WeGET: predicting new genes for molecular systems by weighted co-expression

Adaptive evolution of chloroplast division mechanisms during plant terrestrialization

Despite extensive research, the origin and evolution of the chloroplast division machinery remain unclear. Here, we employ recently sequenced genomes and transcriptomes of Archaeplastida clades to identify the core components of chloroplast division and reconstruct their evolutionary histories, respectively. Our findings show that complete division ring structures emerged in Charophytes. We find that Glaucophytes experienced strong selection pressure, generating diverse variants adapted to the changing terrestrial environments. By integrating the functions of chloroplast division genes (CDGs) annotated in a workflow developed using large-scale multi-omics data, we further show that dispersed duplications acquire more species-specific functions under stronger selection pressures. Notably, PARC6, a dispersed duplicate CDG, regulates leaf color and plant growth in Solanum lycopersicum, demonstrating neofunctionalization. Our findings provide an integrated perspective on the functional evolution of chloroplast division machinery and highlight the potential of dispersed duplicate genes as the primary source of adaptive evolution of chloroplast division.

Integrated network analyses identify MYB4R1 neofunctionalization in the UV-B adaptation of Tartary buckwheat

A hallmark of adaptive evolution is innovation in gene function, which is associated with the development of distinct roles for genes during plant evolution; however, assessing functional innovation over long periods of time is not trivial. Tartary buckwheat (Fagopyrum tataricum) originated in the Himalayan region and has been exposed to intense UV-B radiation for a long time, making it an ideal species for studying novel UV-B response mechanisms in plants. Here, we developed a workflow to obtain a co-functional network of UV-B responses using data from more than 10,000 samples in more than 80 projects with multi-species and multi-omics data. Dissecting the entire network revealed that flavonoid biosynthesis was most significantly related to the UV-B response. Importantly, we found that the regulatory factor MYB4R1, which resides at the core of the network, has undergone neofunctionalization. In vitro and in vivo experiments demonstrated that MYB4R1 regulates flavonoid and anthocyanin accumulation in response to UV-B in buckwheat by binding to L-box motifs in the FtCHS, FtFLS, and FtUFGT promoters. We used deep learning to develop a visual discrimination model of buckwheat flavonoid content based on natural populations exposed to global UV-B radiation. Our study highlights the critical role of gene neofunctionalization in UV-B adaptation.

A cross-species co-functional gene network underlying leaf senescence

The complex leaf senescence process is governed by various levels of transcriptional and translational regulation. Several features of the leaf senescence process are similar across species, yet the extent to which the molecular mechanisms underlying the process of leaf senescence are conserved remains unclear. Currently used experimental approaches permit the identification of individual pathways that regulate various physiological and biochemical processes; however, the large-scale regulatory network underpinning intricate processes like leaf senescence cannot be built using these methods. Here, we discovered a series of conserved genes involved in leaf senescence in a common horticultural crop (Solanum lycopersicum), a monocot plant (Oryza sativa), and a eudicot plant (Arabidopsis thaliana) through analyses of the evolutionary relationships and expression patterns among genes. Our analyses revealed that the genetic basis of leaf senescence is largely conserved across species. We also created a multi-omics workflow using data from more than 10 000 samples from 85 projects and constructed a leaf senescence-associated co-functional gene network with 2769 conserved, high-confidence functions. Furthermore, we found that the mitochondrial unfolded protein response (UPR^mt) is the central biological process underlying leaf senescence. Specifically, UPR^mt responds to leaf senescence by maintaining mitostasis through a few cross-species conserved transcription factors (e.g. NAC13) and metabolites (e.g. ornithine). The co-functional network built in our study indicates that UPR^mt figures prominently in cross-species conserved mechanisms. Generally, the results of our study provide new insights that will aid future studies of leaf senescence.

COXPRESdb v8: an animal gene coexpression database navigating from a global view to detailed investigations

Gene coexpression is synchronization of gene expression across many cellular and environmental conditions and is widely used to infer the biological function of genes. Gene coexpression information is complex, comprising a complete graph of all genes in the genome, and requires appropriate visualization and analysis tools. Since its initial release in 2007, the animal gene expression database COXPRESdb (https://coxpresdb.jp) has been continuously improved by adding new gene coexpression data and analysis tools. Here, we report COXPRESdb version 8, which has been enhanced with new features for an overview, summary, and individual examination of coexpression relationships: CoexMap to display coexpression on a genome scale, pathway enrichment analysis to summarize the function of coexpressed genes, and CoexPub to bridges coexpression and existing knowledge. COXPRESdb also facilitates downstream analyses such as interspecies comparisons by integrating RNAseq and microarray coexpression data in a union-type gene coexpression. COXPRESdb strongly support users with the new coexpression data and enhanced functionality.

Peripheral Ion Channel Gene Screening in Painful- and Painless-Diabetic Neuropathy

Neuropathic pain is common in diabetic peripheral neuropathy (DN), probably caused by pathogenic ion channel gene variants. Therefore, we performed molecular inversion probes-next generation sequencing of 5 transient receptor potential cation channels, 8 potassium channels and 2 calcium-activated chloride channel genes in 222 painful- and 304 painless-DN patients. Twelve painful-DN (5.4%) patients showed potentially pathogenic variants (five nonsense/frameshift, seven missense, one out-of-frame deletion) in ANO3 (n = 3), HCN1 (n = 1), KCNK18 (n = 2), TRPA1 (n = 3), TRPM8 (n = 3) and TRPV4 (n = 1) and fourteen painless-DN patients (4.6%-three nonsense/frameshift, nine missense, one out-of-frame deletion) in ANO1 (n = 1), KCNK18 (n = 3), KCNQ3 (n = 1), TRPA1 (n = 2), TRPM8 (n = 1), TRPV1 (n = 3) and TRPV4 (n = 3). Missense variants were present in both conditions, presumably with loss- or gain-of-functions. KCNK18 nonsense/frameshift variants were found in painless/painful-DN, making a causal role in pain less likely. Surprisingly, premature stop-codons with likely nonsense-mediated RNA-decay were more frequent in painful-DN. Although limited in number, painful-DN patients with ion channel gene variants reported higher maximal pain during the night and day. Moreover, painful-DN patients with TRP variants had abnormal thermal thresholds and more severe pain during the night and day. Our results suggest a role of ion channel gene variants in neuropathic pain, but functional validation is required.

ATTED-II v11: A Plant Gene Coexpression Database Using a Sample Balancing Technique by Subagging of Principal Components

ATTED-II (https://atted.jp) is a gene coexpression database for nine plant species based on publicly available RNAseq and microarray data. One of the challenges in constructing condition-independent coexpression data based on publicly available gene expression data is managing the inherent sampling bias. Here, we report ATTED-II version 11, wherein we adopted a coexpression calculation methodology to balance the samples using principal component analysis and ensemble calculation. This approach has two advantages. First, omitting principal components with low contribution rates reduces the main contributors of noise. Second, balancing large differences in contribution rates enables considering various sample conditions entirely. In addition, based on RNAseq- and microarray-based coexpression data, we provide species-representative, integrated coexpression information to enhance the efficiency of interspecies comparison of the coexpression data. These coexpression data are provided as a standardized z-score to facilitate integrated analysis with different data sources. We believe that with these improvements, ATTED-II is more valuable and powerful for supporting interspecies comparative studies and integrated analyses using heterogeneous data.

Machine learning prediction of novel pectinolytic enzymes in Aspergillus niger through integrating heterogeneous (post-) genomics data

Pectinolytic enzymes are a variety of enzymes involved in breaking down pectin, a complex and abundant plant cell-wall polysaccharide. In nature, pectinolytic enzymes play an essential role in allowing bacteria and fungi to depolymerize and utilize pectin. In addition, pectinases have been widely applied in various industries, such as the food, wine, textile, paper and pulp industries. Due to their important biological function and increasing industrial potential, discovery of novel pectinolytic enzymes has received global interest. However, traditional enzyme characterization relies heavily on biochemical experiments, which are time consuming, laborious and expensive. To accelerate identification of novel pectinolytic enzymes, an automatic approach is needed. We developed a machine learning (ML) approach for predicting pectinases in the industrial workhorse fungus, Aspergillus niger. The prediction integrated a diverse range of features, including evolutionary profile, gene expression, transcriptional regulation and biochemical characteristics. Results on both the training and the independent testing dataset showed that our method achieved over 90 % accuracy, and recalled over 60 % of pectinolytic genes. Application of the ML model on the A. niger genome led to the identification of 83 pectinases, covering both previously described pectinases and novel pectinases that do not belong to any known pectinolytic enzyme family. Our study demonstrated the tremendous potential of ML in discovery of new industrial enzymes through integrating heterogeneous (post-) genomimcs data.

A Prioritized and Validated Resource of Mitochondrial Proteins in Plasmodium Identifies Unique Biology

Plasmodium species have a single mitochondrion that is essential for their survival and has been successfully targeted by antimalarial drugs. Most mitochondrial proteins are imported into this organelle, and our picture of the Plasmodium mitochondrial proteome remains incomplete. Many data sources contain information about mitochondrial localization, including proteome and gene expression profiles, orthology to mitochondrial proteins from other species, coevolutionary relationships, and amino acid sequences, each with different coverage and reliability. To obtain a comprehensive, prioritized list of Plasmodium falciparum mitochondrial proteins, we rigorously analyzed and integrated eight data sets using Bayesian statistics into a predictive score per protein for mitochondrial localization. At a corrected false discovery rate of 25%, we identified 445 proteins with a sensitivity of 87% and a specificity of 97%. They include proteins that have not been identified as mitochondrial in other eukaryotes but have characterized homologs in bacteria that are involved in metabolism or translation. Mitochondrial localization of seven Plasmodium berghei orthologs was confirmed by epitope labeling and colocalization with a mitochondrial marker protein. One of these belongs to a newly identified apicomplexan mitochondrial protein family that in P. falciparum has four members. With the experimentally validated mitochondrial proteins and the complete ranked P. falciparum proteome, which we have named PlasmoMitoCarta, we present a resource to study unique proteins of Plasmodium mitochondria. IMPORTANCE The unique biology and medical relevance of the mitochondrion of the malaria parasite Plasmodium falciparum have made it the subject of many studies. However, we actually do not have a comprehensive assessment of which proteins reside in this organelle. Many omics data are available that are predictive of mitochondrial localization, such as proteomics data and expression data. Individual data sets are, however, rarely complete and can provide conflicting evidence. We integrated a wide variety of available omics data in a manner that exploits the relative strengths of the data sets. Our analysis gave a predictive score for the mitochondrial localization to each nuclear encoded P. falciparum protein and identified 445 likely mitochondrial proteins. We experimentally validated the mitochondrial localization of seven of the new mitochondrial proteins, confirming the quality of the complete list. These include proteins that have not been observed mitochondria before, adding unique mitochondrial functions to P. falciparum.

Whole exome sequencing reveals a homozygous C1QBP deletion as the cause of progressive external ophthalmoplegia and multiple mtDNA deletions

Whole exome sequencing (WES), analyzed with GENESIS and WeGET, revealed a homozygous deletion in the C1QBP gene in a patient with progressive external ophthalmoplegia (PEO) and multiple mtDNA deletions. The gene encodes the mitochondria-located complementary 1 Q subcomponent-binding protein, involved in mitochondrial homeostasis. Biallelic mutations in C1QBP cause mitochondrial cardiomyopathy and/or PEO with variable age of onset. Our patient showed only late-onset PEO-plus syndrome without overt cardiac involvement. Available data suggest that early-onset cardiomyopathy variants localize in important structural domains and PEO-plus variants in the coiled-coil region. Our patient demonstrates that C1QBP mutations should be considered in individuals with PEO with or without cardiomyopathy.

Mining conditions specific hub genes from RNA-Seq gene-expression data via biclustering and their application to drug discovery

Gene‐expression data is being widely used for various clinical research. It represents expression levels of thousands of genes across the various experimental conditions simultaneously. Mining conditions specific hub genes from gene‐expression data is a challenging task. Conditions specific hub genes signify the functional behaviour of bicluster across the subset of conditions and can act as prognostic or diagnostic markers of the diseases. In this study, the authors have introduced a new approach for identifying conditions specific hub genes from the RNA‐Seq data using a biclustering algorithm. In the proposed approach, efficient ‘runibic’ biclustering algorithm, the concept of gene co‐expression network and concept of protein–protein interaction network have been used for getting better performance. The result shows that the proposed approach extracts biologically significant conditions specific hub genes which play an important role in various biological processes and pathways. These conditions specific hub genes can be used as prognostic or diagnostic biomarkers. Conditions specific hub genes will be helpful to reduce the analysis time and increase the accuracy of further research. Also, they summarised application of the proposed approach to the drug discovery process.

Mutations in PDLIM5 are rare in dilated cardiomyopathy but are emerging as potential disease modifiers

BACKGROUND

A causal genetic mutation is found in 40% of families with dilated cardiomyopathy (DCM), leaving a large percentage of families genetically unsolved. This prevents adequate counseling and clear recommendations in these families. We aim to identify novel genes or modifiers associated with DCM.

METHODS

We performed computational ranking of human genes based on coexpression with a predefined set of genes known to be associated with DCM, which allowed us to prioritize gene candidates for their likelihood of being involved in DCM. Top candidates will be checked for variants in the available whole-exome sequencing data of 142 DCM patients. RNA was isolated from cardiac biopsies to investigate gene expression.

RESULTS

PDLIM5 was classified as the top candidate. An interesting heterozygous variant (189_190delinsGG) was found in a DCM patient with a known pathogenic truncating TTN-variant. The PDLIM5 loss-of-function (LoF) variant affected all cardiac-specific isoforms of PDLIM5 and no LoF variants were detected in the same region in a control cohort of 26,000 individuals. RNA expression of PDLIM5 and its direct interactors (MYOT, LDB3, and MYOZ2) was increased in cardiac tissue of this patient, indicating a possible compensatory mechanism. The PDLIM5 variant cosegregated with the TTN-variant and the phenotype, leading to a high disease penetrance in this family. A second patient was an infant with a homozygous 10 kb-deletion of exon 2 in PDLIM5 resulting in early-onset cardiac disease, showing the importance of PDLIM5 in cardiac function.

CONCLUSIONS

Heterozygous PDLIM5 variants are rare and therefore will not have a major contribution in DCM. Although they likely play a role in disease development as this gene plays a major role in contracting cardiomyocytes and homozygous variants lead to early-onset cardiac disease. Other environmental and/or genetic factors are probably necessary to unveil the cardiac phenotype in PDLIM5 mutation carriers.

Identifying gene function and module connections by the integration of multispecies expression compendia

The functions of many eukaryotic genes are still poorly understood. Here, we developed and validated a new method, termed GeneBridge, which is based on two linked approaches to impute gene function and bridge genes with biological processes. First, Gene-Module Association Determination (G-MAD) allows the annotation of gene function. Second, Module-Module Association Determination (M-MAD) allows predicting connectivity among modules. We applied the GeneBridge tools to large-scale multispecies expression compendia—1700 data sets with over 300,000 samples from human, mouse, rat, fly, worm, and yeast—collected in this study. G-MAD identifies novel functions of genes—for example, DDT in mitochondrial respiration and WDFY4 in T cell activation—and also suggests novel components for modules, such as for cholesterol biosynthesis. By applying G-MAD on data sets from respective tissues, tissue-specific functions of genes were identified—for instance, the roles of EHHADH in liver and kidney, as well as SLC6A1 in brain and liver. Using M-MAD, we identified a list of module-module associations, such as those between mitochondria and proteasome, mitochondria and histone demethylation, as well as ribosomes and lipid biosynthesis. The GeneBridge tools together with the expression compendia are available as an open resource, which will facilitate the identification of connections linking genes, modules, phenotypes, and diseases.

A Combined Mass Spectrometry and Data Integration Approach to Predict the Mitochondrial Poly(A) RNA Interacting Proteome

In order to synthesize the 13 oxidative phosphorylation proteins encoded by mammalian mtDNA, a large assortment of nuclear encoded proteins is required. These include mitoribosomal proteins and various RNA processing, modification and degradation enzymes. RNA crosslinking has been successfully applied to identify whole-cell poly(A) RNA-binding proteomes, but this method has not been adapted to identify mitochondrial poly(A) RNA-binding proteomes. Here we developed and compared two related methods that specifically enrich for mitochondrial poly(A) RNA-binding proteins and analyzed bound proteins using mass spectrometry. To obtain a catalog of the mitochondrial poly(A) RNA interacting proteome, we used Bayesian data integration to combine these two mitochondrial-enriched datasets as well as published whole-cell datasets of RNA-binding proteins with various online resources, such as mitochondrial localization from MitoCarta 2.0 and co-expression analyses. Our integrated analyses ranked the complete human proteome for the likelihood of mtRNA interaction. We show that at a specific, inclusive cut-off of the corrected false discovery rate (cFDR) of 69%, we improve the number of predicted proteins from 185 to 211 with our mass spectrometry data as input for the prediction instead of the published whole-cell datasets. The chosen cut-off determines the cFDR: the less proteins included, the lower the cFDR will be. For the top 100 proteins, inclusion of our data instead of the published whole-cell datasets improve the cFDR from 54% to 31%. We show that the mass spectrometry method most specific for mitochondrial RNA-binding proteins involves ex vivo 4-thiouridine labeling followed by mitochondrial isolation with subsequent in organello UV-crosslinking.

COSCEB: Comprehensive search for column-coherent evolution biclusters and its application to hub gene identification

Biclustering is an increasingly used data mining technique for searching groups of co-expressed genes across the subset of experimental conditions from the gene-expression data. The group of co-expressed genes is present in the form of various patterns called a bicluster. A bicluster provides significant insights related to the functionality of genes and plays an important role in various clinical applications such as drug discovery, biomarker discovery, gene network analysis, gene identification, disease diagnosis, pathway analysis etc. This paper presents a novel unsupervised approach 'COmprehensive Search for Column-Coherent Evolution Biclusters (COSCEB)' for a comprehensive search of biologically significant column-coherent evolution biclusters. The concept of column subspace extraction from each gene pair and Longest Common Contiguous Subsequence (LCCS) is employed to identify significant biclusters. The experiments have been performed on both synthetic as well as real datasets. The performance of COSCEB is evaluated with the help of key issues. The issues are comprehensive search, Deep OPSM bicluster, bicluster types, bicluster accuracy, bicluster size, noise, overlapping, output nature, computational complexity and biologically significant biclusters. The performance of COSCEB is compared with six all-time famous biclustering algorithms SAMBA, OPSM, xMotif, Bimax, Deep OPSM- and UniBic. The result shows that the proposed approach performs effectively on most of the issues and extracts all possible biologically significant column-coherent evolution biclusters which are far more than other biclustering algorithms. Along with the proposed approach, we have also presented the case study which shows the application of significant biclusters for hub gene identification.

CiliaCarta: An integrated and validated compendium of ciliary genes

The cilium is an essential organelle at the surface of mammalian cells whose dysfunction causes a wide range of genetic diseases collectively called ciliopathies. The current rate at which new ciliopathy genes are identified suggests that many ciliary components remain undiscovered. We generated and rigorously analyzed genomic, proteomic, transcriptomic and evolutionary data and systematically integrated these using Bayesian statistics into a predictive score for ciliary function. This resulted in 285 candidate ciliary genes. We generated independent experimental evidence of ciliary associations for 24 out of 36 analyzed candidate proteins using multiple cell and animal model systems (mouse, zebrafish and nematode) and techniques. For example, we show that OSCP1, which has previously been implicated in two distinct non-ciliary processes, causes ciliogenic and ciliopathy-associated tissue phenotypes when depleted in zebrafish. The candidate list forms the basis of CiliaCarta, a comprehensive ciliary compendium covering 956 genes. The resource can be used to objectively prioritize candidate genes in whole exome or genome sequencing of ciliopathy patients and can be accessed at http://bioinformatics.bio.uu.nl/john/syscilia/ciliacarta/.

Latent network-based representations for large-scale gene expression data analysis

Background

With the recent advancements in high-throughput experimental procedures, biologists are gathering huge quantities of data. A main priority in bioinformatics and computational biology is to provide system level analytical tools capable of meeting an ever-growing production of high-throughput biological data while taking into account its biological context. In gene expression data analysis, genes have widely been considered as independent components. However, a systemic view shows that they act synergistically in living cells, forming functional complexes and more generally a biological system.

Results

In this paper, we propose LatNet, a signal transformation framework that, starting from an initial large-scale gene expression data, allows to generate new representations based on latent network-based relationships between the genes. LatNet aims to leverage system level relations between the genes as an underlying hidden structure to derive the new transformed latent signals. We present a concrete implementation of our framework, based on a gene regulatory network structure and two signal transformation approaches, to quantify latent network-based activity of regulators, as well as gene perturbation signals. The new gene/regulator signals are at the level of each sample of the input data and, thus, could directly be used instead of the initial expression signals for major bioinformatics analysis, including diagnosis and personalized medicine.

Conclusion

Multiple patterns could be hidden or weakly observed in expression data. LatNet helps in uncovering latent signals that could emphasize hidden patterns based on the relations between the genes and, thus, enhancing the performance of gene expression-based analysis algorithms. We use LatNet for the analysis of real-world gene expression data of bladder cancer and we show the efficiency of our transformation framework as compared to using the initial expression data.

Identification of Signaling Pathways Targeted by the Food Contaminant FB1: Transcriptome and Kinome Analysis of Samples from Pig Liver and Intestine

SCOPE

Fumonisin B1 (FB1) is a mycotoxin produced by Fusarium species. In mammals, this toxin causes widespread organ-specific damage; it promotes hepatotoxicity, is immunotoxic, alters intestinal functions etc. Despite its inhibitory effect on de novo ceramide synthesis, its molecular mechanism of action and toxicity is not totally elucidated.

METHODS AND RESULTS

To explore the mechanism of FB1 toxicity, we analyzed the transcriptome and the kinome of two organs targeted by FB1: the liver and the jejunum. Pigs were fed for 4 weeks a control diet or a FB1-contaminated diet (10 mg/kg). As expected, FB1-exposed pigs gained less weight and displayed a higher sphinganine/sphingosine ratio. Comparison of the transcriptomes and the kinomes of treated versus control pigs showed striking differences. Among the disrupted pathways in liver and jejunum, we highlight Protein Kinase B (AKT) / Phosphatase and tensin homolog (PTEN) at the intersection of the FB1-modulated pathways.

CONCLUSION

Most of the effects of FB1 are mediated by the regulation of ceramide level, which influences protein phosphatase 2 (PP2A) and the phosphoinositide 3-kinase (PI3K)/AKT signaling pathway. This pathway might be a new target to counteract toxic effect of Fumonisin B1, which is one of the most spread food contaminant in the world.

CLIC, a tool for expanding biological pathways based on co-expression across thousands of datasets

In recent years, there has been a huge rise in the number of publicly available transcriptional profiling datasets. These massive compendia comprise billions of measurements and provide a special opportunity to predict the function of unstudied genes based on co-expression to well-studied pathways. Such analyses can be very challenging, however, since biological pathways are modular and may exhibit co-expression only in specific contexts. To overcome these challenges we introduce CLIC, CLustering by Inferred Co-expression. CLIC accepts as input a pathway consisting of two or more genes. It then uses a Bayesian partition model to simultaneously partition the input gene set into coherent co-expressed modules (CEMs), while assigning the posterior probability for each dataset in support of each CEM. CLIC then expands each CEM by scanning the transcriptome for additional co-expressed genes, quantified by an integrated log-likelihood ratio (LLR) score weighted for each dataset. As a byproduct, CLIC automatically learns the conditions (datasets) within which a CEM is operative. We implemented CLIC using a compendium of 1774 mouse microarray datasets (28628 microarrays) or 1887 human microarray datasets (45158 microarrays). CLIC analysis reveals that of 910 canonical biological pathways, 30% consist of strongly co-expressed gene modules for which new members are predicted. For example, CLIC predicts a functional connection between protein C7orf55 (FMC1) and the mitochondrial ATP synthase complex that we have experimentally validated. CLIC is freely available at www.gene-clic.org. We anticipate that CLIC will be valuable both for revealing new components of biological pathways as well as the conditions in which they are active.

Genome-wide association studies identified multiple genetic loci for body size at four growth stages in Chinese Holstein cattle

The growth and maturity of cattle body size affect not only feed efficiency, but also productivity and longevity. Dissecting the genetic architecture of body size is critical for cattle breeding to improve both efficiency and productivity. The volume and weight of body size are indicated by several measurements. Among them, Heart Girth (HG) and Hip Height (HH) are the most important traits. They are widely used as predictors of body weight (BW). Few association studies have been conducted for HG and HH in cattle focusing on single growth stage. In this study, we extended the Genome-wide association studies to a full spectrum of four growth stages (6-, 12-, 18-, and 24-months after birth) in Chinese Holstein heifers. The whole genomic single nucleotide polymorphisms (SNPs) were obtained from the Illumina BovineSNP50 v2 BeadChip genotyped on 3,325 individuals. Estimated breeding values (EBVs) were derived for both HG and HH at the four different ages and analyzed separately for GWAS by using the Fixed and random model Circuitous Probability Unification (FarmCPU) method. In total, 27 SNPs were identified to be significantly associated with HG and HH at different growth stages. We found 66 candidate genes located nearby the associated SNPs, including nine genes that were known as highly related to development and skeletal and muscular growth. In addition, biological function analysis was performed by Ingenuity Pathway Analysis and an interaction network related to development was obtained, which contained 16 genes out of the 66 candidates. The set of putative genes provided valuable resources and can help elucidate the genomic architecture and mechanisms underlying growth traits in dairy cattle.

Exploiting single-cell expression to characterize co-expression replicability

Background

Co-expression networks have been a useful tool for functional genomics, providing important clues about the cellular and biochemical mechanisms that are active in normal and disease processes. However, co-expression analysis is often treated as a black box with results being hard to trace to their basis in the data. Here, we use both published and novel single-cell RNA sequencing (RNA-seq) data to understand fundamental drivers of gene-gene connectivity and replicability in co-expression networks.

Results

We perform the first major analysis of single-cell co-expression, sampling from 31 individual studies. Using neighbor voting in cross-validation, we find that single-cell network connectivity is less likely to overlap with known functions than co-expression derived from bulk data, with functional variation within cell types strongly resembling that also occurring across cell types. To identify features and analysis practices that contribute to this connectivity, we perform our own single-cell RNA-seq experiment of 126 cortical interneurons in an experimental design targeted to co-expression. By assessing network replicability, semantic similarity and overall functional connectivity, we identify technical factors influencing co-expression and suggest how they can be controlled for. Many of the technical effects we identify are expression-level dependent, making expression level itself highly predictive of network topology. We show this occurs generally through re-analysis of the BrainSpan RNA-seq data.

Conclusions

Technical properties of single-cell RNA-seq data create confounds in co-expression networks which can be identified and explicitly controlled for in any supervised analysis. This is useful both in improving co-expression performance and in characterizing single-cell data in generally applicable terms, permitting cross-laboratory comparison within a common framework.

Electronic supplementary material

The online version of this article (doi:10.1186/s13059-016-0964-6) contains supplementary material, which is available to authorized users.