Gene co-expression network analysis identifies trait-related modules in Arabidopsis thaliana

A method for mining condition-specific co-expressed genes in Camellia sinensis based on k-means clustering

BACKGROUND

As one of the world's most important beverage crops, tea plants (Camellia sinensis) are renowned for their unique flavors and numerous beneficial secondary metabolites, attracting researchers to investigate the formation of tea quality. With the increasing availability of transcriptome data on tea plants in public databases, conducting large-scale co-expression analyses has become feasible to meet the demand for functional characterization of tea plant genes. However, as the multidimensional noise increases, larger-scale co-expression analyses are not always effective. Analyzing a subset of samples generated by effectively downsampling and reorganizing the global sample set often leads to more accurate results in co-expression analysis. Meanwhile, global-based co-expression analyses are more likely to overlook condition-specific gene interactions, which may be more important and worthy of exploration and research.

RESULTS

Here, we employed the k-means clustering method to organize and classify the global samples of tea plants, resulting in clustered samples. Metadata annotations were then performed on these clustered samples to determine the "conditions" represented by each cluster. Subsequently, we conducted gene co-expression network analysis (WGCNA) separately on the global samples and the clustered samples, resulting in global modules and cluster-specific modules. Comparative analyses of global modules and cluster-specific modules have demonstrated that cluster-specific modules exhibit higher accuracy in co-expression analysis. To measure the degree of condition specificity of genes within condition-specific clusters, we introduced the correlation difference value (CDV). By incorporating the CDV into co-expression analyses, we can assess the condition specificity of genes. This approach proved instrumental in identifying a series of high CDV transcription factor encoding genes upregulated during sustained cold treatment in Camellia sinensis leaves and buds, and pinpointing a pair of genes that participate in the antioxidant defense system of tea plants under sustained cold stress.

CONCLUSIONS

To summarize, downsampling and reorganizing the sample set improved the accuracy of co-expression analysis. Cluster-specific modules were more accurate in capturing condition-specific gene interactions. The introduction of CDV allowed for the assessment of condition specificity in gene co-expression analyses. Using this approach, we identified a series of high CDV transcription factor encoding genes related to sustained cold stress in Camellia sinensis. This study highlights the importance of considering condition specificity in co-expression analysis and provides insights into the regulation of the cold stress in Camellia sinensis.

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.

Comparative transcriptome profiling and co-expression network analysis uncover the key genes associated with pear petal defense responses against Monilinia laxa infection

Pear brown rot and blossom blight caused by Monilinia laxa seriously affect pear production worldwide. Here, we compared the transcriptomic profiles of petals after inoculation with M. laxa using two pear cultivars with different levels of sensitivity to disease (Sissy, a relatively tolerant cultivar, and Kristalli, a highly susceptible cultivar). Physiological indexes were also monitored in the petals of both cultivars at 2 h and 48 h after infection (2 HAI and 48 HAI). RNA-seq data and weighted gene co-expression network analysis (WGCNA) allowed the identification of key genes and pathways involved in immune- and defense-related responses that were specific for each cultivar in a time-dependent manner. In particular, in the Kristalli cultivar, a significant transcriptome reprogramming occurred early at 2 HAI and was accompanied either by suppression of key differentially expressed genes (DEGs) involved in the modulation of any defense responses or by activation of DEGs acting as sensitivity factors promoting susceptibility. In contrast to the considerably high number of DEGs induced early in the Kristalli cultivar, upregulation of specific DEGs involved in pathogen perception and signal transduction, biosynthesis of secondary and primary metabolism, and other defense-related responses was delayed in the Sissy cultivar, occurring at 48 HAI. The WGCNA highlighted one module that was significantly and highly correlated to the relatively tolerant cultivar. Six hub genes were identified within this module, including three WRKY transcription factor-encoding genes: WRKY 65 (pycom05g27470), WRKY 71 (pycom10g22220), and WRKY28 (pycom17g13130), which may play a crucial role in enhancing the tolerance of pear petals to M. laxa. Our results will provide insights into the interplay of the molecular mechanisms underlying immune responses of petals at the pear-M. laxa pathosystem.

Exploring large-scale gene coexpression networks in peach (Prunus persica L.): a new tool for predicting gene function

Peach is a model for Prunus genetics and genomics, however, identifying and validating genes associated to peach breeding traits is a complex task. A gene coexpression network (GCN) capable of capturing stable gene-gene relationships would help researchers overcome the intrinsic limitations of peach genetics and genomics approaches and outline future research opportunities. In this study, we created four GCNs from 604 Illumina RNA-Seq libraries. We evaluated the performance of every GCN in predicting functional annotations using an algorithm based on the 'guilty-by-association' principle. The GCN with the best performance was COO300, encompassing 21 956 genes. To validate its performance predicting gene function, we performed two case studies. In case study 1, we used two genes involved in fruit flesh softening: the endopolygalacturonases PpPG21 and PpPG22. Genes coexpressing with both genes were extracted and referred to as melting flesh (MF) network. Finally, we performed an enrichment analysis of MF network and compared the results with the current knowledge regarding peach fruit softening. The MF network mostly included genes involved in cell wall expansion and remodeling, and with expressions triggered by ripening-related phytohormones, such as ethylene, auxin, and methyl jasmonate. In case study 2, we explored potential targets of the anthocyanin regulator PpMYB10.1 by comparing its gene-centered coexpression network with that of its grapevine orthologues, identifying a common regulatory network. These results validated COO300 as a powerful tool for peach and Prunus research. This network, renamed as PeachGCN v1.0, and the scripts required to perform a function prediction analysis are available at https://github.com/felipecobos/PeachGCN.

Trehalose metabolism coordinates transcriptional regulatory control and metabolic requirements to trigger the onset of cassava storage root initiation

Cassava storage roots (SR) are an important source of food energy and raw material for a wide range of applications. Understanding SR initiation and the associated regulation is critical to boosting tuber yield in cassava. Decades of transcriptome studies have identified key regulators relevant to SR formation, transcriptional regulation and sugar metabolism. However, there remain uncertainties over the roles of the regulators in modulating the onset of SR development owing to the limitation of the widely applied differential gene expression analysis. Here, we aimed to investigate the regulation underlying the transition from fibrous (FR) to SR based on Dynamic Network Biomarker (DNB) analysis. Gene expression analysis during cassava root initiation showed the transition period to SR happened in FR during 8 weeks after planting (FR8). Ninety-nine DNB genes associated with SR initiation and development were identified. Interestingly, the role of trehalose metabolism, especially trehalase1 (TRE1), in modulating metabolites abundance and coordinating regulatory signaling and carbon substrate availability via the connection of transcriptional regulation and sugar metabolism was highlighted. The results agree with the associated DNB characters of TRE1 reported in other transcriptome studies of cassava SR initiation and Attre1 loss of function in literature. The findings help fill the knowledge gap regarding the regulation underlying cassava SR initiation.

Transcriptomic analysis reveals the gene regulatory networks involved in leaf and root response to osmotic stress in tomato

Introduction

Tomato (Solanum lycopersicum L.) is a major horticultural crop that is cultivated worldwide and is characteristic of the Mediterranean agricultural system. It represents a key component of the diet of billion people and an important source of vitamins and carotenoids. Tomato cultivation in open field often experiences drought episodes, leading to severe yield losses, since most modern cultivars are sensitive to water deficit. Water stress leads to changes in the expression of stress-responsive genes in different plant tissues, and transcriptomics can support the identification of genes and pathways regulating this response.

Methods

Here, we performed a transcriptomic analysis of two tomato genotypes, M82 and Tondo, in response to a PEG-mediated osmotic treatment. The analysis was conducted separately on leaves and roots to characterize the specific response of these two organs.

Results

A total of 6,267 differentially expressed transcripts related to stress response was detected. The construction of gene co-expression networks defined the molecular pathways of the common and specific responses of leaf and root. The common response was characterized by ABA-dependent and ABA-independent signaling pathways, and by the interconnection between ABA and JA signaling. The root-specific response concerned genes involved in cell wall metabolism and remodeling, whereas the leaf-specific response was principally related to leaf senescence and ethylene signaling. The transcription factors representing the hubs of these regulatory networks were identified. Some of them have not yet been characterized and can represent novel candidates for tolerance.

Discussion

This work shed new light on the regulatory networks occurring in tomato leaf and root under osmotic stress and set the base for an in-depth characterization of novel stress-related genes that may represent potential candidates for improving tolerance to abiotic stress in tomato.

Transcriptome and Regional Association Analyses Reveal the Effects of Oleosin Genes on the Accumulation of Oil Content in Brassica napus

Rapeseed stores lipids in the form of oil bodies. Oil bodies in the seeds of higher plants are surrounded by oleosins. Adjusting oleosin protein levels can prevent the fusion of oil bodies and maintain oil body size during seed development. However, oil contents are affected by many factors, and studies on the complex molecular regulatory mechanisms underlying the variations in seed oil contents of B. napus are limited. In this study, a total of 53 BnOLEO (B. napus oleosin) genes were identified in the genome of B. napus through a genome-wide analysis. The promoter sequences of oleosin genes consisted of various light-, hormone-, and stress-related cis-acting elements, along with transcription factor (TF) binding sites, for 25 TF families in 53 BnOLEO genes. The differentially expressed oleosin genes between two high- and two low-oil-content accessions were explored. BnOLEO3-C09, BnOLEO4-A02, BnOLEO4-A09, BnOLEO2-C04, BnOLEO1-C01, and BnOLEO7-A03 showed higher expressions in the high-oil-content accessions than in low-oil-content accessions, at 25, 35, and 45 days after pollination (DAP) in two different environments. A regional association analysis of 50 re-sequenced rapeseed accessions was used to further analyze these six BnOLEO genes, and it revealed that the nucleotide variations in the BnOLEO1-C01 and BnOLEO7-A03 gene regions were related to the phenotypic variations in seed oil content. Moreover, a co-expression network analysis revealed that the BnOLEO genes were directly linked to lipid/fatty acid metabolism, TF, lipid transport, and carbohydrate genes, thus forming a molecular network involved in seed oil accumulation. These favorable haplotypes can be utilized in molecular marker-assisted selection in order to further improve seed oil contents in rapeseed.

Comprehensive functional analysis of the PYL-PP2C-SnRK2s family in Bletilla striata reveals that BsPP2C22 and BsPP2C38 interact with BsPYLs and BsSnRK2s in response to multiple abiotic stresses

As the core regulation network for the abscisic acid (ABA) signaling pathway, the PYL-PP2C-SnRK2s family commonly exists in many species. For this study, a total of 9 BsPYLs, 66 BsPP2Cs, and 7 BsSnRK2s genes were identified based on the genomic databases of Bletilla striata, which were classified into 3, 10, and 3 subgroups, respectively. Basic bioinformatics analysis completed, including the physicochemical properties of proteins, gene structures, protein motifs and conserved domains. Multiple cis-acting elements related to stress responses and plant growth were found in promoter regions. Further, 73 genes were localized on 16 pseudochromosomes and 29 pairs of paralogous genes were found via intraspecific collinearity analysis. Furthermore, tissue-specific expression was found in different tissues and germination stages. There were two BsPYLs, 10 BsPP2Cs, and four BsSnRK2 genes that exhibited a difference in response to multiple abiotic stresses. Moreover, subcellular localization analysis revealed six important proteins BsPP2C22, BsPP2C38, BsPP2C64, BsPYL2, BsPYL8, and BsSnRK2.4 which were localized in the nucleus and plasma membrane. Finally, yeast two-hybrid (Y2H) and bimolecular fluorescence complementation (BiFC) assays suggested that BsPP2C22 and BsPP2C38 could interact with multiple BsPYLs and BsSnRK2s proteins. This study systematically reported on the identification and characterization of the PYL-PP2C-SnRK2s family in B. striata, which provided a conceptual basis for deep insights into the functionality of ABA core signal pathways in Orchidaceae.

Integrated physiological and transcriptomic analyses reveal the molecular mechanism behind the response to cultivation in Quercus mongolica

Quercus mongolica, a common tree species for building and landscaping in northern China, has great commercial and ecological value. The seedlings of Q. mongolica grow poorly and develop chlorosis when introduced from high-altitude mountains to low-altitude plains. Effective cultivation measures are key to improving the quality of seedlings. To investigate the complex responses of Q. mongolica to different cultivation measures, we compared the adaptability of 3-year-old Q. mongolica seedlings to pruning (P), irrigation (W), and fertilization [F (nitro compound fertilizer with 16N-16P-16K)]. Physiological measurements and transcriptome sequencing were performed on leaves collected under the P treatments (control, cutting, removal of all lateral branches, and removal of base branches to one-third of seedling height), the W treatments (0, 1, 2, 3, 4, or 5 times in sequence), and the F treatments (0, 2, 4, and 6 g/plant). Analyses of the physiological data showed that P was more effective than W or F for activating intracellular antioxidant systems. By contrast, W and F were more beneficial than P for inducing the accumulation of soluble sugar. OPLS-DA identified superoxide dismutase, malondialdehyde, and peroxidase as critical physiological indices for the three cultivation measures. Transcriptome analyses revealed 1,012 differentially expressed genes (DEGs) in the P treatment, 1,035 DEGs in the W treatment, and 1,175 DEGs in the F treatment; these DEGs were mainly enriched in Gene Ontology terms related to the stress response and signal transduction. Weighted gene coexpression network analyses indicated that specific gene modules were significantly correlated with MDA (one module) and soluble sugar (four modules). Functional annotation of the hub genes differentially expressed in MDA and soluble sugar-related modules revealed that Q. mongolica responded and adapted to different cultivation measures by altering signal transduction, hormone levels, reactive oxygen species, metabolism, and transcription factors. The hub genes HOP3, CIPK11, WRKY22, and BHLH35 in the coexpression networks may played a central role in responses to the cultivation practices. These results reveal the mechanism behind the response of Q. mongolica to different cultivation measures at the physiological and molecular levels and provide insight into the response of plants to cultivation measures.

Transcriptomics and metabolomics changes triggered by exogenous 6-benzylaminopurine in relieving epicotyl dormancy of Polygonatum cyrtonema Hua seeds

Polygonatum cyrtonema Hua is one of the most useful herbs in traditional Chinese medicine and widely used in medicinal and edible perennial plant. However, the seeds have the characteristics of epicotyl dormancy. In this study, the molecular basis for relieving epicotyl dormancy of P. cyrtonema seeds under exogenous 6-benzylaminopurine (6-BA) treatment was revealed for the first time through transcriptome and metabolomics analysis. We determined the elongation of epicotyl buds as a critical period for dormancy release and found that the content of trans-zeatin, proline, auxin and gibberellin was higher, while flavonoids and arginine were lower in the treatment group. Transcriptome analysis showed that there were significant differences in gene expression in related pathways, and the expression patterns were highly consistent with the change of metabolites in corresponding pathways. Co-expression analysis showed that cytokinin dehydrogenase of P. cyrtonema (PcCKXs) and pelargonidin in flavonoid biosynthesis, as well as L-proline, L-ornithine, and L-citrulline in arginine and proline metabolism form network modules, indicating that they have related regulatory roles. Above all, our findings provide new insight into the exogenous 6-BA relieving epicotyl dormancy of P. cyrtonema seeds.

Systems biology-based analysis indicates that PHO1;H10 positively modulates high light-induced anthocyanin biosynthesis in Arabidopsis leaves

Arabidopsis PHO1;H10 is a member of the PHO1 gene family with SPX and EXS domains, and its functions remain largely unknown. As shown in PCSD database, the upstream region of PHO1;H10 gene is in the active chromatin states, with high DHS accessibility and binding sites of multiple transcription factors, especially ABI5, SPCH and HY5. Co-expression network and data-mining analyses showed PHO1;H10 and co-expression genes were with activation under high light stress. We did wet-lab experiments, and found that the detached leaves of PHO1;H10 overexpression lines accumulated more anthocyanin than those of WT and mutant under high light treatment. RNA-seq results showed overexpression of PHO1;H10 up-regulated many anthocyanin biosynthetic genes. The GSEA analysis result showed that the functional module related to anthocyanin pathway was significantly enriched. In summary, we conducted systems biology approach, combining dry- and wet-lab analyses, and discovered that PHO1;H10 might play an essential role during modulating high light-induced anthocyanin accumulation in the Arabidopsis detached leaves.

Applications of cell- and tissue-specific 'omics to improve plant productivity

The individual tissues and cell types of plants each have characteristic properties that contribute to the function of the plant as a whole. These are reflected by unique patterns of gene expression, protein and metabolite content, which enable cell-type-specific patterns of growth, development and physiology. Gene regulatory networks act within the cell types to govern the production and activity of these components. For the broader organism to grow and reproduce successfully, cell-type-specific activity must also function within the context of surrounding cell types, which is achieved by coordination of signalling pathways. We can investigate how gene regulatory networks are constructed and function using integrative 'omics technologies. Historically such experiments in plant biological research have been performed at the bulk tissue level, to organ resolution at best. In this review, we describe recent advances in cell- and tissue-specific 'omics technologies that allow investigation at much improved resolution. We discuss the advantages of these approaches for fundamental and translational plant biology, illustrated through the examples of specialised metabolism in medicinal plants and seed germination. We also discuss the challenges that must be overcome for such approaches to be adopted widely by the community.

Transcriptional Landscape of Cotton Fiber Development and Its Alliance With Fiber-Associated Traits

Cotton fiber development is still an intriguing question to understand fiber commitment and development. At different fiber developmental stages, many genes change their expression pattern and have a pivotal role in fiber quality and yield. Recently, numerous studies have been conducted for transcriptional regulation of fiber, and raw data were deposited to the public repository for comprehensive integrative analysis. Here, we remapped > 380 cotton RNAseq data with uniform mapping strategies that span ∼400 fold coverage to the genome. We identified stage-specific features related to fiber cell commitment, initiation, elongation, and Secondary Cell Wall (SCW) synthesis and their putative cis-regulatory elements for the specific regulation in fiber development. We also mined Exclusively Expressed Transcripts (EETs) that were positively selected during cotton fiber evolution and domestication. Furthermore, the expression of EETs was validated in 100 cotton genotypes through the nCounter assay and correlated with different fiber-related traits. Thus, our data mining study reveals several important features related to cotton fiber development and improvement, which were consolidated in the "CottonExpress-omics" database.

Gene co-expression analysis identifies modules related to insufficient sleep in humans

BACKGROUND

Insufficient sleep and circadian rhythm disruption may cause cancer, obesity, cardiovascular disease, and cognitive impairment. The underlying mechanisms need to be elucidated.

METHOD

Weighted gene co-expression network analysis (WGCNA) was used to identify co-expressed modules. Connectivity Map tool was used to identify candidate drugs based on top connected genes. R ptestg package was utilized to detected module rhythmicity alteration. A hypergeometric test was used to test the enrichment of insomnia SNP signals in modules. Google Scholar was used to validate the modules and hub genes by literature.

RESULTS

We identified a total of 45 co-expressed modules. These modules were stable and preserved. Eight modules were correlated with sleep restriction duration. Module rhythmicity was disrupted in sleep restriction subjects. Hub genes that involve in insufficient sleep also play important roles in sleep disorders. Insomnia GWAS signals were enriched in six modules. Finally, eight drugs associated with sleep disorders were identified.

CONCLUSION

Systems biology method was used to identify sleep-related modules, hub genes, and candidate drugs. Module rhythmicity was altered in sleep insufficient subjects. Thiamphenicol, lisuride, timolol, and piretanide are novel candidates for sleep disorders.

Identification of key genes associated with secondary metabolites biosynthesis by system network analysis in Valeriana officinalis

Valeriana officinalis is a medicinal plant, a source of bioactive chemical compounds and secondary metabolites which are applied in pharmaceutical industries. The advent of ethnomedicine has provided alternatives for disease treatment and has increased demands for natural products and bioactive compounds. A set of preliminary steps to answers for such demands can include integrative omics for systems metabolic engineering, as an approach that contributes to the understanding of cellular metabolic status. There is a growing trend of this approach for genetically engineering metabolic pathways in plant systems, by which natural and synthetic compounds can be produced. As in the case of most medicinal plants, there are no sufficient information about molecular mechanisms involved in the regulation of metabolic pathways in V. officinalis. In this research, systems biology was performed on the RNA-seq transcriptome and metabolome data to find key genes that contribute to the synthesis of major secondary metabolites in V. officinalis. The R Package Weighted Gene Co-Expression Network Analysis (WGCNA) was employed to analyze the data. Based on the results, some major modules and hub genes were identified to be associated with the valuable secondary metabolites. In addition, some TF-encoding genes, including AP2/ERF-ERF, WRKY and NAC TF families, as well as some regulatory factors including protein kinases and transporters were identified. The results showed that several novel hub genes, such as PCMP-H24, RPS24B, ANX1 and PXL1, may play crucial roles in metabolic pathways. The current findings provide an overall insight into the metabolic pathways of V. officinalis and can expand the potential for engineering genome-scale pathways and systems metabolic engineering to increase the production of bioactive compounds by plants.

Genome-Wide Mining of MYB Transcription Factors in the Anthocyanin Biosynthesis Pathway of Gossypium Hirsutum

The MYB family, one of the largest transcription factor (TF) families, plays an important role in plant growth, development, and stress response. Although genome-wide analysis of the MYB family has been performed in many species based on sequence similarity, predicting the potential functions of the MYB genes and classifying the regulators into specific metabolic pathways remains difficult. In this study, using a hidden Markov model search and co-expression regulatory network analysis, we demonstrated a process to screen and identify potential MYB TFs in the anthocyanin biosynthesis pathway of Gossypium hirsutum. As a result, we identified 617 and 784 MYB genes (812 in total) from the previously reported and recently released genomes, respectively. Using 126 structural genes involved in the anthocyanin biosynthesis pathway as targets for several co-expression network analyses, we sorted out 31 R2R3-MYB genes, which are potential regulators in the specific pathway. Phylogenetic and collinearity analyses indicated that 83.9% of the 31 MYB genes originated from whole genome duplication or polyploidization. In addition, we revealed relatively specific regulatory relationships between the MYB TFs and their target structural genes. Approximately, 71% of the MYBs could regulate only a single anthocyanin-related structural gene. Moreover, we found that the A- and D- subgenome homoeologs of MYB TFs in G. hirsutum rarely co-regulate the same target gene. The current study not only demonstrated an easy method to rapidly predict potential TFs in a specific metabolic pathway, but also enhanced our understanding of the evolution, gene characteristics, expression, and regulatory pattern of MYB TFs in G. hirsutum.

Co-expression Networks From Gene Expression Variability Between Genetically Identical Seedlings Can Reveal Novel Regulatory Relationships

Co-expression networks are a powerful tool to understand gene regulation. They have been used to identify new regulation and function of genes involved in plant development and their response to the environment. Up to now, co-expression networks have been inferred using transcriptomes generated on plants experiencing genetic or environmental perturbation, or from expression time series. We propose a new approach by showing that co-expression networks can be constructed in the absence of genetic and environmental perturbation, for plants at the same developmental stage. For this, we used transcriptomes that were generated from genetically identical individual plants that were grown under the same conditions and for the same amount of time. Twelve time points were used to cover the 24-h light/dark cycle. We used variability in gene expression between individual plants of the same time point to infer a co-expression network. We show that this network is biologically relevant and use it to suggest new gene functions and to identify new targets for the transcriptional regulators GI, PIF4, and PRR5. Moreover, we find different co-regulation in this network based on changes in expression between individual plants, compared to the usual approach requiring environmental perturbation. Our work shows that gene co-expression networks can be identified using variability in gene expression between individual plants, without the need for genetic or environmental perturbations. It will allow further exploration of gene regulation in contexts with subtle differences between plants, which could be closer to what individual plants in a population might face in the wild.

CoCoCoNet: conserved and comparative co-expression across a diverse set of species

Co-expression analysis has provided insight into gene function in organisms from Arabidopsis to zebrafish. Comparison across species has the potential to enrich these results, for example by prioritizing among candidate human disease genes based on their network properties or by finding alternative model systems where their co-expression is conserved. Here, we present CoCoCoNet as a tool for identifying conserved gene modules and comparing co-expression networks. CoCoCoNet is a resource for both data and methods, providing gold standard networks and sophisticated tools for on-the-fly comparative analyses across 14 species. We show how CoCoCoNet can be used in two use cases. In the first, we demonstrate deep conservation of a nucleolus gene module across very divergent organisms, and in the second, we show how the heterogeneity of autism mechanisms in humans can be broken down by functional groups and translated to model organisms. CoCoCoNet is free to use and available to all at https://milton.cshl.edu/CoCoCoNet, with data and R scripts available at ftp://milton.cshl.edu/data.

CoFly: A gene coexpression database for the fruit fly Drosophila melanogaster

The fruit fly Drosophila melanogaster can be used as a model organism for studying various problems in biomedicine and pest management. A large number of fruit fly transcriptomes have been profiled in various cell types, tissues, development stages, toxicological exposures, and other conditions by microarray. Until now, there are still no database developed for exploring those precious data. Microarray data for 4,367 samples from National Center for Biotechnology Information Gene Expression Omnibus was collected, and analyzed by weighted gene coexpression network analysis algorithm. Fifty one gene coexpression modules that are related to cell types, tissues, development stages, and other experimental conditions were identified. The high dimensional gene expression was reduced to tens of modules that were associated with experiments/traits, representing signatures for phenotypes. Six modules were enriched with genomic regions of clustered genes. Hub genes could also be screened by intramodule connectivity. By analyzing higher order module networks, we found that cell signaling modules are more connected than other modules. Module-based gene function identification may help to discover novel gene function. An easy-to-use database was developed, which provides a new source for gene function study in the fruit fly (http://bioinformatics.fafu.edu.cn/fly/).

SemanticGO: a tool for gene functional similarity analysis in Arabidopsis thaliana and rice

Gene or pathway functional similarities are important information for researchers. However, these similarities are often described sparsely and qualitatively. The latent semantic analysis of Arabidopsis thaliana (Arabidopsis) Gene Ontology (GO) data produced a set of 200-dimension feature vectors for each gene. Pathways were represented by summing the vectors of the pathway member genes. Thus, the similarities between genes and pathways were assessed. Additionally, the gene feature vectors were correlated with external gene data, including gene expression and gene network connectivity, to elucidate the associated functions. The gene feature vectors were decoded, and their applications were demonstrated. A simple online tool, SemanticGO (http://bioinformatics.fafu.edu.cn/semanticGO/), is herein provided to enable researchers to explore the similarities between genes and pathways in both Arabidopsis and rice.

Large-scale analysis of zebrafish (Danio rerio) transcriptomes identifies functional modules associated with phenotypes

Zebrafish (Danio rerio) is an excellent model for biomedicine research due to its genetic accessibility and optical transparency. A large number of microarray based transcriptomes of zebrafish have been profiled in various cell types, tissues, development stages, toxicological exposures and other conditions. However, there is still no easy-to-use web tool to explore those precious data. We downloaded 1434 microarray data from National Center for Biotechnology Information Gene Expression Omnibus (NCBI GEO), constructed weighted gene co-expression network, and identified 50 modules of co-expressed genes that correspond to different cell types, tissues, development stages, and other experimental conditions. These modules were associated with experiments/traits, and may serve signature modules for phenotypes. Hub genes were screened by intra-modular connectivity. Higher-order module networks analysis suggested that nucleus and cell cycle modules are densely connected. Module-based gene function identification may help to discover novel gene function. Our web tool provides a new resource for gene function study in zebrafish (http://bioinformatics.fafu.edu.cn/zebrafish/).

A Co-Expression Network in Hexaploid Wheat Reveals Mostly Balanced Expression and Lack of Significant Gene Loss of Homeologous Meiotic Genes Upon Polyploidization

Polyploidization has played an important role in plant evolution. However, upon polyploidization, the process of meiosis must adapt to ensure the proper segregation of increased numbers of chromosomes to produce balanced gametes. It has been suggested that meiotic gene (MG) duplicates return to a single copy following whole genome duplication to stabilize the polyploid genome. Therefore, upon the polyploidization of wheat, a hexaploid species with three related (homeologous) genomes, the stabilization process may have involved rapid changes in content and expression of MGs on homeologous chromosomes (homeologs). To examine this hypothesis, sets of candidate MGs were identified in wheat using co-expression network analysis and orthology informed approaches. In total, 130 RNA-Seq samples from a range of tissues including wheat meiotic anthers were used to define co-expressed modules of genes. Three modules were significantly correlated with meiotic tissue samples but not with other tissue types. These modules were enriched for GO terms related to cell cycle, DNA replication, and chromatin modification and contained orthologs of known MGs. Overall, 74.4% of genes within these meiosis-related modules had three homeologous copies which was similar to other tissue-related modules. Amongst wheat MGs identified by orthology, rather than co-expression, the majority (93.7%) were either retained in hexaploid wheat at the same number of copies (78.4%) or increased in copy number (15.3%) compared to ancestral wheat species. Furthermore, genes within meiosis-related modules showed more balanced expression levels between homeologs than genes in non-meiosis-related modules. Taken together, our results do not support extensive gene loss nor changes in homeolog expression of MGs upon wheat polyploidization. The construction of the MG co-expression network allowed identification of hub genes and provided key targets for future studies.