Identification of co-expression gene networks, regulatory genes and pathways for obesity based on adipose tissue RNA Sequencing in a porcine model

Gene co-expression analysis identifies gene clusters associated with isotropic and polarized growth in Aspergillus fumigatus conidia

Aspergillus fumigatus is a saprophytic fungus that extensively produces conidia. These microscopic asexually reproductive structures are small enough to reach the lungs. Germination of conidia followed by hyphal growth inside human lungs is a key step in the establishment of infection in immunocompromised patients. RNA-Seq was used to analyze the transcriptome of dormant and germinating A. fumigatus conidia. Construction of a gene co-expression network revealed four gene clusters (modules) correlated with a growth phase (dormant, isotropic growth, polarized growth). Transcripts levels of genes encoding for secondary metabolites were high in dormant conidia. During isotropic growth, transcript levels of genes involved in cell wall modifications increased. Two modules encoding for growth and cell cycle/DNA processing were associated with polarized growth. In addition, the co-expression network was used to identify highly connected intermodular hub genes. These genes may have a pivotal role in the respective module and could therefore be compelling therapeutic targets. Generally, cell wall remodeling is an important process during isotropic and polarized growth, characterized by an increase of transcripts coding for hyphal growth and cell cycle/DNA processing when polarized growth is initiated.

Integrative analysis of microRNAs and mRNAs revealed regulation of composition and metabolism in Nelore cattle

BACKGROUND

The amount of intramuscular fat can influence the sensory characteristics and nutritional value of beef, thus the selection of animals with adequate fat deposition is important to the consumer. There is growing knowledge about the genes and pathways that control the biological processes involved in fat deposition in muscle. MicroRNAs (miRNAs) belong to a well-conserved class of non-coding small RNAs that modulate gene expression across a range of biological functions in animal development and physiology. The aim of this study was to identify differentially expressed (DE) miRNAs, regulatory candidate genes and co-expression networks related to intramuscular fat (IMF) deposition. To achieve this, we used mRNA and miRNA expression data from the Longissimus dorsi muscle of 30 Nelore steers with high (H) and low (L) genomic estimated breeding values (GEBV) for IMF deposition.

RESULTS

Differential miRNA expression analysis between animals with extreme GEBV values for IMF identified six DE miRNAs (FDR 10%). Functional annotation of the target genes for these microRNAs indicated that the PPARs signaling pathway is involved with IMF deposition. Candidate regulatory genes such as SDHAF4, FBXO17, ALDOA and PKM were identified by partial correlation with information theory (PCIT), phenotypic impact factor (PIF) and regulatory impact factor (RIF) co-expression approaches from integrated miRNA-mRNA expression data. Two DE miRNAs (FDR 10%), bta-miR-143 and bta-miR-146b, which were upregulated in the Low IMF group, were correlated with regulatory candidate genes, which were functionally enriched for fatty acid oxidation GO terms. Co-expression patterns obtained by weighted correlation network analysis (WGCNA), which showed possible interaction and regulation between mRNAs and miRNAs, identified several modules related to immune system function, protein metabolism, energy metabolism and glucose catabolism according to in silico analysis performed herein.

CONCLUSION

In this study, several genes and miRNAs were identified as candidate regulators of IMF by analyzing DE miRNAs using two different miRNA-mRNA co-expression network methods. This study contributes to the understanding of potential regulatory mechanisms of gene signaling networks involved in fat deposition processes measured in muscle. Glucose metabolism and inflammation processes were the main pathways found in silico to influence intramuscular fat deposition in beef cattle in the integrative mRNA-miRNA co-expression analysis.

Meta-Analysis of Aedes aegypti Expression Datasets: Comparing Virus Infection and Blood-Fed Transcriptomes to Identify Markers of Virus Presence

The mosquito Aedes aegypti (L.) is vector of several arboviruses including dengue, yellow fever, chikungunya, and more recently zika. Previous transcriptomic studies have been performed to elucidate altered pathways in response to viral infection. However, the intrinsic coupling between alimentation and infection were unappreciated in these studies. Feeding is required for the initial mosquito contact with the virus and these events are highly dependent. Addressing this relationship, we reinterrogated datasets of virus-infected mosquitoes with two different diet schemes (fed and unfed mosquitoes), evaluating the metabolic cross-talk during both processes. We constructed coexpression networks with the differentially expressed genes of these comparison: virus-infected versus blood-fed mosquitoes and virus-infected versus unfed mosquitoes. Our analysis identified one module with 110 genes that correlated with infection status (representing ~0.7% of the A. aegypti genome). Furthermore, we performed a machine-learning approach and summarized the infection status using only four genes (AAEL012128, AAEL014210, AAEL002477, and AAEL005350). While three of the four genes were annotated as hypothetical proteins, AAEL012128 gene is a membrane amino acid transporter correlated with viral envelope binding. This gene alone is able to discriminate all infected samples and thus should have a key role to discriminate viral infection in the A. aegypti mosquito. Moreover, validation using external datasets found this gene as differentially expressed in four transcriptomic experiments. Therefore, these genes may serve as a proxy of viral infection in the mosquito and the others 106 identified genes provides a framework to future studies.

Association between polymorphisms in SLC15A1 and PLA2G16 genes and development of obesity in Chinese subjects

INTRODUCTION

The small peptide transporter 1 (PepT-1) and adipose phospholipase A2 (AdPLA) play a key role in the development of obesity. However, there are no data assessing the impact of PepT-1 (SLC15A1) and AdPLA (PLA2G16) variants on obesity susceptibility. Therefore, we assessed the contribution of 9 single-nucleotide polymorphisms (SNPs) between these two genes on obesity susceptibility in Chinese subjects.

MATERIALS AND METHODS

A total of 611 participants were enrolled in the study, and 9 SNPs in the SLC15A1 and PLA2G16 genes were selected. Blood samples were collected for genotyping. Overweight and obesity were established by body mass index. Regression analyses were performed to test for any association of genetic polymorphisms with weight abnormality.

RESULTS

The genotype frequencies (P=0.04 for rs9557029, P=0.027 for rs1289389) were significantly different between obese or overweight subjects and healthy controls. However, no significant difference in allele was found between these three groups (P>0.05). Further logistic regression analyses adjusted for age and sex also failed to reveal significant associations between overweight, obesity, and the selected SNPs (P>0.05).

CONCLUSION

Data indicate that the selected 9 SNPs in SLC15A1 and PLA2G16 genes were not related to obesity susceptibility in the Han Chinese population.

Differential expression and co-expression gene networks reveal candidate biomarkers of boar taint in non-castrated pigs

Boar taint (BT) is an offensive odour or taste observed in pork from a proportion of non-castrated male pigs. Surgical castration is effective in avoiding BT, but animal welfare issues have created an incentive for alternatives such as genomic selection. In order to find candidate biomarkers, gene expression profiles were analysed from tissues of non-castrated pigs grouped by their genetic merit of BT. Differential expression analysis revealed substantial changes with log-transformed fold changes of liver and testis from -3.39 to 2.96 and -7.51 to 3.53, respectively. Co-expression network analysis revealed one module with a correlation of -0.27 in liver and three modules with correlations of 0.31, -0.44 and -0.49 in testis. Differential expression and co-expression analysis revealed candidate biomarkers with varying biological functions: phase I (COQ3, COX6C, CYP2J2, CYP2B6, ACOX2) and phase II metabolism (GSTO1, GSR, FMO3) of skatole and androstenone in liver to steroidgenesis (HSD17B7, HSD17B8, CYP27A1), regulation of steroidgenesis (STARD10, CYB5R3) and GnRH signalling (MAPK3, MAP2K2, MAP3K2) in testis. Overrepresented pathways included "Ribosome", "Protein export" and "Oxidative phosphorylation" in liver and "Steroid hormone biosynthesis" and "Gap junction" in testis. Future work should evaluate the biomarkers in large populations to ensure their usefulness in genomic selection programs.

Transcriptomic profiling in muscle and adipose tissue identifies genes related to growth and lipid deposition

Growth performance and meat quality are important traits for the pig industry and consumers. Adipose tissue is the main site at which fat storage and fatty acid synthesis occur. Therefore, we combined high-throughput transcriptomic sequencing in adipose and muscle tissues with the quantification of corresponding phenotypic features using seven Chinese indigenous pig breeds and one Western commercial breed (Yorkshire). We obtained data on 101 phenotypic traits, from which principal component analysis distinguished two groups: one associated with the Chinese breeds and one with Yorkshire. The numbers of differentially expressed genes between all Chinese breeds and Yorkshire were shown to be 673 and 1056 in adipose and muscle tissues, respectively. Functional enrichment analysis revealed that these genes are associated with biological functions and canonical pathways related to oxidoreductase activity, immune response, and metabolic process. Weighted gene coexpression network analysis found more coexpression modules significantly correlated with the measured phenotypic traits in adipose than in muscle, indicating that adipose regulates meat and carcass quality. Using the combination of differential expression, QTL information, gene significance, and module hub genes, we identified a large number of candidate genes potentially related to economically important traits in pig, which should help us improve meat production and quality.

Evaluation and improvement of the regulatory inference for large co-expression networks with limited sample size

BACKGROUND

Co-expression has been widely used to identify novel regulatory relationships using high throughput measurements, such as microarray and RNA-seq data. Evaluation studies on co-expression network analysis methods mostly focus on networks of small or medium size of up to a few hundred nodes. For large networks, simulated expression data usually consist of hundreds or thousands of profiles with different perturbations or knock-outs, which is uncommon in real experiments due to their cost and the amount of work required. Thus, the performances of co-expression network analysis methods on large co-expression networks consisting of a few thousand nodes, with only a small number of profiles with a single perturbation, which more accurately reflect normal experimental conditions, are generally uncharacterized and unknown.

METHODS

We proposed a novel network inference methods based on Relevance Low order Partial Correlation (RLowPC). RLowPC method uses a two-step approach to select on the high-confidence edges first by reducing the search space by only picking the top ranked genes from an intial partial correlation analysis and, then computes the partial correlations in the confined search space by only removing the linear dependencies from the shared neighbours, largely ignoring the genes showing lower association.

RESULTS

We selected six co-expression-based methods with good performance in evaluation studies from the literature: Partial correlation, PCIT, ARACNE, MRNET, MRNETB and CLR. The evaluation of these methods was carried out on simulated time-series data with various network sizes ranging from 100 to 3000 nodes. Simulation results show low precision and recall for all of the above methods for large networks with a small number of expression profiles. We improved the inference significantly by refinement of the top weighted edges in the pre-inferred partial correlation networks using RLowPC. We found improved performance by partitioning large networks into smaller co-expressed modules when assessing the method performance within these modules.

CONCLUSIONS

The evaluation results show that current methods suffer from low precision and recall for large co-expression networks where only a small number of profiles are available. The proposed RLowPC method effectively reduces the indirect edges predicted as regulatory relationships and increases the precision of top ranked predictions. Partitioning large networks into smaller highly co-expressed modules also helps to improve the performance of network inference methods. The RLowPC R package for network construction, refinement and evaluation is available at GitHub: https://github.com/wyguo/RLowPC .

Gene co-expression network analysis identifies porcine genes associated with variation in metabolizing fenbendazole and flunixin meglumine in the liver

Identifying individual genetic variation in drug metabolism pathways is of importance not only in livestock, but also in humans in order to provide the ultimate goal of giving the right drug at the right dose at the right time. Our objective was to identify individual genes and gene networks involved in metabolizing fenbendazole (FBZ) and flunixin meglumine (FLU) in swine liver. The population consisted of female and castrated male pigs that were sired by boars represented by 4 breeds. Progeny were randomly placed into groups: no drug (UNT), FLU or FBZ administered. Liver transcriptome profiles from 60 animals with extreme (i.e. fast or slow drug metabolism) pharmacokinetic (PK) profiles were generated from RNA sequencing. Multiple cytochrome P450 (CYP1A1, CYP2A19 and CYP2C36) genes displayed different transcript levels across treated versus UNT. Weighted gene co-expression network analysis identified 5 and 3 modules of genes correlated with PK parameters and a portion of these were enriched for biological processes relevant to drug metabolism for FBZ and FLU, respectively. Genes within identified modules were shown to have a higher transcript level relationship (i.e. connectivity) in treated versus UNT animals. Investigation into the identified genes would allow for greater insight into FBZ and FLU metabolism.

RNA-Seq transcriptomics and pathway analyses reveal potential regulatory genes and molecular mechanisms in high- and low-residual feed intake in Nordic dairy cattle

Background

The selective breeding of cattle with high-feed efficiencies (FE) is an important goal of beef and dairy cattle producers. Global gene expression patterns in relevant tissues can be used to study the functions of genes that are potentially involved in regulating FE. In the present study, high-throughput RNA sequencing data of liver biopsies from 19 dairy cows were used to identify differentially expressed genes (DEGs) between high- and low-FE groups of cows (based on Residual Feed Intake or RFI). Subsequently, a profile of the pathways connecting the DEGs to FE was generated, and a list of candidate genes and biomarkers was derived for their potential inclusion in breeding programmes to improve FE.

Results

The bovine RNA-Seq gene expression data from the liver was analysed to identify DEGs and, subsequently, identify the molecular mechanisms, pathways and possible candidate biomarkers of feed efficiency. On average, 57 million reads (short reads or short mRNA sequences < ~200 bases) were sequenced, 52 million reads were mapped, and 24,616 known transcripts were quantified according to the bovine reference genome. A comparison of the high- and low-RFI groups revealed 70 and 19 significantly DEGs in Holstein and Jersey cows, respectively. The interaction analysis (high vs. low RFI x control vs. high concentrate diet) showed no interaction effects in the Holstein cows, while two genes showed interaction effects in the Jersey cows. The analyses showed that DEGs act through certain pathways to affect or regulate FE, including steroid hormone biosynthesis, retinol metabolism, starch and sucrose metabolism, ether lipid metabolism, arachidonic acid metabolism and drug metabolism cytochrome P450.

Conclusion

We used RNA-Seq-based liver transcriptomic profiling of high- and low-RFI dairy cows in two breeds and identified significantly DEGs, their molecular mechanisms, their interactions with other genes and functional enrichments of different molecular pathways. The DEGs that were identified were the CYP’s and GIMAP genes for the Holstein and Jersey cows, respectively, which are related to the primary immunodeficiency pathway and play a major role in feed utilization and the metabolism of lipids, sugars and proteins.

Electronic supplementary material

The online version of this article (doi:10.1186/s12864-017-3622-9) contains supplementary material, which is available to authorized users.

Inter-Tissue Gene Co-Expression Networks between Metabolically Healthy and Unhealthy Obese Individuals

BACKGROUND

Obesity is associated with severe co-morbidities such as type 2 diabetes and nonalcoholic steatohepatitis. However, studies have shown that 10-25 percent of the severely obese individuals are metabolically healthy. To date, the identification of genetic factors underlying the metabolically healthy obese (MHO) state is limited. Systems genetics approaches have led to the identification of genes and pathways in complex diseases. Here, we have used such approaches across tissues to detect genes and pathways involved in obesity-induced disease development.

METHODS

Expression data of 60 severely obese individuals was accessible, of which 28 individuals were MHO and 32 were metabolically unhealthy obese (MUO). A whole genome expression profile of four tissues was available: liver, muscle, subcutaneous adipose tissue and visceral adipose tissue. Using insulin-related genes, we used the weighted gene co-expression network analysis (WGCNA) method to build within- and inter-tissue gene networks. We identified genes that were differentially connected between MHO and MUO individuals, which were further investigated by homing in on the modules they were active in. To identify potentially causal genes, we integrated genomic and transcriptomic data using an eQTL mapping approach.

RESULTS

Both IL-6 and IL1B were identified as highly differentially co-expressed genes across tissues between MHO and MUO individuals, showing their potential role in obesity-induced disease development. WGCNA showed that those genes were clustering together within tissues, and further analysis showed different co-expression patterns between MHO and MUO subnetworks. A potential causal role for metabolic differences under similar obesity state was detected for PTPRE, IL-6R and SLC6A5.

CONCLUSIONS

We used a novel integrative approach by integration of co-expression networks across tissues to elucidate genetic factors related to obesity-induced metabolic disease development. The identified genes and their interactions give more insight into the genetic architecture of obesity and the association with co-morbidities.

Genome-Wide Analysis and Functional Characterization of the Polyadenylation Site in Pigs Using RNAseq Data

Polyadenylation, a critical step in the production of mature mRNA for translation in most eukaryotes, involves cleavage and poly(A) tail addition at the 3′ end of mRNAs at the polyadenylation site (PAS). Sometimes, one gene can have more than one PAS, which can produce the alternative polyadenylation (APA) phenomenon and affect the stability, localization and translation of the mRNA. In this study, we discovered 28,363 PASs using pig RNAseq data, with 13,033 located in 7,403 genes. Among the genes, 41% were identified to have more than one PAS. PAS distribution analysis indicated that the PAS position was highly variable in genes. Additionally, the analysis of RNAseq data from the liver and testis showed a difference in their PAS number and usage. RT-PCR and qRT-PCR were performed to confirm our findings by detecting the expression of 3′UTR isoforms for five candidate genes. The analysis of RNAseq data under a different androstenone level and salmonella inoculation indicated that the functional usage of PAS might participate in the immune response and may be related to the androstenone level in pigs. This study provides new insights into pig PAS and facilitates further functional research of PAS.

Gene co-expression network analysis of dysferlinopathy: Altered cellular processes and functional prediction of TOR1AIP1, a novel muscular dystrophy gene

Dysferlinopathy, caused by a dysferlin gene mutation, is a clinically heterogeneous autosomal recessive muscle disease characterized by progressive muscle degeneration. The dysferlin protein's functions and dysferlinopathy disease pathogenesis are not fully explored, and there is no specific treatment available that can alter the disease progression. This study uses publicly available dysferlinopathy patient microarray data to construct a gene co-expression network and investigates significant cellular pathways and their key players in dysferlinopathy pathogenesis. Extracellular matrix deposition, inflammation, mitochondrial abnormalities and protein degradation were found to be important in dysferlinopathy. Out of the hub genes, OXR1 and TIMP1 were selected through literature search as candidate genes for possible biomarker and molecular therapeutic target studies. A recently identified muscular dystrophy gene TOR1AIP1 was detected as a hub gene in dysferlinopathy. Co-expression and protein sequence feature analysis were adopted to predict TOR1AIP1's function. Our results suggest that LAP1 protein encoded by TOR1AIP1 may play a role in protein degradation possibly through transcriptional regulation in muscle tissue. These findings extend dysferlinopathy pathogenesis by presenting key genes and also suggest a novel function for a poorly characterized gene.

RNAseq Analyses Identify Tumor Necrosis Factor-Mediated Inflammation as a Major Abnormality in ALS Spinal Cord

ALS is a rapidly progressive, devastating neurodegenerative illness of adults that produces disabling weakness and spasticity arising from death of lower and upper motor neurons. No meaningful therapies exist to slow ALS progression, and molecular insights into pathogenesis and progression are sorely needed. In that context, we used high-depth, next generation RNA sequencing (RNAseq, Illumina) to define gene network abnormalities in RNA samples depleted of rRNA and isolated from cervical spinal cord sections of 7 ALS and 8 CTL samples. We aligned >50 million 2X150 bp paired-end sequences/sample to the hg19 human genome and applied three different algorithms (Cuffdiff2, DEseq2, EdgeR) for identification of differentially expressed genes (DEG's). Ingenuity Pathways Analysis (IPA) and Weighted Gene Co-expression Network Analysis (WGCNA) identified inflammatory processes as significantly elevated in our ALS samples, with tumor necrosis factor (TNF) found to be a major pathway regulator (IPA) and TNFα-induced protein 2 (TNFAIP2) as a major network "hub" gene (WGCNA). Using the oPOSSUM algorithm, we analyzed transcription factors (TF) controlling expression of the nine DEG/hub genes in the ALS samples and identified TF's involved in inflammation (NFkB, REL, NFkB1) and macrophage function (NR1H2::RXRA heterodimer). Transient expression in human iPSC-derived motor neurons of TNFAIP2 (also a DEG identified by all three algorithms) reduced cell viability and induced caspase 3/7 activation. Using high-density RNAseq, multiple algorithms for DEG identification, and an unsupervised gene co-expression network approach, we identified significant elevation of inflammatory processes in ALS spinal cord with TNF as a major regulatory molecule. Overexpression of the DEG TNFAIP2 in human motor neurons, the population most vulnerable to die in ALS, increased cell death and caspase 3/7 activation. We propose that therapies targeted to reduce inflammatory TNFα signaling may be helpful in ALS patients.

NeuroRDF: semantic integration of highly curated data to prioritize biomarker candidates in Alzheimer's disease

BACKGROUND

Neurodegenerative diseases are incurable and debilitating indications with huge social and economic impact, where much is still to be learnt about the underlying molecular events. Mechanistic disease models could offer a knowledge framework to help decipher the complex interactions that occur at molecular and cellular levels. This motivates the need for the development of an approach integrating highly curated and heterogeneous data into a disease model of different regulatory data layers. Although several disease models exist, they often do not consider the quality of underlying data. Moreover, even with the current advancements in semantic web technology, we still do not have cure for complex diseases like Alzheimer's disease. One of the key reasons accountable for this could be the increasing gap between generated data and the derived knowledge.

RESULTS

In this paper, we describe an approach, called as NeuroRDF, to develop an integrative framework for modeling curated knowledge in the area of complex neurodegenerative diseases. The core of this strategy lies in the usage of well curated and context specific data for integration into one single semantic web-based framework, RDF. This increases the probability of the derived knowledge to be novel and reliable in a specific disease context. This infrastructure integrates highly curated data from databases (Bind, IntAct, etc.), literature (PubMed), and gene expression resources (such as GEO and ArrayExpress). We illustrate the effectiveness of our approach by asking real-world biomedical questions that link these resources to prioritize the plausible biomarker candidates. Among the 13 prioritized candidate genes, we identified MIF to be a potential emerging candidate due to its role as a pro-inflammatory cytokine. We additionally report on the effort and challenges faced during generation of such an indication-specific knowledge base comprising of curated and quality-controlled data.

CONCLUSION

Although many alternative approaches have been proposed and practiced for modeling diseases, the semantic web technology is a flexible and well established solution for harmonized aggregation. The benefit of this work, to use high quality and context specific data, becomes apparent in speculating previously unattended biomarker candidates around a well-known mechanism, further leveraged for experimental investigations.

Integrative Analysis of Genomics and Transcriptome Data to Identify Potential Functional Genes of BMDs in Females

Osteoporosis is known to be highly heritable. However, to date, the findings from more than 20 genome-wide association studies (GWASs) have explained less than 6% of genetic risks. Studies suggest that the missing heritability data may be because of joint effects among genes. To identify novel heritability for osteoporosis, we performed a system-level study on bone mineral density (BMD) by weighted gene coexpression network analysis (WGCNA), using the largest GWAS data set for BMD in the field, Genetic Factors for Osteoporosis Consortium (GEFOS-2), and a transcriptomic gene expression data set generated from transiliac bone biopsies in women. A weighted gene coexpression network was generated for 1574 genes with GWAS nominal evidence of association (p ≤ 0.05) based on dissimilarity measurement on the expression data. Twelve distinct gene modules were identified, and four modules showed nominally significant associations with BMD (p ≤ 0.05), but only one module, the yellow module, demonstrated a good correlation between module membership (MM) and gene significance (GS), suggesting that the yellow module serves an important biological role in bone regulation. Interestingly, through characterization of module content and topology, the yellow module was found to be significantly enriched with contractile fiber part (GO:044449), which is widely recognized as having a close relationship between muscle and bone. Furthermore, detailed submodule analyses of important candidate genes (HOMER1, SPTBN1) by all edges within the yellow module implied significant enrichment of functional connections between bone and cytoskeletal protein binding. Our study yielded novel information from system genetics analyses of GWAS data jointly with transcriptomic data. The findings highlighted a module and several genes in the model as playing important roles in the regulation of bone mass in females, which may yield novel insights into the genetic basis of osteoporosis. © 2016 American Society for Bone and Mineral Research.

Multi-omic data integration and analysis using systems genomics approaches: methods and applications in animal production, health and welfare

In the past years, there has been a remarkable development of high-throughput omics (HTO) technologies such as genomics, epigenomics, transcriptomics, proteomics and metabolomics across all facets of biology. This has spearheaded the progress of the systems biology era, including applications on animal production and health traits. However, notwithstanding these new HTO technologies, there remains an emerging challenge in data analysis. On the one hand, different HTO technologies judged on their own merit are appropriate for the identification of disease-causing genes, biomarkers for prevention and drug targets for the treatment of diseases and for individualized genomic predictions of performance or disease risks. On the other hand, integration of multi-omic data and joint modelling and analyses are very powerful and accurate to understand the systems biology of healthy and sustainable production of animals. We present an overview of current and emerging HTO technologies each with a focus on their applications in animal and veterinary sciences before introducing an integrative systems genomics framework for analysing and integrating multi-omic data towards improved animal production, health and welfare. We conclude that there are big challenges in multi-omic data integration, modelling and systems-level analyses, particularly with the fast emerging HTO technologies. We highlight existing and emerging systems genomics approaches and discuss how they contribute to our understanding of the biology of complex traits or diseases and holistic improvement of production performance, disease resistance and welfare.

Coexpression network analysis of the genes regulated by two types of resistance responses to powdery mildew in wheat

Powdery mildew disease caused by Blumeria graminis f. sp. tritici (Bgt) inflicts severe economic losses in wheat crops. A systematic understanding of the molecular mechanisms involved in wheat resistance to Bgt is essential for effectively controlling the disease. Here, using the diploid wheat Triticum urartu as a host, the genes regulated by immune (IM) and hypersensitive reaction (HR) resistance responses to Bgt were investigated through transcriptome sequencing. Four gene coexpression networks (GCNs) were developed using transcriptomic data generated for 20 T. urartu accessions showing IM, HR or susceptible responses. The powdery mildew resistance regulated (PMRR) genes whose expression was significantly correlated with Bgt resistance were identified, and they tended to be hubs and enriched in six major modules. A wide occurrence of negative regulation of PMRR genes was observed. Three new candidate immune receptor genes (TRIUR3_13045, TRIUR3_01037 and TRIUR3_06195) positively associated with Bgt resistance were discovered. Finally, the involvement of TRIUR3_01037 in Bgt resistance was tentatively verified through cosegregation analysis in a F2 population and functional expression assay in Bgt susceptible leaf cells. This research provides insights into the global network properties of PMRR genes. Potential molecular differences between IM and HR resistance responses to Bgt are discussed.

Bioinformatic analyses in early host response to Porcine Reproductive and Respiratory Syndrome virus (PRRSV) reveals pathway differences between pigs with alternate genotypes for a major host response QTL

Background

A region on Sus scrofa chromosome 4 (SSC4) surrounding single nucleotide polymorphism (SNP) marker WUR10000125 (WUR) has been reported to be strongly associated with both weight gain and serum viremia in pigs after infection with PRRS virus (PRRSV). A proposed causal mutation in the guanylate binding protein 5 gene (GBP5) is predicted to truncate the encoded protein. To investigate transcriptional differences between WUR genotypes in early host response to PRRSV infection, an RNA-seq experiment was performed on globin depleted whole blood RNA collected on 0, 4, 7, 10 and 14 days post-infection (dpi) from eight littermate pairs with one AB (favorable) and one AA (unfavorable) WUR genotype animal per litter.

Results

Gene Ontology (GO) enrichment analysis of transcripts that were differentially expressed (DE) between dpi across both genotypes revealed an inflammatory response for all dpi when compared to day 0. However, at the early time points of 4 and 7dpi, several GO terms had higher enrichment scores compared to later dpi, including inflammatory response (p < 10^-7), specifically regulation of NFkappaB (p < 0.01), cytokine, and chemokine activity (p < 0.01). At 10 and 14dpi, GO term enrichment indicated a switch to DNA damage response, cell cycle checkpoints, and DNA replication. Few transcripts were DE between WUR genotypes on individual dpi or averaged over all dpi, and little enrichment of any GO term was found. However, there were differences in expression patterns over time between AA and AB animals, which was confirmed by genotype-specific expression patterns of several modules that were identified in weighted gene co-expression network analyses (WGCNA). Minor differences between AA and AB animals were observed in immune response and DNA damage response (p = 0.64 and p = 0.11, respectively), but a significant effect between genotypes pointed to a difference in ion transport/homeostasis and the participation of G-coupled protein receptors (p = 8e-4), which was reinforced by results from regulatory and phenotypic impact factor analyses between genotypes.

Conclusion

We propose these pathway differences between WUR genotypes are the result of the inability of the truncated GBP5 of the AA genotyped pigs to inhibit viral entry and replication as quickly as the intact GBP5 protein of the AB genotyped pigs.

Electronic supplementary material

The online version of this article (doi:10.1186/s12864-016-2547-z) contains supplementary material, which is available to authorized users.

Transcriptional profiling of subcutaneous adipose tissue in Italian Large White pigs divergent for backfat thickness

Fat deposition is a widely studied trait in pigs because of its implications with animal growth efficiency, technological and nutritional characteristics of meat products, but the global framework of the biological and molecular processes regulating fat deposition in pigs is still incomplete. This study describes the backfat tissue transcription profile in Italian Large White pigs and reports genes differentially expressed between fat and lean animals according to RNA-seq data. The backfat transcription profile was characterised by the expression of 23 483 genes, of which 54.1% were represented by known genes. Of 63 418 expressed transcripts, about 80% were non-previously annotated isoforms. By comparing the expression level of fat vs. lean pigs, we detected 86 robust differentially expressed transcripts, 72 more highly expressed (e.g. ACP5, BCL2A1, CCR1, CD163, CD1A, EGR2, ENPP1, GPNMB, INHBB, LYZ, MSR1, OLR1, PIK3AP1, PLIN2, SPP1, SLC11A1, STC1) and 14 lower expressed (e.g. ADSSL1, CDO1, DNAJB1, HSPA1A, HSPA1B, HSPA2, HSPB8, IGFBP5, OLFML3) in fat pigs. The main functional categories enriched in differentially expressed genes were immune system process, response to stimulus, cell activation and skeletal system development, for the overexpressed genes, and unfolded protein binding and stress response, for the underexpressed genes, which included five heat shock proteins. Adipose tissue alterations and impaired stress response are linked to inflammation and, in turn, to adipose tissue secretory activity, similar to what is observed in human obesity. Our results provide the opportunity to identify biomarkers of carcass fat traits to improve the pig production chain and to identify genetic factors that regulate the observed differential expression.

Liver transcriptomic networks reveal main biological processes associated with feed efficiency in beef cattle

BACKGROUND

The selection of beef cattle for feed efficiency (FE) traits is very important not only for productive and economic efficiency but also for reduced environmental impact of livestock. Considering that FE is multifactorial and expensive to measure, the aim of this study was to identify biological functions and regulatory genes associated with this phenotype.

RESULTS

Eight genes were differentially expressed between high and low feed efficient animals (HFE and LFE, respectively). Co-expression analyses identified 34 gene modules of which 4 were strongly associated with FE traits. They were mainly enriched for inflammatory response or inflammation-related terms. We also identified 463 differentially co-expressed genes which were functionally enriched for immune response and lipid metabolism. A total of 8 key regulators of gene expression profiles affecting FE were found. The LFE animals had higher feed intake and increased subcutaneous and visceral fat deposition. In addition, LFE animals showed higher levels of serum cholesterol and liver injury biomarker GGT. Histopathology of the liver showed higher percentage of periportal inflammation with mononuclear infiltrate.

CONCLUSION

Liver transcriptomic network analysis coupled with other results demonstrated that LFE animals present altered lipid metabolism and increased hepatic periportal lesions associated with an inflammatory response composed mainly by mononuclear cells. We are now focusing to identify the causes of increased liver lesions in LFE animals.

Transcriptome analysis of an apple (Malus × domestica) yellow fruit somatic mutation identifies a gene network module highly associated with anthocyanin and epigenetic regulation

Using RNA-seq, this study analysed an apple (Malus×domestica) anthocyanin-deficient yellow-skin somatic mutant 'Blondee' (BLO) and its red-skin parent 'Kidd's D-8' (KID), the original name of 'Gala', to understand the molecular mechanisms underlying the mutation. A total of 3299 differentially expressed genes (DEGs) were identified between BLO and KID at four developmental stages and/or between two adjacent stages within BLO and/or KID. A weighted gene co-expression network analysis (WGCNA) of the DEGs uncovered a network module of 34 genes highly correlated (r=0.95, P=9.0×10(-13)) with anthocyanin contents. Although 12 of the 34 genes in the WGCNA module were characterized and known of roles in anthocyanin, the remainder 22 appear to be novel. Examining the expression of ten representative genes in the module in 14 diverse apples revealed that at least eight were significantly correlated with anthocyanin variation. MdMYB10 (MDP0000259614) and MdGST (MDP0000252292) were among the most suppressed module member genes in BLO despite being undistinguishable in their corresponding sequences between BLO and KID. Methylation assay of MdMYB10 and MdGST in fruit skin revealed that two regions (MR3 and MR7) in the MdMYB10 promoter exhibited remarkable differences between BLO and KID. In particular, methylation was high and progressively increased alongside fruit development in BLO while was correspondingly low and constant in KID. The methylation levels in both MR3 and MR7 were negatively correlated with anthocyanin content as well as the expression of MdMYB10 and MdGST. Clearly, the collective repression of the 34 genes explains the loss-of-colour in BLO while the methylation in MdMYB10 promoter is likely causal for the mutation.

An integrative systems genetics approach reveals potential causal genes and pathways related to obesity

Background

Obesity is a multi-factorial health problem in which genetic factors play an important role. Limited results have been obtained in single-gene studies using either genomic or transcriptomic data. RNA sequencing technology has shown its potential in gaining accurate knowledge about the transcriptome, and may reveal novel genes affecting complex diseases. Integration of genomic and transcriptomic variation (expression quantitative trait loci [eQTL] mapping) has identified causal variants that affect complex diseases. We integrated transcriptomic data from adipose tissue and genomic data from a porcine model to investigate the mechanisms involved in obesity using a systems genetics approach.

Methods

Using a selective gene expression profiling approach, we selected 36 animals based on a previously created genomic Obesity Index for RNA sequencing of subcutaneous adipose tissue. Differential expression analysis was performed using the Obesity Index as a continuous variable in a linear model. eQTL mapping was then performed to integrate 60 K porcine SNP chip data with the RNA sequencing data. Results were restricted based on genome-wide significant single nucleotide polymorphisms, detected differentially expressed genes, and previously detected co-expressed gene modules. Further data integration was performed by detecting co-expression patterns among eQTLs and integration with protein data.

Results

Differential expression analysis of RNA sequencing data revealed 458 differentially expressed genes. The eQTL mapping resulted in 987 cis-eQTLs and 73 trans-eQTLs (false discovery rate < 0.05), of which the cis-eQTLs were associated with metabolic pathways. We reduced the eQTL search space by focusing on differentially expressed and co-expressed genes and disease-associated single nucleotide polymorphisms to detect obesity-related genes and pathways. Building a co-expression network using eQTLs resulted in the detection of a module strongly associated with lipid pathways. Furthermore, we detected several obesity candidate genes, for example, ENPP1, CTSL, and ABHD12B.

Conclusions

To our knowledge, this is the first study to perform an integrated genomics and transcriptomics (eQTL) study using, and modeling, genomic and subcutaneous adipose tissue RNA sequencing data on obesity in a porcine model. We detected several pathways and potential causal genes for obesity. Further validation and investigation may reveal their exact function and association with obesity.

Electronic supplementary material

The online version of this article (doi:10.1186/s13073-015-0229-0) contains supplementary material, which is available to authorized users.

Uncovering co-expression gene network modules regulating fruit acidity in diverse apples

Background

Acidity is a major contributor to fruit quality. Several organic acids are present in apple fruit, but malic acid is predominant and determines fruit acidity. The trait is largely controlled by the Malic acid (Ma) locus, underpinning which Ma1 that putatively encodes a vacuolar aluminum-activated malate transporter1 (ALMT1)-like protein is a strong candidate gene. We hypothesize that fruit acidity is governed by a gene network in which Ma1 is key member. The goal of this study is to identify the gene network and the potential mechanisms through which the network operates.

Results

Guided by Ma1, we analyzed the transcriptomes of mature fruit of contrasting acidity from six apple accessions of genotype Ma_ (MaMa or Mama) and four of mama using RNA-seq and identified 1301 fruit acidity associated genes, among which 18 were most significant acidity genes (MSAGs). Network inferring using weighted gene co-expression network analysis (WGCNA) revealed five co-expression gene network modules of significant (P < 0.001) correlation with malate. Of these, the Ma1 containing module (Turquoise) of 336 genes showed the highest correlation (0.79). We also identified 12 intramodular hub genes from each of the five modules and 18 enriched gene ontology (GO) terms and MapMan sub-bines, including two GO terms (GO:0015979 and GO:0009765) and two MapMap sub-bins (1.3.4 and 1.1.1.1) related to photosynthesis in module Turquoise. Using Lemon-Tree algorithms, we identified 12 regulator genes of probabilistic scores 35.5–81.0, including MDP0000525602 (a LLR receptor kinase), MDP0000319170 (an IQD2-like CaM binding protein) and MDP0000190273 (an EIN3-like transcription factor) of greater interest for being one of the 18 MSAGs or one of the 12 intramodular hub genes in Turquoise, and/or a regulator to the cluster containing Ma1.

Conclusions

The most relevant finding of this study is the identification of the MSAGs, intramodular hub genes, enriched photosynthesis related processes, and regulator genes in a WGCNA module Turquoise that not only encompasses Ma1 but also shows the highest modular correlation with acidity. Overall, this study provides important insight into the Ma1-mediated gene network controlling acidity in mature apple fruit of diverse genetic background.

Electronic supplementary material

The online version of this article (doi:10.1186/s12864-015-1816-6) contains supplementary material, which is available to authorized users.

Potential role of lncRNA cyp2c91-protein interactions on diseases of the immune system

With unprecedented increase in next generation sequencing technologies, there has been a persistent interest on transcript profiles of long non-coding RNAs (lncRNAs) and protein-coding genes forming an interaction network. Apart from protein–protein interaction (PPI), gene network models such as Weighted Gene Co-expression Network Analysis (WGCNA) are used to functionally annotate lncRNAs in identifying their potential disease associations. To address this, studies have led to characterizing transcript structures and understanding expression profiles mediating regulatory roles. In the current exploratory analysis, we show how a lncRNA – cyp2c91 contributes to the transcriptional regulation localized to cytoplasm thereby making refractory environment for transcription. By applying network methods and pathway analyses on genes related to a disease such as obesity and systemic lupus erythematosus, we show that we can gain deeper insight in biological processes such as the perturbances in immune system, and get a better understanding of the systems biology of diseases.

Complex network models reveal correlations among network metrics, exercise intensity and role of body changes in the fatigue process

The aims of the present study were analyze the fatigue process at distinct intensity efforts and to investigate its occurrence as interactions at distinct body changes during exercise, using complex network models. For this, participants were submitted to four different running intensities until exhaustion, accomplished in a non-motorized treadmill using a tethered system. The intensities were selected according to critical power model. Mechanical (force, peak power, mean power, velocity and work) and physiological related parameters (heart rate, blood lactate, time until peak blood lactate concentration (lactate time), lean mass, anaerobic and aerobic capacities) and IPAQ score were obtained during exercises and it was used to construction of four complex network models. Such models have both, theoretical and mathematical value, and enables us to perceive new insights that go beyond conventional analysis. From these, we ranked the influences of each node at the fatigue process. Our results shows that nodes, links and network metrics are sensibility according to increase of efforts intensities, been the velocity a key factor to exercise maintenance at models/intensities 1 and 2 (higher time efforts) and force and power at models 3 and 4, highlighting mechanical variables in the exhaustion occurrence and even training prescription applications.

Reduced Representation Libraries from DNA Pools Analysed with Next Generation Semiconductor Based-Sequencing to Identify SNPs in Extreme and Divergent Pigs for Back Fat Thickness

The aim of this study was to identify single nucleotide polymorphisms (SNPs) that could be associated with back fat thickness (BFT) in pigs. To achieve this goal, we evaluated the potential and limits of an experimental design that combined several methodologies. DNA samples from two groups of Italian Large White pigs with divergent estimating breeding value (EBV) for BFT were separately pooled and sequenced, after preparation of reduced representation libraries (RRLs), on the Ion Torrent technology. Taking advantage from SNAPE for SNPs calling in sequenced DNA pools, 39,165 SNPs were identified; 1/4 of them were novel variants not reported in dbSNP. Combining sequencing data with Illumina PorcineSNP60 BeadChip genotyping results on the same animals, 661 genomic positions overlapped with a good approximation of minor allele frequency estimation. A total of 54 SNPs showing enriched alleles in one or in the other RRLs might be potential markers associated with BFT. Some of these SNPs were close to genes involved in obesity related phenotypes.