

Revealing differential modules in uveal melanoma by analyzing differential networks

The aim of the present study was to investigate differential modules (DMs) between uveal melanoma (UM) and normal conditions by examining differential networks. Based on a gene expression profile collected from the ArrayExpress database, the inference of DMs involved three steps: The first step was construction of a differential co‑expression network (DCN); second, the module algorithm was adapted to identify the DMs presented in DCN; finally, the statistical significance of DMs were assessed based on the null score distribution of DMs generated using randomized networks. A DCN with 309 nodes and 3,729 edges was obtained, and 30 seed genes from the DCN were examined. Subsequently, one DM, which had 179 nodes and 3,068 edges, was investigated. By utilizing randomized networks, the P‑value for DM was 0.034, therefore, the DM was statically significant between UM and baseline conditions. In conclusion, the present study successfully identified one DM in UM based on DCN and module algorithm, and this DM may be beneficial in revealing the pathological mechanism of UM and provide insight for future investigation of UM.

Investigating ego modules and pathways in osteosarcoma by integrating the EgoNet algorithm and pathway analysis

Osteosarcoma (OS) is the most common primary bone malignancy, but current therapies are far from effective for all patients. A better understanding of the pathological mechanism of OS may help to achieve new treatments for this tumor. Hence, the objective of this study was to investigate ego modules and pathways in OS utilizing EgoNet algorithm and pathway-related analysis, and reveal pathological mechanisms underlying OS. The EgoNet algorithm comprises four steps: constructing background protein-protein interaction (PPI) network (PPIN) based on gene expression data and PPI data; extracting differential expression network (DEN) from the background PPIN; identifying ego genes according to topological features of genes in reweighted DEN; and collecting ego modules using module search by ego gene expansion. Consequently, we obtained 5 ego modules (Modules 2, 3, 4, 5, and 6) in total. After applying the permutation test, all presented statistical significance between OS and normal controls. Finally, pathway enrichment analysis combined with Reactome pathway database was performed to investigate pathways, and Fisher's exact test was conducted to capture ego pathways for OS. The ego pathway for Module 2 was CLEC7A/inflammasome pathway, while for Module 3 a tetrasaccharide linker sequence was required for glycosaminoglycan (GAG) synthesis, and for Module 6 was the Rho GTPase cycle. Interestingly, genes in Modules 4 and 5 were enriched in the same pathway, the 2-LTR circle formation. In conclusion, the ego modules and pathways might be potential biomarkers for OS therapeutic index, and give great insight of the molecular mechanism underlying this tumor.

Microarray and network-based identification of functional modules and pathways of active tuberculosis

Diagnose of active tuberculosis (TB) is challenging and treatment response is also difficult to efficiently monitor. The aim of this study was to use an integrated analysis of microarray and network-based method to the samples from publically available datasets to obtain a diagnostic module set and pathways in active TB. Towards this goal, background protein-protein interactions (PPI) network was generated based on global PPI information and gene expression data, following by identification of differential expression network (DEN) from the background PPI network. Then, ego genes were extracted according to the degree features in DEN. Next, module collection was conducted by ego gene expansion based on EgoNet algorithm. After that, differential expression of modules between active TB and controls was evaluated using random permutation test. Finally, biological significance of differential modules was detected by pathways enrichment analysis based on Reactome database, and Fisher's exact test was implemented to extract differential pathways for active TB. Totally, 47 ego genes and 47 candidate modules were identified from the DEN. By setting the cutoff-criteria of gene size >5 and classification accuracy ≥0.9, 7 ego modules (Module 4, Module 7, Module 9, Module 19, Module 25, Module 38 and Module 43) were extracted, and all of them had the statistical significance between active TB and controls. Then, Fisher's exact test was conducted to capture differential pathways for active TB. Interestingly, genes in Module 4, Module 25, Module 38, and Module 43 were enriched in the same pathway, formation of a pool of free 40S subunits. Significant pathway for Module 7 and Module 9 was eukaryotic translation termination, and for Module 19 was nonsense mediated decay enhanced by the exon junction complex (EJC). Accordingly, differential modules and pathways might be potential biomarkers for treating active TB, and provide valuable clues for better understanding of molecular mechanism of active TB.

Strategic Integration of Multiple Bioinformatics Resources for System Level Analysis of Biological Networks

Recent technological advances in genomics allow the production of biological data at unprecedented tera- and petabyte scales. Efficient mining of these vast and complex datasets for the needs of biomedical research critically depends on a seamless integration of the clinical, genomic, and experimental information with prior knowledge about genotype-phenotype relationships. Such experimental data accumulated in publicly available databases should be accessible to a variety of algorithms and analytical pipelines that drive computational analysis and data mining.We present an integrated computational platform Lynx (Sulakhe et al., Nucleic Acids Res 44:D882-D887, 2016) ( http://lynx.cri.uchicago.edu ), a web-based database and knowledge extraction engine. It provides advanced search capabilities and a variety of algorithms for enrichment analysis and network-based gene prioritization. It gives public access to the Lynx integrated knowledge base (LynxKB) and its analytical tools via user-friendly web services and interfaces. The Lynx service-oriented architecture supports annotation and analysis of high-throughput experimental data. Lynx tools assist the user in extracting meaningful knowledge from LynxKB and experimental data, and in the generation of weighted hypotheses regarding the genes and molecular mechanisms contributing to human phenotypes or conditions of interest. The goal of this integrated platform is to support the end-to-end analytical needs of various translational projects.

Investigating ego modules involved in TGFβ3-induced chondrogenesis in mesenchymal stem cells based on ego network

OBJECTIVE

This paper aimed to investigate ego modules for TGFβ3-induced chondrogenesis in mesenchymal stem cells (MSCs) using ego network algorithm.

METHODS

The ego network algorithm comprised three parts, extracting differential expression network (DEN) based on gene expression data and protein-protein interaction (PPI) data; exploring ego genes by reweighting DEN; and searching ego modules by ego gene expansions. Subsequently, permutation test was carried out to evaluate the statistical significance of the ego modules. Finally, pathway enrichment analysis was conducted to investigate ego pathways enriched by the ego modules.

RESULTS

A total of 15 ego genes were obtained from the DEN, such as PSMA4, HNRNPM and WDR77. Starting with each ego genes, 15 candidate modules were gained. When setting the thresholds of the area under the receiver operating characteristics curve (AUC) ≥0.9 and gene size ≥4, three ego modules (Module 3, Module 8 and Module 14) were identified, and all of them had statistical significances between normal and TGFβ3-induced chondrogenesis in MSCs. By mapping module genes to confirmed pathway database, their ego pathways were detected, Cdc20:Phospho-APC/C mediated degradation of Cyclin A for Module 3, Mitotic G1-G1/S phases for Module 8, and mRNA Splicing for Module 14.

CONCLUSIONS

We have successfully identified three ego modules, evaluated their statistical significances and investigated their functional enriched ego pathways. The findings might provide potential biomarkers and give great insights to reveal molecular mechanism underlying this process.

Anesthetic Propofol-Induced Gene Expression Changes in Patients Undergoing Coronary Artery Bypass Graft Surgery Based on Dynamical Differential Coexpression Network Analysis

We aimed to determine the influence of anesthetic propofol on gene expression in patients treated by coronary artery bypass graft (CABG) surgery based on differential coexpression network (DCN) and to further reveal the novel mechanisms of the cardioprotective effects of propofol. Firstly, we constructed the DCN for disease condition based on Pearson correlation coefficient (PCC) and weight value. Secondly, the inference of modules was applied to search modules from DCN with same members but varied connectivity. Furthermore, we measured the statistical significance of the modules for selecting differential modules (DMs). Finally, attract method was used for DMs analysis to select key modules. Based on the δ value, 11928 edges and 2956 nodes were chosen to construct DCNs. A total of 29 seed genes were selected. Moreover, by quantifying connectivity changes in shared gene modules across different conditions, 8 DMs with higher connectivity dynamics were identified. Then, we extracted key modules using attract method, there were 8 key modules, and the top 3 modules were module 1, 2, and 3. Furthermore, GCG, PPY, and PON1 were initial seed genes of these 3 key modules, respectively. Accordingly, GCG and PON1 might exert important roles in the cardioprotective effects of propofol during CABG.

Alzheimer's as a Systems-Level Disease Involving the Interplay of Multiple Cellular Networks

Alzheimer's disease (AD), and many neurodegenerative disorders, are multifactorial in nature. They involve a combination of genomic, epigenomic, interactomic and environmental factors. Progress is being made, and these complex diseases are beginning to be understood as having their origin in altered states of biological networks at the cellular level. In the case of AD, genomic susceptibility and mechanisms leading to (or accompanying) the impairment of the central Amyloid Precursor Protein (APP) processing and tau networks are widely accepted as major contributors to the diseased state. The derangement of these networks may result in both the gain and loss of functions, increased generation of toxic species (e.g., toxic soluble oligomers and aggregates) and imbalances, whose effects can propagate to supra-cellular levels. Although well sustained by empirical data and widely accepted, this global perspective often overlooks the essential roles played by the main counteracting homeostatic networks (e.g., protein quality control/proteostasis, unfolded protein response, protein folding chaperone networks, disaggregases, ER-associated degradation/ubiquitin proteasome system, endolysosomal network, autophagy, and other stress-protective and clearance networks), whose relevance to AD is just beginning to be fully realized. In this chapter, an integrative perspective is presented. Alzheimer's disease is characterized to be a result of: (a) intrinsic genomic/epigenomic susceptibility and, (b) a continued dynamic interplay between the deranged networks and the central homeostatic networks of nerve cells. This interplay of networks will underlie both the onset and rate of progression of the disease in each individual. Integrative Systems Biology approaches are required to effect its elucidation. Comprehensive Systems Biology experiments at different 'omics levels in simple model organisms, engineered to recapitulate the basic features of AD may illuminate the onset and sequence of events underlying AD. Indeed, studies of models of AD in simple organisms, differentiated cells in culture and rodents are beginning to offer hope that the onset and progression of AD, if detected at an early stage, may be stopped, delayed, or even reversed, by activating or modulating networks involved in proteostasis and the clearance of toxic species. In practice, the incorporation of next-generation neuroimaging, high-throughput and computational approaches are opening the way towards early diagnosis well before irreversible cell death. Thus, the presence or co-occurrence of: (a) accumulation of toxic Aβ oligomers and tau species; (b) altered splicing and transcriptome patterns; (c) impaired redox, proteostatic, and metabolic networks together with, (d) compromised homeostatic capacities may constitute relevant 'AD hallmarks at the cellular level' towards reliable and early diagnosis. From here, preventive lifestyle changes and tailored therapies may be investigated, such as combined strategies aimed at both lowering the production of toxic species and potentiating homeostatic responses, in order to prevent or delay the onset, and arrest, alleviate, or even reverse the progression of the disease.

Discovering distinct functional modules of specific cancer types using protein-protein interaction networks

Background. The molecular profiles exhibited in different cancer types are very different; hence, discovering distinct functional modules associated with specific cancer types is very important to understand the distinct functions associated with them. Protein-protein interaction networks carry vital information about molecular interactions in cellular systems, and identification of functional modules (subgraphs) in these networks is one of the most important applications of biological network analysis. Results. In this study, we developed a new graph theory based method to identify distinct functional modules from nine different cancer protein-protein interaction networks. The method is composed of three major steps: (i) extracting modules from protein-protein interaction networks using network clustering algorithms; (ii) identifying distinct subgraphs from the derived modules; and (iii) identifying distinct subgraph patterns from distinct subgraphs. The subgraph patterns were evaluated using experimentally determined cancer-specific protein-protein interaction data from the Ingenuity knowledgebase, to identify distinct functional modules that are specific to each cancer type. Conclusion. We identified cancer-type specific subgraph patterns that may represent the functional modules involved in the molecular pathogenesis of different cancer types. Our method can serve as an effective tool to discover cancer-type specific functional modules from large protein-protein interaction networks.

Multi-Facial, Non-Peptidic α-Helix Mimetics

α-Helices often recognize their target proteins at protein–protein interfaces through more than one recognition face. This review describes the state-of-the-art in the design of non-peptidic α-helix mimetics that reproduce functionality from multiple faces of an α-helix.

Towards more drug-like proteomimetics: two-faced, synthetic α-helix mimetics based on a purine scaffold

Mimicry of two faces of an α-helix might yield more potent and more selective inhibitors of aberrant, helix-mediated protein-protein interactions (PPI). Herein, we demonstrate that a 2,6,9-tri-substituted purine is capable of disrupting the Mcl-1-Bak-BH3 PPI through effective mimicry of key residues on opposing faces of the Bak-BH3 α-helix.

Pathway mapping and development of disease-specific biomarkers: protein-based network biomarkers

It is known that a disease is rarely a consequence of an abnormality of a single gene, but reflects the interactions of various processes in a complex network. Annotated molecular networks offer new opportunities to understand diseases within a systems biology framework and provide an excellent substrate for network-based identification of biomarkers. The network biomarkers and dynamic network biomarkers (DNBs) represent new types of biomarkers with protein-protein or gene-gene interactions that can be monitored and evaluated at different stages and time-points during development of disease. Clinical bioinformatics as a new way to combine clinical measurements and signs with human tissue-generated bioinformatics is crucial to translate biomarkers into clinical application, validate the disease specificity, and understand the role of biomarkers in clinical settings. In this article, the recent advances and developments on network biomarkers and DNBs are comprehensively reviewed. How network biomarkers help a better understanding of molecular mechanism of diseases, the advantages and constraints of network biomarkers for clinical application, clinical bioinformatics as a bridge to the development of diseases-specific, stage-specific, severity-specific and therapy predictive biomarkers, and the potentials of network biomarkers are also discussed.

Identification of hub subnetwork based on topological features of genes in breast cancer

The aim of this study was to provide functional insight into the identification of hub subnetworks by aggregating the behavior of genes connected in a protein-protein interaction (PPI) network. We applied a protein network-based approach to identify subnetworks which may provide new insight into the functions of pathways involved in breast cancer rather than individual genes. Five groups of breast cancer data were downloaded and analyzed from the Gene Expression Omnibus (GEO) database of high-throughput gene expression data to identify gene signatures using the genome-wide global significance (GWGS) method. A PPI network was constructed using Cytoscape and clusters that focused on highly connected nodes were obtained using the molecular complex detection (MCODE) clustering algorithm. Pathway analysis was performed to assess the functional relevance of selected gene signatures based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. Topological centrality was used to characterize the biological importance of gene signatures, pathways and clusters. The results revealed that, cluster1, as well as the cell cycle and oocyte meiosis pathways were significant subnetworks in the analysis of degree and other centralities, in which hub nodes mostly distributed. The most important hub nodes, with top ranked centrality, were also similar with the common genes from the above three subnetwork intersections, which was viewed as a hub subnetwork with more reproducible than individual critical genes selected without network information. This hub subnetwork attributed to the same biological process which was essential in the function of cell growth and death. This increased the accuracy of identifying gene interactions that took place within the same functional process and was potentially useful for the development of biomarkers and networks for breast cancer.

Efficient prediction of human protein-protein interactions at a global scale

BACKGROUND

Our knowledge of global protein-protein interaction (PPI) networks in complex organisms such as humans is hindered by technical limitations of current methods.

RESULTS

On the basis of short co-occurring polypeptide regions, we developed a tool called MP-PIPE capable of predicting a global human PPI network within 3 months. With a recall of 23% at a precision of 82.1%, we predicted 172,132 putative PPIs. We demonstrate the usefulness of these predictions through a range of experiments.

CONCLUSIONS

The speed and accuracy associated with MP-PIPE can make this a potential tool to study individual human PPI networks (from genomic sequences alone) for personalized medicine.

Recapitulating the α-helix: nonpeptidic, low-molecular-weight ligands for the modulation of helix-mediated protein–protein interactions

Protein–protein interactions play critical roles in a wide variety of biological processes, and their dysregulations contribute to the pathogenesis of several diseases, including cancer. Chemical entities that can abrogate aberrant protein–protein interactions may provide novel therapeutic agents. A large number of protein–protein interactions are mediated by protein secondary structure, the most commonly encountered form of which is the α-helix. Accordingly, over the last decade, there has been a flood of nonpeptidic small molecules that recapitulate the projection and chemical nature of key side chains of the canonical α-helix as a strategy to disrupt helix-mediated protein–protein interactions. In this review, we discuss recent advances (post 2006) in the design of synthetic α-helix mimetics, which include single-faced and two-faced/amphipathic structures, for the modulation of protein–protein interactions.