Integrating Gene Co-expression Research Articles

Harnessing the predicted maize pan-interactome for putative gene function prediction and prioritization of candidate genes for important traits.

The recent assembly and annotation of the 26 maize nested association mapping population founder inbreds have enabled large-scale pan-genomic comparative studies. These studies have expanded our understanding of agronomically important traits by integrating pan-transcriptomic data with trait-specific gene candidates from previous association mapping results. In contrast to the availability of pan-transcriptomic data, obtaining reliable protein-protein interaction (PPI) data has remained a challenge due to its high cost and complexity. We generated predicted PPI networks for each of the 26 genomes using the established STRING database. The individual genome-interactomes were then integrated to generate core- and pan-interactomes. We deployed the PPI clustering algorithm ClusterONE to identify numerous PPI clusters that were functionally annotated using gene ontology (GO) functional enrichment, demonstrating a diverse range of enriched GO terms across different clusters. Additional cluster annotations were generated by integrating gene coexpression data and gene description annotations, providing additional useful information. We show that the functionally annotated PPI clusters establish a useful framework for protein function prediction and prioritization of candidate genes of interest. Our study not only provides a comprehensive resource of predicted PPI networks for 26 maize genomes but also offers annotated interactome clusters for predicting protein functions and prioritizing gene candidates. The source code for the Python implementation of the analysis workflow and a standalone web application for accessing the analysis results are available at https://github.com/eporetsky/PanPPI.

Identification of diagnostic model in heart failure with myocardial fibrosis and conduction block by integrated gene co-expression network analysis

BackgroundDespite the advancements in heart failure(HF) research, the early diagnosis of HF continues to be a challenging issue in clinical practice. This study aims to investigate the genes related to myocardial fibrosis and conduction block, with the goal of developing a diagnostic model for early treatment of HF in patients.MethodThe gene expression profiles of GSE57345, GSE16499, and GSE9128 were obtained from the Gene Expression Omnibus (GEO) database. After merging the expression profile data and adjusting for batch effects, differentially expressed genes (DEGs) associated with conduction block and myocardial fibrosis were identified. Gene Ontology (GO) resources, Kyoto Encyclopedia of Genes and Genomes (KEGG) resources, and gene set enrichment analysis (GSEA) were utilized for functional enrichment analysis. A protein-protein interaction network (PPI) was constructed using a string database. Potential key genes were selected based on the bioinformatics information mentioned above. SVM and LASSO were employed to identify hub genes and construct the module associated with HF. The mRNA levels of TAC mice and external datasets (GSE141910 and GSE59867) are utilized for validating the diagnostic model. Additionally, the study explores the relationship between the diagnostic model and immune cell infiltration.ResultsA total of 395 genes exhibiting differential expression were identified. Functional enrichment analysis revealed that these specific genes primarily participate in biological processes and pathways associated with the constituents of the extracellular matrix (ECM), immune system processes, and inflammatory responses. We identified a diagnostic model consisting of 16 hub genes, and its predictive performance was validated using external data sets and a transverse aortic coarctation (TAC) mouse model. In addition, we observed significant differences in mRNA expression of 7 genes in the TAC mouse model. Interestingly, our study also unveiled a correlation between these model genes and immune cell infiltration.ConclusionsWe identified sixteen key genes associated with myocardial fibrosis and conduction block, as well as diagnostic models for heart failure. Our findings have significant implications for the intensive management of individuals with potential genetic variants associated with heart failure, especially in the context of advancing cell-targeted therapy for myocardial fibrosis.

Integrated gene co-expression network analysis and experimental validation revealed potential targets of human urine extract CDA-II in treating chronic myeloid leukemia

BackgroundCell differentiation agent II (CDA-II) exhibits potent anti-proliferative and apoptosis-inducing properties against a variety of cancer cells. However, its mechanism of action in chronic myeloid leukemia (CML) remains unclear. MethodsCell counting Kit 8 (CCK-8) and flow cytometry were used to investigate the effects of CDA-II on the biological characteristics of K562 cells. Gene (mRNA and lncRNA) expression profiles were analyzed by bioinformatics to screen differentially expressed genes and to perform enrichment analysis. The Pearson correlation coefficients of lncRNAs and mRNAs were calculated using gene expression values, and a lncRNA/mRNA co-expression network was constructed. The MCODE and cytoHubba plugins were used to analyze the co-expression network. ResultsThe Results, derived from CCK-8 and flow cytometry, indicated that CDA-II exerts dual effects on K562 cells: it inhibits their proliferation and induces apoptosis. From bioinformatics analysis, we identified 316 mRNAs and 32 lncRNAs. These mRNAs were predominantly related to the meiotic cell cycle, DNA methylation, transporter complex and peptidase regulator activity, complement and coagulation cascades, protein digestion and absorption, and cell adhesion molecule signaling pathways. The co-expression network comprised of 163 lncRNA/mRNA interaction pairs. Notably, our analysis results implicated clustered histone gene families and five lncRNAs in the biological effects of CDA-II on K562 cells. ConclusionThis study highlights the hub gene and lncRNA/mRNA co-expression network as crucial elements in the context of CDA-II treatment of CML. This insight not only enriches our understanding of CDA-II's mechanism of action but also might provide valuable clues for subsequent experimental studies of CDA-II, and potentially contribute to the discovery of new therapeutic targets for CML.

GeCoNet-Tool: a software package for gene co-expression network construction and analysis

BackgroundNetwork analysis is a powerful tool for studying gene regulation and identifying biological processes associated with gene function. However, constructing gene co-expression networks can be a challenging task, particularly when dealing with a large number of missing values.ResultsWe introduce GeCoNet-Tool, an integrated gene co-expression network construction and analysis tool. The tool comprises two main parts: network construction and network analysis. In the network construction part, GeCoNet-Tool offers users various options for processing gene co-expression data derived from diverse technologies. The output of the tool is an edge list with the option of weights associated with each link. In network analysis part, the user can produce a table that includes several network properties such as communities, cores, and centrality measures. With GeCoNet-Tool, users can explore and gain insights into the complex interactions between genes.

Investigating Key Targets of Dajianzhong Decoction for Treating Crohn’s Disease Using Weighted Gene Co-Expression Network

Background: Crohn’s disease (CD) is an inflammatory bowel disease, cases of which have substantially increased in recent years. The classical formula Dajianzhong decoction (DD, Japanese: Daikenchuto) is often used to treat CD, but few studies have evaluated related therapeutic mechanisms. In this study, we investigated the potential targets and mechanisms of DD used for treating CD at the molecular level through the weighted gene co-expression network. Methods: The main chemical components of the three DD herbs (Zanthoxylum bungeanum Maxim., Zingiber officinale (Willd.) Rosc., and Ginseng Radix et Rhizoma) were searched for using the HERB database. The targets for each component were identified using the SwissTargetPrediction and HERB databases, whereas the disease targets for CD were retrieved from the GeneCards and DisGeNET databases. The functional enrichment analysis was performed on the common targets of DD and CD. High-throughput sequencing data for CD patients were retrieved from the Gene Expression Omnibus database, and WGCNA was performed to identify the key targets. The association between the key targets and DD ingredients was verified using molecular docking. Results: By analyzing the interaction targets between DD and CD, 196 overlapping genes were identified. The enrichment results indicated that the PI3K-AKT, TNF, MAPK, and IL-17 signaling pathways influenced the mechanism of action of DD in counteracting CD. Combined with WGCNA, four differentially expressed genes (SLC6A4, NOS2, SHBG, and ABCB1) and their corresponding 24 compounds were closely related to the occurrence of CD. Conclusions: By integrating gene co-expression network analysis, this study preliminarily reveals the internal molecular mechanism of DD in treating CD from a systematic perspective, validated by molecular docking. However, these findings require further validation.

Detection of new candidate genes controlling seed weight by integrating gene coexpression analysis and QTL mapping in Brassica napus L.

Seed weight is a component of seed yield in rapeseed (Brassica napus L.). Although quantitative trait loci (QTL) for seed weight have been reported in rapeseed, only a few causal quantitative trait genes (QTGs) have been identified, resulting in a limitation in understanding of seed weight regulation. We constructed a gene coexpression network at the early seed developmental stage using transcripts of 20,408 genes in QTL intervals and 1017 rapeseed homologs of known genes from other species. Among the 10 modules in this gene coexpression network, modules 1 and 2 were core modules and contained genes involved in source–flow–sink processes such as synthesis and transportation of fatty acid and protein, and photosynthesis. A hub gene SERINE CARBOXYPEPTIDASE-LIKE 19 (SCPL19) was identified by candidate gene association analysis in rapeseed and functionally investigated using Arabidopsis T-DNA mutant and overexpression lines. Our study demonstrates the power of gene coexpression analysis to prioritize candidate genes from large candidate QTG sets and enhances the understanding of molecular mechanism for seed weight at the early developmental stage in rapeseed.

Integrated gene co-expression network analysis reveals unique developmental processes of Aurelia aurita

The typical life cycle of the moon jellyfish (Aurelia aurita) includes the planula, polyp, strobila, ephyra, and medusa developmental stages. These stages exhibit huge differences in both external morphology and internal physiological functions. However, the gene co-expression network involved in these post-embryonic developmental processes has not been studied yet. Here, based on 15 RNA sequencing samples covering all five stages of the A. aurita life cycle, we systematically analyzed the gene co-expression network and obtained 35 relevant modules. Furthermore, we identified the highly correlated modules and hub genes for each stage. These hub genes are implicated to play important roles in the developmental processes of A. aurita, which should help improve our understanding of the jellyfish life cycle.

Identifying significant genes and functionally enriched pathways in familial hypercholesterolemia using integrated gene co-expression network analysis

Familial hypercholesterolemia (FH) is a monogenic lipid disorder which promotes atherosclerosis and cardiovascular diseases. Owing to the lack of sufficient published information, this study aims to identify the potential genetic biomarkers for FH by studying the global gene expression profile of blood cells. The microarray expression data of FH patients and controls was analyzed by different computational biology methods like differential expression analysis, protein network mapping, hub gene identification, functional enrichment of biological pathways, and immune cell restriction analysis. Our results showed the dysregulated expression of 115 genes connected to lipid homeostasis, immune responses, cell adhesion molecules, canonical Wnt signaling, mucin type O-glycan biosynthesis pathways in FH patients. The findings from expanded protein interaction network construction with known FH genes and subsequent Gene Ontology (GO) annotations have also supported the above findings, in addition to identifying the involvement of dysregulated thyroid hormone and ErbB signaling pathways in FH patients. The genes like CSNK1A1, JAK3, PLCG2, RALA, and ZEB2 were found to be enriched under all GO annotation categories. The subsequent phenotype ontology results have revealed JAK3I, PLCG2, and ZEB2 as key hub genes contributing to the inflammation underlying cardiovascular and immune response related phenotypes. Immune cell restriction findings show that above three genes are highly expressed by T-follicular helper CD4+ T cells, naïve B cells, and monocytes, respectively. These findings not only provide a theoretical basis to understand the role of immune dysregulations underlying the atherosclerosis among FH patients but may also pave the way to develop genomic medicine for cardiovascular diseases.

Integrated Gene Co-expression Analysis and Metabolites Profiling Highlight the Important Role of ZmHIR3 in Maize Resistance to Gibberella Stalk Rot.

Gibberella stalk rot (GSR) caused by Fusarium graminearum is one of the most devastating diseases causing significant yield loss of maize, and GSR resistance is a quantitative trait controlled by multiple genes. Although a few quantitative trait loci/resistance genes have been identified, the molecular mechanisms underlying GSR resistance remain largely unexplored. To identify potential resistance genes and to better understand the molecular mechanism of GSR resistance, a joint analysis using a comparative transcriptomic and metabolomic approaches was conducted using two inbred lines with contrasting GSR resistance, K09 (resistant) and A08 (susceptible), upon infection with F. graminearum. While a substantial number of differentially expressed genes associated with various defense-related signaling pathways were identified between two lines, multiple hub genes likely associated with GSR resistance were pinpointed using Weighted Gene Correlation Network Analysis and K-means clustering. Moreover, a core set of metabolites, including anthocyanins, associated with the hub genes was determined. Among the complex co-expression networks, ZmHIR3 showed strong correlation with multiple key genes, and genetic and histological studies showed that zmhir3 mutant is more susceptible to GSR, accompanied by enhanced cell death in the stem in response to infection with F. graminearum. Taken together, our study identified differentially expressed key genes and metabolites, as well as co-expression networks associated with distinct infection stages of F. graminearum. Moreover, ZmHIR3 likely plays a positive role in disease resistance to GSR, probably through the transcriptional regulation of key genes, functional metabolites, and the control of cell death.

Single-cell RNA-seq dissects the intratumoral heterogeneity of triple-negative breast cancer based on gene regulatory networks

Triple-negative breast cancer (TNBC) is a subtype of breast cancer with high intratumoral heterogeneity. Recent studies revealed that TNBC patients might comprise cells with distinct molecular subtypes. In addition, gene regulatory networks (GRNs) constructed based on single-cell RNA sequencing (scRNA-seq) data have demonstrated the significance for decoding the key regulators. We performed a comprehensive analysis of the GRNs for the intrinsic subtypes of TNBC patients using scRNA-seq. The copy number variations (CNVs) were inferred from scRNA-seq data and identified 545 malignant cells. The subtypes of the malignant cells were assigned based on the PAM50 model. The cell-cell communication analysis revealed that the macrophage plays a dominant role in the tumor microenvironment. Next, the GRN for each subtype was constructed through integrating gene co-expression and enrichment of transcription-binding motifs. Then, we identified the critical genes based on the centrality metrics of genes. Importantly, the critical gene ETV6 was ubiquitously upregulated in all subtypes, but it exerted diverse roles in each subtype through regulating different target genes. In conclusion, the construction of GRNs based on scRNA-seq data could help us to dissect the intratumoral heterogeneity and identify the critical genes of TNBC.

Integrated time-serial transcriptome networks reveal common innate and tissue-specific adaptive immune responses to PRRSV infection

Porcine reproductive and respiratory syndrome virus (PRRSV) infection is the most important viral disease causing severe economic losses in the swine industry. However, mechanisms underlying gene expression control in immunity-responsible tissues at different time points during PRRSV infection are poorly understood. We constructed an integrated gene co-expression network and identified tissue- and time-dependent biological mechanisms of PRRSV infection through bioinformatics analysis using three tissues (lungs, bronchial lymph nodes [BLNs], and tonsils) via RNA-Seq. Three groups with specific expression patterns (i.e., the 3-dpi, lung, and BLN groups) were discovered. The 3 dpi-specific group showed antiviral and innate-immune signalling similar to the case for influenza A infection. Moreover, we observed adaptive immune responses in the lung-specific group based on various cytokines, while the BLN-specific group showed down-regulated AMPK signalling related to viral replication. Our study may provide comprehensive insights into PRRSV infection, as well as useful information for vaccine development.

Genetic Dissection of Hypertrophic Cardiomyopathy with Myocardial RNA-Seq

Background: Hypertrophic cardiomyopathy (HCM) is an inherited disorder of the myocardium, and pathogenic mutations in the sarcomere genes MYH7 and MYBPC3 explain 60-70% of observed clinical cases. The heterogeneity of phenotypes observed in HCM patients, however, suggests that novel causative genes or genetic modifiers likely exist and that further elucidation of these genes is needed. Methods: In order to identify genes and potential networks that affect the HCM phenotype, we systemically evaluated RNA-seq data from 28 HCM patients and 9 healthy controls with pathogenic variant identification, differential expression analysis, and gene co-expression and protein-protein interaction network analyses. Findings: We identified 43 potential pathogenic variants in 19 genes in 24 HCM patients. Genes with more than one variant included the following: MYBPC3, TTN, MYH7, PSEN2, and LDB3. A total of 2538 protein-coding genes, 6 microRNAs (miRNAs), and 1617 long noncoding RNAs (lncRNAs) were identified differentially expressed between the groups, including several well-characterized cardiomyopathy-related genes (ANKRD1, FHL2, MYH6, TGFB3, miR-30d, and miR-154). Gene enrichment analysis revealed that those genes are significantly involved in heart development and physiology. Furthermore, by integrating gene co-expression and protein-protein interaction network analyses, we highlighted four subnetworks: Mitochondrial DNA (mtDNA)-subnetwork, Desmoplakin (DSP) -subnetwork, myosin heavy chain 7 (MYH7) -subnetwork, and cardiac myosin-binding protein C (MYBPC3)-subnetwork, which could play significant roles in the progression of HCM. Interpretation: Our findings further illustrate that HCM is a complex disease, which results from mutations in multiple protein-coding genes, modulation by non-coding RNAs and perturbations in gene networks. Funding Statement: The authors received no specific funding for this work. Declaration of Interests: All other authors declare no competing interests. Ethics Approval Statement: Not required.

Detecting Diagnostic Biomarkers of Alzheimer's Disease by Integrating Gene Expression Data in Six Brain Regions.

Alzheimer's disease (AD) is a neurodegenerative and progressive disease, which often causes irreversible damages to the cerebrum. The pathogenesis of AD is far from being fully understood, while there are some popular hypotheses. So far, the diagnosis of AD relies only on clinical screening in the form of imaging techniques or cerebrospinal fluid analysis, which may lead to inaccurate evaluation and then cause the delay of suitable treatments. While molecular biomarkers provide promising alternatives of establishing correct relationships between genotypes and phenotypes of clinical symptoms. In this paper, we propose a machine-learning-based method of identifying potential diagnostic biomarkers of AD based on gene coexpression network by integrating gene expression profiles in six brain regions. After building an integrated gene coexpression network of multiple brain regions, we decompose the differential network into some subnetwork modules. The module candidates from these coexpressed gene communities are then identified by screening their discriminative powers in control from disease samples. The potential biomarkers are then validated by multiple cross-validations and functional enrichment analyses. If the biomarkers successfully pass clinical significance tests, they can be used as a reference for clinical diagnosis after wet-experimental validations.

A novel integrated gene coexpression analysis approach reveals a prognostic three-transcription-factor signature for glioma molecular subtypes.

BackgroundGlioma is the most common brain tumor and it has very high mortality rate due to its infiltration and heterogeneity. Precise classification of glioma subtype is essential for proper therapeutic treatment and better clinical prognosis. However, the molecular mechanism of glioma is far from clear and the classical classification methods based on traditional morphologic and histopathologic knowledge are subjective and inconsistent. Recently, classification methods based on molecular characteristics are developed with rapid progress of high throughput technology.MethodsIn the present study, we designed a novel integrated gene coexpression analysis approach, which involves differential coexpression and differential regulation analysis (DCEA and DRA), to investigate glioma prognostic biomarkers and molecular subtypes based on six glioma transcriptome data sets.ResultsWe revealed a novel three-transcription-factor signature including AHR, NFIL3 and ZNF423 for glioma molecular subtypes. This three-TF signature clusters glioma patients into three major subtypes (ZG, NG and IG subtypes) which are significantly different in patient survival as well as transcriptomic patterns. Notably, ZG subtype is featured with higher expression of ZNF423 and has better prognosis with younger age at diagnosis. NG subtype is associated with higher expression of NFIL3 and AHR, and has worse prognosis with elder age at diagnosis. According to our inferred differential networking information and previously reported signalling knowledge, we suggested testable hypotheses on the roles of AHR and NFIL3 in glioma carcinogenesis.ConclusionsWith so far the least biomarkers, our approach not only provides a novel glioma prognostic molecular classification scheme, but also helps to explore its dysregulation mechanisms. Our work is extendable to prognosis-related classification and signature identification in other cancer researches.Electronic supplementary materialThe online version of this article (doi:10.1186/s12918-016-0315-y) contains supplementary material, which is available to authorized users.

Abstract 1937: In-depth genomic analysis reveals complex functional consequence of driver mutations in breast cancers

Abstract Understanding the extensive functional consequences of molecular drivers of a cancer is challenging. We developed an in-depth genomic analytical framework to identify the gene and pathway targets dysregulated by somatic driver mutations. Using The Cancer Genome Atlas (TCGA) breast cancer data, we systemically analyzed all high frequent mutation events including amplification, deletion and point mutation across tumor tissues. Those include the known molecular driver genes such as PIK3CA, GATA3, PTEN and TP53. Utilizing bagging with distance-based regression for differential gene expression analysis, we identified several thousands gene targets associated with the somatic mutations. Integrating gene co-expression regulatory network and the association dependence structure among the somatic mutations, we explored a complex association network of molecular pathway modules or sub-modules that are specific to a mutational event or common to a set of events. Interestingly, we found those modules commonly associated with sets of mutation events were significantly enriched in well-known breast cancer related pathways such as ERBB2, PI3AK and TGFB1. Combined with the publicly available data of a genome-wide RNA-mediated interference screen in breast cancer, we further demonstrated that those modules enriched with cell proliferation essential genes. Our analysis strategy provides a novel view of complex functional mechanism driven by the driver mutation events and it has high potential to provide novel insights to the understanding of the tumorigenesis or the development of therapeutic targets. Citation Format: Yaomin Xu, Xingyi Guo, Shilin Zhao, Eric Torstenson, Todd Edwards, Yu Shyr. In-depth genomic analysis reveals complex functional consequence of driver mutations in breast cancers. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 1937. doi:10.1158/1538-7445.AM2015-1937

GroupRank: rank candidate genes in PPI network by differentially expressed gene groups.

Many cell activities are organized as a network, and genes are clustered into co-expressed groups if they have the same or closely related biological function or they are co-regulated. In this study, based on an assumption that a strong candidate disease gene is more likely close to gene groups in which all members coordinately differentially express than individual genes with differential expression, we developed a novel disease gene prioritization method GroupRank by integrating gene co-expression and differential expression information generated from microarray data as well as PPI network. A candidate gene is ranked high using GroupRank if it is differentially expressed in disease and control or is close to differentially co-expressed groups in PPI network. We tested our method on data sets of lung, kidney, leukemia and breast cancer. The results revealed GroupRank could efficiently prioritize disease genes with significantly improved AUC value in comparison to the previous method with no consideration of co-exprssed gene groups in PPI network. Moreover, the functional analyses of the major contributing gene group in gene prioritization of kidney cancer verified that our algorithm GroupRank not only ranks disease genes efficiently but also could help us identify and understand possible mechanisms in important physiological and pathological processes of disease.

Analysis of regulatory networks constructed based on gene coexpression in pituitary adenoma

Gene coexpression patterns can reveal gene collections with functional consistency. This study systematically constructs regulatory networks for pituitary tumours by integrating gene coexpression, transcriptional and posttranscriptional regulation. Through network analysis, we elaborate the incidence mechanism of pituitary adenoma. The Pearson's correlation coefficient was utilized to calculate the level of gene coexpression. By comparing pituitary adenoma samples with normal samples, pituitary adenoma-specific gene coexpression patterns were identified. For pituitary adenoma-specific coexpressed genes, we integrated transcription factor (TF) and microRNA (miRNA) regulation to construct a complex regulatory network from the transcriptional and posttranscriptional perspectives. Network module analysis identified the synergistic regulation of genes by miRNAs and TFs in pituitary adenoma. We identified 142 pituitary adenoma-specific active genes, including 43 TFs and 99 target genes of TFs. Functional enrichment of these 142 genes revealed that the occurrence of pituitary adenoma induced abnormalities in intracellular metabolism and angiogenesis process. These 142 genes were also significantly enriched in adenoma pathway. Module analysis of the systematic regulatory network found that three modules contained elements that were closely related to pituitary adenoma, such as FGF2 and SP1, as well as transcription factors and miRNAs involved in the tumourigenesis. These results show that in the occurrence of pituitary adenoma, miRNA, TF and genes interact with each other. Based on gene expression, the proposed method integrates interaction information from different levels and systematically explains the occurrence of pituitary tumours. It facilitates the tracing of the origin of the disease and can provide basis for early diagnosis of complex diseases or cancer without obvious symptoms.

Network analysis for gene discovery in plant‐specialized metabolism

Recent omics technologies provide information on multiple components of biological networks. Web-based data mining tools are continuously being developed. Because genes involved in specialized (secondary) metabolism are often co-ordinately regulated at the transcriptional level, a number of gene discovery studies have been successfully conducted using network analysis, especially by integrating gene co-expression network analysis and metabolomic investigation. In addition, next-generation sequencing technologies are currently utilized in functional genomics investigations of Arabidopsis and non-model plant species including medicinal plants. Systems-based approaches are expected to gain importance in medicinal plant research. This review discussed network analysis in Arabidopsis and gene discovery in plant-specialized metabolism in non-model plants.

Integrated gene co-expression network analysis in the growth phase of Mycobacterium tuberculosis reveals new potential drug targets

We have carried out weighted gene co-expression network analysis of Mycobacterium tuberculosis to gain insights into gene expression architecture during log phase growth. The differentially expressed genes between at least one pair of 11 different M. tuberculosis strains as source of biological variability were used for co-expression network analysis. This data included genes with highest coefficient of variation in expression. Five distinct modules were identified using topological overlap based clustering. All the modules together showed significant enrichment in biological processes: fatty acid biosynthesis, cell membrane, intracellular membrane bound organelle, DNA replication, Quinone biosynthesis, cell shape and peptidoglycan biosynthesis, ribosome and structural constituents of ribosome and transposition. We then extracted the co-expressed connections which were supported either by transcriptional regulatory network or STRING database or high edge weight of topological overlap. The genes trpC, nadC, pitA, Rv3404c, atpA, pknA, Rv0996, purB, Rv2106 and Rv0796 emerged as top hub genes. After overlaying this network on the iNJ661 metabolic network, the reactions catalyzed by 15 highly connected metabolic genes were knocked down in silico and evaluated by Flux Balance Analysis. The results showed that in 12 out of 15 cases, in 11 more than 50% of reactions catalyzed by genes connected through co-expressed connections also had altered fluxes. The modules 'Turquoise', 'Blue' and 'Red' also showed enrichment in essential genes. We could map 152 of the previously known or proposed drug targets in these modules and identified 15 new potential drug targets based on their high degree of co-expressed connections and strong correlation with module eigengenes.

Co-expression network with protein–protein interaction and transcription regulation in malaria parasite Plasmodium falciparum

BackgroundMalaria continues to be one of the most severe global infectious diseases, as a major threat to human health and economic development. Network-based biological analysis is a promising approach to uncover key genes and biological processes from a network viewpoint, which could not be recognized from individual gene-based signatures. ResultsWe integrated gene co-expression profile with protein–protein interaction and transcriptional regulation information to construct a comprehensive gene co-expression network of Plasmodium falciparum. Based on this network, we identified 10 core modules by using ICE (Iterative Clique Enumeration) algorithm, which were essential for malaria parasite development in intraerythrocytic developmental cycle (IDC) stages. In each module, all genes were highly correlated probably due to co-regulation or formation of a protein complex. Some of these genes were recognized to be differentially coexpressed among three close-by IDC stages. The gene of prpf8 (PFD0265w) encoding pre-mRNA processing splicing factor 8 product was identified as DCGs (differentially co-expressed genes) among IDC stages, although this gene function was seldom reported in previous researches. Integrating the species-specific gene prediction and differential co-expression gene detection, we found some modules could perform species-specific functions according to some of genes in these modules were species-specific genes, like the module 10.Furthermore, in order to reveal the underlying mechanisms of the erythrocyte invasion by P. falciparum, Steiner Tree algorithm was employed to identify the invasion subnetwork from our gene co-expression network. The subnetwork-based analysis indicated that some important Plasmodium parasite specific genes could corporate with each other and be co-regulated during the parasite invasion process, which including a head-to-head gene pair of PfRH2a (PF13_0198) and PfRH2b (MAL13P1.176). ConclusionsThis study based on gene co-expression network could shed new insights on the mechanisms of pathogenesis, even virulence and P. falciparum development.