Revealing the architecture of gene regulation: the promise of eQTL studies

DeepCORE: An interpretable multi-view deep neural network model to detect co-operative regulatory elements

Gene transcription is an essential process involved in all aspects of cellular functions with significant impact on biological traits and diseases. This process is tightly regulated by multiple elements that co-operate to jointly modulate the transcription levels of target genes. To decipher the complicated regulatory network, we present a novel multi-view attention-based deep neural network that models the relationship between genetic, epigenetic, and transcriptional patterns and identifies co-operative regulatory elements (COREs). We applied this new method, named DeepCORE, to predict transcriptomes in various tissues and cell lines, which outperformed the state-of-the-art algorithms. Furthermore, DeepCORE contains an interpreter that extracts the attention values embedded in the deep neural network, maps the attended regions to putative regulatory elements, and infers COREs based on correlated attentions. The identified COREs are significantly enriched with known promoters and enhancers. Novel regulatory elements discovered by DeepCORE showed epigenetic signatures consistent with the status of histone modification marks.

SPATS2L is a positive feedback regulator of the type I interferon signaling pathway and plays a vital role in lupus

Through genome-wide association studies (GWAS) and integrated expression quantitative trait locus (eQTL) analyses, numerous susceptibility genes ("eGenes", whose expressions are significantly associated with common variants) associated with systemic lupus erythematosus (SLE) have been identified. Notably, a subset of these eGenes is correlated with disease activity. However, the precise mechanisms through which these genes contribute to the initiation and progression of the disease remain to be fully elucidated. In this investigation, we initially identify SPATS2L as an SLE eGene correlated with disease activity. eSignaling and transcriptomic analyses suggest its involvement in the type I interferon (IFN) pathway. We observe a significant increase in SPATS2L expression following type I IFN stimulation, and the expression levels are dependent on both the concentration and duration of stimulation. Furthermore, through dual-luciferase reporter assays, western blot analysis, and imaging flow cytometry, we confirm that SPATS2L positively modulates the type I IFN pathway, acting as a positive feedback regulator. Notably, siRNA-mediated intervention targeting SPATS2L, an interferon-inducible gene, in peripheral blood mononuclear cells (PBMCs) from patients with SLE reverses the activation of the interferon pathway. In conclusion, our research highlights the pivotal role of SPATS2L as a positive-feedback regulatory molecule within the type I IFN pathway. Our findings suggest that SPATS2L plays a critical role in the onset and progression of SLE and may serve as a promising target for disease activity assessment and intervention strategies.

Integrative Multi-omics Analysis to Characterize Herpes Virus Infection Increases the Risk of Alzheimer's Disease

Evidence suggests that herpes virus infection is associated with an increased risk of Alzheimer's disease (AD), and innate and adaptive immunity plays an important role in the association. Although there have been many studies, the mechanism of the association is still unclear. This study aims to reveal the underlying molecular and immune regulatory network through multi-omics data and provide support for the study of the mechanism of infection and AD in the future. Here, we found that the herpes virus infection significantly increased the risk of AD. Genes associated with the occurrence and development of AD and genetically regulated by herpes virus infection are mainly enrichment in immune-related pathways. The 22 key regulatory genes identified by machine learning are mainly immune genes. They are also significantly related to the infiltration changes of 3 immune cell in AD. Furthermore, many of these genes have previously been reported to be linked, or potentially linked, to the pathological mechanisms of both herpes virus infection and AD. In conclusion, this study contributes to the study of the mechanisms related to herpes virus infection and AD, and indicates that the regulation of innate and adaptive immunity may be an effective strategy for preventing and treating herpes virus infection and AD. Additionally, the identified key regulatory genes, whether previously studied or newly discovered, may serve as valuable targets for prevention and treatment strategies.

Animal board invited review: Heritability of nitrogen use efficiency in fattening pigs: Current state and possible directions

Pork, an important component of human nutrition worldwide, contributes considerably to anthropogenic nitrogen and greenhouse gas emissions. Reducing the environmental impact of pig production is therefore essential. This can be achieved through system-level strategies, such as optimising resource use, improving manure management and recycling leftovers from human food production, and at the individual animal level by maintaining pig health and fine-tuning dietary protein levels to individual requirements. Breeding, coupled with nutritional strategies, offers a lasting solution to improve nitrogen use efficiency (NUE) - the ratio of nitrogen retained in the body to nitrogen ingested. With a heritability as high as 0.54, incorporating NUE into breeding programmes appears promising. Nitrogen use efficiency involves multiple tissues and metabolic processes, and is influenced by the environment and individual animal characteristics, including its genetic background. Heritable genetic variation in NUE may therefore occur in many different processes, including the central nervous regulation of feed intake, the endocrine system, the gastrointestinal tract where digestion and absorption take place, and the composition of the gut microbiome. An animal's postabsorptive protein metabolism might also harbour important genetic variation, especially in the maintenance requirements of tissues and organs. Precise phenotyping, although challenging and costly, is essential for successful breeding. Various measurement techniques, such as imaging techniques and mechanistic models, are being explored for their potential in genetic analysis. Despite the difficulties in phenotyping, some studies have estimated the heritability and genetic correlations of NUE. These studies suggest that direct selection for NUE is more effective than indirect methods through feed efficiency. The complexity of NUE indicates a polygenic trait architecture, which has been confirmed by genome-wide association studies that have been unable to identify significant quantitative trait loci. Building sufficiently large reference populations to train genomic prediction models is an important next step. However, this will require the development of truly high-throughput phenotyping methods. In conclusion, breeding pigs with higher NUE is both feasible and necessary but will require increased efforts in high-throughput phenotyping and improved genome annotation.

A functional schizophrenia-associated genetic variant near the TSNARE1 and ADGRB1 genes

Recent collaborative genome-wide association studies (GWAS) have identified >200 independent loci contributing to risk for schizophrenia (SCZ). The genes closest to these loci have diverse functions, supporting the potential involvement of multiple relevant biological processes, yet there is no direct evidence that individual variants are functional or directly linked to specific genes. Nevertheless, overlap with certain epigenetic marks suggest that most GWAS-implicated variants are regulatory. Based on the strength of association with SCZ and the presence of regulatory epigenetic marks, we chose one such variant near TSNARE1 and ADGRB1, rs4129585, to test for functional potential and assay differences that may drive the pathogenicity of the risk allele. We observed that the variant-containing sequence drives reporter expression in relevant neuronal populations in zebrafish. Next, we introduced each allele into human induced pluripotent cells and differentiated four isogenic clones homozygous for the risk allele and five clones homozygous for the non-risk allele into neural progenitor cells. Employing RNA sequencing, we found that the two alleles yield significant transcriptional differences in the expression of 109 genes at a false discovery rate (FDR) of <0.05 and 259 genes at a FDR of <0.1. We demonstrate that these genes are highly interconnected in pathways enriched for synaptic proteins, axon guidance, and regulation of synapse assembly. Exploration of genes near rs4129585 suggests that this variant does not regulate TSNARE1 transcripts, as previously thought, but may regulate the neighboring ADGRB1, a regulator of synaptogenesis. Our results suggest that rs4129585 is a functional common variant that functions in specific pathways likely involved in SCZ risk.

Expression genome-wide association study identifies key regulatory variants enriched with metabolic and immune functions in four porcine tissues

BACKGROUND

Integration of high throughput DNA genotyping and RNA-sequencing data enables the discovery of genomic regions that regulate gene expression, known as expression quantitative trait loci (eQTL). In pigs, efforts to date have been mainly focused on purebred lines for traits with commercial relevance as such growth and meat quality. However, little is known on genetic variants and mechanisms associated with the robustness of an animal, thus its overall health status. Here, the liver, lung, spleen, and muscle transcriptomes of 100 three-way crossbred female finishers were studied, with the aim of identifying novel eQTL regulatory regions and transcription factors (TFs) associated with regulation of porcine metabolism and health-related traits.

RESULTS

An expression genome-wide association study with 535,896 genotypes and the expression of 12,680 genes in liver, 13,310 genes in lung, 12,650 genes in spleen, and 12,595 genes in muscle resulted in 4,293, 10,630, 4,533, and 6,871 eQTL regions for each of these tissues, respectively. Although only a small fraction of the eQTLs were annotated as cis-eQTLs, these presented a higher number of polymorphisms per region and significantly stronger associations with their target gene compared to trans-eQTLs. Between 20 and 115 eQTL hotspots were identified across the four tissues. Interestingly, these were all enriched for immune-related biological processes. In spleen, two TFs were identified: ERF and ZNF45, with key roles in regulation of gene expression.

CONCLUSIONS

This study provides a comprehensive analysis with more than 26,000 eQTL regions identified that are now publicly available. The genomic regions and their variants were mostly associated with tissue-specific regulatory roles. However, some shared regions provide new insights into the complex regulation of genes and their interactions that are involved with important traits related to metabolism and immunity.

Response eQTLs, chromatin accessibility, and 3D chromatin structure in chondrocytes provide mechanistic insight into osteoarthritis risk

Osteoarthritis (OA) poses a significant healthcare burden with limited treatment options. While genome-wide association studies (GWAS) have identified over 100 OA-associated loci, translating these findings into therapeutic targets remains challenging. Integrating expression quantitative trait loci (eQTL), 3D chromatin structure, and other genomic approaches with OA GWAS data offers a promising approach to elucidate disease mechanisms; however, comprehensive eQTL maps in OA-relevant tissues and conditions remain scarce. We mapped gene expression, chromatin accessibility, and 3D chromatin structure in primary human articular chondrocytes in both resting and OA-mimicking conditions. We identified thousands of differentially expressed genes, including those associated with differences in sex and age. RNA-seq in chondrocytes from 101 donors across two conditions uncovered 3782 unique eGenes, including 420 that exhibited strong and significant condition-specific effects. Colocalization with OA GWAS signals revealed 13 putative OA risk genes, 10 of which have not been previously identified. Chromatin accessibility and 3D chromatin structure provided insights into the mechanisms and conditional specificity of these variants. Our findings shed light on OA pathogenesis and highlight potential targets for therapeutic development.

Highlights

∘ Comprehensive analysis of sex- and age-related global gene expression in human chondrocytes revealed differences that correlate with osteoarthritis ∘ First response eQTLs in chondrocytes treated with an OA-related stimulus ∘ Deeply sequenced Hi-C in resting and activated chondrocytes helps connect OA risk variants to their putative causal genes ∘ Colocalization analysis reveals 13 (including 10 novel) putative OA risk genes.

The genetic regulation of the gastric transcriptome is associated with metabolic and obesity-related traits and diseases

Tissue-specific gene expression and gene regulation lead to a better understanding of tissue-specific physiology and pathophysiology. We analyzed the transcriptome and genetic regulatory profiles of two distinct gastric sites, corpus and antrum, to identify tissue-specific gene expression and its regulation. Gastric corpus and antrum mucosa biopsies were collected during routine gastroscopies from up to 431 healthy individuals. We obtained genotype and transcriptome data and performed transcriptome profiling and expression quantitative trait locus (eQTL) studies. We further used data from genome-wide association studies (GWAS) of various diseases and traits to partition their heritability and to perform transcriptome-wide association studies (TWAS). The transcriptome data from corpus and antral mucosa highlights the heterogeneity of gene expression in the stomach. We identified enriched pathways revealing distinct and common physiological processes in gastric corpus and antrum. Furthermore, we found an enrichment of the single nucleotide polymorphism (SNP)-based heritability of metabolic, obesity-related, and cardiovascular traits and diseases by considering corpus- and antrum-specifically expressed genes. Particularly, we could prioritize gastric-specific candidate genes for multiple metabolic traits, like NQO1 which is involved in glucose metabolism, MUC1 which contributes to purine and protein metabolism or RAB27B being a regulator of weight and body composition. Our findings show that gastric corpus and antrum vary in their transcriptome and genetic regulatory profiles indicating physiological differences which are mostly related to digestion and epithelial protection. Moreover, our findings demonstrate that the genetic regulation of the gastric transcriptome is linked to biological mechanisms associated with metabolic, obesity-related, and cardiovascular traits and diseases. NEW & NOTEWORTHY We analyzed the transcriptomes and genetic regulatory profiles of gastric corpus and for the first time also of antrum mucosa in 431 healthy individuals. Through tissue-specific gene expression and eQTL analyses, we uncovered unique and common physiological processes across both primary gastric sites. Notably, our findings reveal that stomach-specific eQTLs are enriched in loci associated with metabolic traits and diseases, highlighting the pivotal role of gene expression regulation in gastric physiology and potential pathophysiology.

Incorporating genetic similarity of auxiliary samples into eGene identification under the transfer learning framework

BACKGROUND

The term eGene has been applied to define a gene whose expression level is affected by at least one independent expression quantitative trait locus (eQTL). It is both theoretically and empirically important to identify eQTLs and eGenes in genomic studies. However, standard eGene detection methods generally focus on individual cis-variants and cannot efficiently leverage useful knowledge acquired from auxiliary samples into target studies.

METHODS

We propose a multilocus-based eGene identification method called TLegene by integrating shared genetic similarity information available from auxiliary studies under the statistical framework of transfer learning. We apply TLegene to eGene identification in ten TCGA cancers which have an explicit relevant tissue in the GTEx project, and learn genetic effect of variant in TCGA from GTEx. We also adopt TLegene to the Geuvadis project to evaluate its usefulness in non-cancer studies.

RESULTS

We observed substantial genetic effect correlation of cis-variants between TCGA and GTEx for a larger number of genes. Furthermore, consistent with the results of our simulations, we found that TLegene was more powerful than existing methods and thus identified 169 distinct candidate eGenes, which was much larger than the approach that did not consider knowledge transfer across target and auxiliary studies. Previous studies and functional enrichment analyses provided empirical evidence supporting the associations of discovered eGenes, and it also showed evidence of allelic heterogeneity of gene expression. Furthermore, TLegene identified more eGenes in Geuvadis and revealed that these eGenes were mainly enriched in cells EBV transformed lymphocytes tissue.

CONCLUSION

Overall, TLegene represents a flexible and powerful statistical method for eGene identification through transfer learning of genetic similarity shared across auxiliary and target studies.

Gene expression variation underlying tissue-specific responses to copper stress in Drosophila melanogaster

Copper is one of a handful of biologically necessary heavy metals that is also a common environmental pollutant. Under normal conditions, copper ions are required for many key physiological processes. However, in excess, copper results in cell and tissue damage ranging in severity from temporary injury to permanent neurological damage. Because of its biological relevance, and because many conserved copper-responsive genes respond to nonessential heavy metal pollutants, copper resistance in Drosophila melanogaster is a useful model system with which to investigate the genetic control of the heavy metal stress response. Because heavy metal toxicity has the potential to differently impact specific tissues, we genetically characterized the control of the gene expression response to copper stress in a tissue-specific manner in this study. We assessed the copper stress response in head and gut tissue of 96 inbred strains from the Drosophila Synthetic Population Resource using a combination of differential expression analysis and expression quantitative trait locus mapping. Differential expression analysis revealed clear patterns of tissue-specific expression. Tissue and treatment specific responses to copper stress were also detected using expression quantitative trait locus mapping. Expression quantitative trait locus associated with MtnA, Mdr49, Mdr50, and Sod3 exhibited both genotype-by-tissue and genotype-by-treatment effects on gene expression under copper stress, illuminating tissue- and treatment-specific patterns of gene expression control. Together, our data build a nuanced description of the roles and interactions between allelic and expression variation in copper-responsive genes, provide valuable insight into the genomic architecture of susceptibility to metal toxicity, and highlight candidate genes for future functional characterization.

Omics approaches in asthma research: Challenges and opportunities

Asthma, a chronic respiratory disease with a global prevalence of approximately 300 million individuals, presents a significant societal and economic burden. This multifaceted syndrome exhibits diverse clinical phenotypes and pathogenic endotypes influenced by various factors. The advent of omics technologies has revolutionized asthma research by delving into the molecular foundation of the disease to unravel its underlying mechanisms. Omics technologies are employed to systematically screen for potential biomarkers, encompassing genes, transcripts, methylation sites, proteins, and even the microbiome components. This review provides an insightful overview of omics applications in asthma research, with a special emphasis on genetics, transcriptomics, epigenomics, and the microbiome. We explore the cutting-edge methods, discoveries, challenges, and potential future directions in the realm of asthma omics research. By integrating multi-omics and non-omics data through advanced statistical techniques, we aspire to advance precision medicine in asthma, guiding diagnosis, risk assessment, and personalized treatment strategies for this heterogeneous condition.

Accounting for isoform expression increases power to identify genetic regulation of gene expression

A core problem in genetics is molecular quantitative trait locus (QTL) mapping, in which genetic variants associated with changes in the molecular phenotypes are identified. One of the most-studied molecular QTL mapping problems is expression QTL (eQTL) mapping, in which the molecular phenotype is gene expression. It is common in eQTL mapping to compute gene expression by aggregating the expression levels of individual isoforms from the same gene and then performing linear regression between SNPs and this aggregated gene expression level. However, SNPs may regulate isoforms from the same gene in different directions due to alternative splicing, or only regulate the expression level of one isoform, causing this approach to lose power. Here, we examine a broader question: which genes have at least one isoform whose expression level is regulated by genetic variants? In this study, we propose and evaluate several approaches to answering this question, demonstrating that "isoform-aware" methods-those that account for the expression levels of individual isoforms-have substantially greater power to answer this question than standard "gene-level" eQTL mapping methods. We identify settings in which different approaches yield an inflated number of false discoveries or lose power. In particular, we show that calling an eGene if there is a significant association between a SNP and any isoform fails to control False Discovery Rate, even when applying standard False Discovery Rate correction. We show that similar trends are observed in real data from the GEUVADIS and GTEx studies, suggesting the possibility that similar effects are present in these consortia.

Comparative transcriptome analysis of hepatopancreas reveals the potential mechanism of shrimp resistant to Vibrio parahaemolyticus infection

Vibrio parahaemolyticus carrying a pathogenic plasmid (VPAHPND) is one of the main causative agents of acute hepatopancreatic necrosis disease (AHPND) in shrimp aquaculture. Knowledge about the mechanism of shrimp resistant to VPAHPND is very helpful for developing efficient strategy for breeding AHPND resistant shrimp. In order to learn the mechanism of shrimp resistant to AHPND, comparative transcriptome was applied to analyze the different expressions of genes in the hepatopancreas of shrimp from different families with different resistance to VPAHPND. Through comparative analysis on the hepatopancreas of shrimp from VPAHPND resistant family and susceptible family, we found that differentially expressed genes (DEGs) were mainly involved in immune and metabolic processes. Most of the immune-related genes among DEGs were highly expressed in the hepatopancreas of shrimp from resistant family, involved in recognition of pathogen-associated molecular patterns, phagocytosis and elimination of pathogens, maintenance of reactive oxygen species homeostasis and other immune processes etc. However, most metabolic-related genes were highly expressed in the hepatopancreas of shrimp from susceptible family, involved in metabolism of lipid, vitamin, cofactors, glucose, carbohydrate and serine. Interestingly, when we analyzed the expression of above DEGs in the shrimp after VPAHPND infection, we found that the most of identified immune-related genes remained at high expression levels in the hepatopancreas of shrimp from the VPAHPND resistant family, and most of the identified metabolic-related genes were still at high expression levels in the hepatopancreas of shrimp from the VPAHPND susceptible family. The data suggested that the differential expression of these immune-related and metabolic-related genes in hepatopancreas might contribute to the resistance variations of shrimp to VPAHPND. These results provided valuable information for understanding the resistant mechanism of shrimp to VPAHPND.

Overexpression of REDUCED WALL ACETYLATION C increases xylan acetylation and biomass recalcitrance in Populus

Plant lignocellulosic biomass, i.e. secondary cell walls of plants, is a vital alternative source for bioenergy. However, the acetylation of xylan in secondary cell walls impedes the conversion of biomass to biofuels. Previous studies have shown that REDUCED WALL ACETYLATION (RWA) proteins are directly involved in the acetylation of xylan but the regulatory mechanism of RWAs is not fully understood. In this study, we demonstrate that overexpression of a Populus trichocarpa PtRWA-C gene increases the level of xylan acetylation and increases the lignin content and S/G ratio, ultimately yielding poplar woody biomass with reduced saccharification efficiency. Furthermore, through gene coexpression network and expression quantitative trait loci (eQTL) analysis, we found that PtRWA-C was regulated not only by the secondary cell wall hierarchical regulatory network but also by an AP2 family transcription factor HARDY (HRD). Specifically, HRD activates PtRWA-C expression by directly binding to the PtRWA-C promoter, which is also the cis-eQTL for PtRWA-C. Taken together, our findings provide insights into the functional roles of PtRWA-C in xylan acetylation and consequently saccharification and shed light on synthetic biology approaches to manipulate this gene and alter cell wall properties. These findings have substantial implications for genetic engineering of woody species, which could be used as a sustainable source of biofuels, valuable biochemicals, and biomaterials.

A Functional Schizophrenia-associated genetic variant near the TSNARE1 and ADGRB1 genes

Recent collaborative genome wide association studies (GWAS) have identified >200 independent loci contributing to risk for schizophrenia (SCZ). The genes closest to these loci have diverse functions, supporting the potential involvement of multiple relevant biological processes; yet there is no direct evidence that individual variants are functional or directly linked to specific genes. Nevertheless, overlap with certain epigenetic marks suggest that most GWAS-implicated variants are regulatory. Based on the strength of association with SCZ and the presence of regulatory epigenetic marks, we chose one such variant near TSNARE1 and ADGRB1, rs4129585, to test for functional potential and assay differences that may drive the pathogenicity of the risk allele. We observed that the variant-containing sequence drives reporter expression in relevant neuronal populations in zebrafish. Next, we introduced each allele into human induced pluripotent cells and differentiated 4 isogenic clones homozygous for the risk allele and 5 clones homozygous for the non-risk allele into neural precursor cells. Employing RNA-seq, we found that the two alleles yield significant transcriptional differences in the expression of 109 genes at FDR <0.05 and 259 genes at FDR <0.1. We demonstrate that these genes are highly interconnected in pathways enriched for synaptic proteins, axon guidance, and regulation of synapse assembly. Exploration of genes near rs4129585 suggests that this variant does not regulate TSNARE1 transcripts, as previously thought, but may regulate the neighboring ADGRB1, a regulator of synaptogenesis. Our results suggest that rs4129585 is a functional common variant that functions in specific pathways likely involved in SCZ risk.

Integrative approaches based on genomic techniques in the functional studies on enhancers

With the development of sequencing technology and the dramatic drop in sequencing cost, the functions of noncoding genes are being characterized in a wide variety of fields (e.g. biomedicine). Enhancers are noncoding DNA elements with vital transcription regulation functions. Tens of thousands of enhancers have been identified in the human genome; however, the location, function, target genes and regulatory mechanisms of most enhancers have not been elucidated thus far. As high-throughput sequencing techniques have leapt forwards, omics approaches have been extensively employed in enhancer research. Multidimensional genomic data integration enables the full exploration of the data and provides novel perspectives for screening, identification and characterization of the function and regulatory mechanisms of unknown enhancers. However, multidimensional genomic data are still difficult to integrate genome wide due to complex varieties, massive amounts, high rarity, etc. To facilitate the appropriate methods for studying enhancers with high efficacy, we delineate the principles, data processing modes and progress of various omics approaches to study enhancers and summarize the applications of traditional machine learning and deep learning in multi-omics integration in the enhancer field. In addition, the challenges encountered during the integration of multiple omics data are addressed. Overall, this review provides a comprehensive foundation for enhancer analysis.

DeepGAMI: deep biologically guided auxiliary learning for multimodal integration and imputation to improve genotype-phenotype prediction

BACKGROUND

Genotypes are strongly associated with disease phenotypes, particularly in brain disorders. However, the molecular and cellular mechanisms behind this association remain elusive. With emerging multimodal data for these mechanisms, machine learning methods can be applied for phenotype prediction at different scales, but due to the black-box nature of machine learning, integrating these modalities and interpreting biological mechanisms can be challenging. Additionally, the partial availability of these multimodal data presents a challenge in developing these predictive models.

METHOD

To address these challenges, we developed DeepGAMI, an interpretable neural network model to improve genotype-phenotype prediction from multimodal data. DeepGAMI leverages functional genomic information, such as eQTLs and gene regulation, to guide neural network connections. Additionally, it includes an auxiliary learning layer for cross-modal imputation allowing the imputation of latent features of missing modalities and thus predicting phenotypes from a single modality. Finally, DeepGAMI uses integrated gradient to prioritize multimodal features for various phenotypes.

RESULTS

We applied DeepGAMI to several multimodal datasets including genotype and bulk and cell-type gene expression data in brain diseases, and gene expression and electrophysiology data of mouse neuronal cells. Using cross-validation and independent validation, DeepGAMI outperformed existing methods for classifying disease types, and cellular and clinical phenotypes, even using single modalities (e.g., AUC score of 0.79 for Schizophrenia and 0.73 for cognitive impairment in Alzheimer's disease).

CONCLUSION

We demonstrated that DeepGAMI improves phenotype prediction and prioritizes phenotypic features and networks in multiple multimodal datasets in complex brains and brain diseases. Also, it prioritized disease-associated variants, genes, and regulatory networks linked to different phenotypes, providing novel insights into the interpretation of gene regulatory mechanisms. DeepGAMI is open-source and available for general use.

Mendelian Randomization Using the Druggable Genome Reveals Genetically Supported Drug Targets for Psychiatric Disorders

BACKGROUND AND HYPOTHESIS

Psychiatric disorders impose a huge health and economic burden on modern society. However, there is currently no proven completely effective treatment available, partly owing to the inefficiency of drug target identification and validation. We aim to identify therapeutic targets relevant to psychiatric disorders by conducting Mendelian randomization (MR) analysis.

STUDY DESIGN

We performed genome-wide MR analysis by integrating expression quantitative trait loci (eQTL) of 4479 actionable genes that encode druggable proteins and genetic summary statistics from genome-wide association studies of psychiatric disorders. After conducting colocalization analysis on the brain MR findings, we employed protein quantitative trait loci (pQTL) data as genetic proposed instruments for intersecting the colocalized genes to provide further genetic evidence.

STUDY RESULTS

By performing MR and colocalization analysis with eQTL genetic instruments, we obtained 31 promising drug targets for psychiatric disorders, including 21 significant genes for schizophrenia, 7 for bipolar disorder, 2 for depression, 1 for attention deficit and hyperactivity (ADHD) and none for autism spectrum disorder. Combining MR results using pQTL genetic instruments, we finally proposed 8 drug-targeting genes supported by the strongest MR evidence, including gene ACE, BTN3A3, HAPLN4, MAPK3 and NEK4 for schizophrenia, gene NEK4 and HAPLN4 for bipolar disorder, and gene TIE1 for ADHD.

CONCLUSIONS

Our findings with genetic support were more likely to be to succeed in clinical trials. In addition, our study prioritizes approved drug targets for the development of new therapies and provides critical drug reuse opportunities for psychiatric disorders.

Gene expression variation underlying tissue-specific responses to copper stress in Drosophila melanogaster

Copper is one of a handful of biologically necessary heavy metals that is also a common environmental pollutant. Under normal conditions, copper ions are required for many key physiological processes. However, in excess, copper quickly results in cell and tissue damage that can range in severity from temporary injury to permanent neurological damage. Because of its biological relevance, and because many conserved copper-responsive genes also respond to other non-essential heavy metal pollutants, copper resistance in Drosophila melanogaster is a useful model system with which to investigate the genetic control of the response to heavy metal stress. Because heavy metal toxicity has the potential to differently impact specific tissues, we genetically characterized the control of the gene expression response to copper stress in a tissue-specific manner in this study. We assessed the copper stress response in head and gut tissue of 96 inbred strains from the Drosophila Synthetic Population Resource (DSPR) using a combination of differential expression analysis and expression quantitative trait locus (eQTL) mapping. Differential expression analysis revealed clear patterns of tissue-specific expression, primarily driven by a more pronounced gene expression response in gut tissue. eQTL mapping of gene expression under control and copper conditions as well as for the change in gene expression following copper exposure (copper response eQTL) revealed hundreds of genes with tissue-specific local cis-eQTL and many distant trans-eQTL. eQTL associated with MtnA, Mdr49, Mdr50, and Sod3 exhibited genotype by environment effects on gene expression under copper stress, illuminating several tissue- and treatment-specific patterns of gene expression control. Together, our data build a nuanced description of the roles and interactions between allelic and expression variation in copper-responsive genes, provide valuable insight into the genomic architecture of susceptibility to metal toxicity, and highlight many candidate genes for future functional characterization.

Combined genome-wide association studies and expression quantitative trait locus analysis uncovers a genetic regulatory network of floral organ number in a tree peony (Paeonia suffruticosa Andrews) breeding population

Great progress has been made in our understanding of floral organ identity determination and its regulatory network in many species; however, the quantitative genetic basis of floral organ number variation is far less well understood for species-specific traits from the perspective of population variation. Here, using a tree peony (Paeonia suffruticosa Andrews, Paeoniaceae) cultivar population as a model, the phenotypic polymorphism and genetic variation based on genome-wide association studies (GWAS) and expression quantitative trait locus (eQTL) analysis were analyzed. Based on 24 phenotypic traits of 271 representative cultivars, the transcript profiles of 119 cultivars were obtained, which indicated abundant genetic variation in tree peony. In total, 86 GWAS-related cis-eQTLs and 3188 trans-eQTL gene pairs were found to be associated with the numbers of petals, stamens, and carpels. In addition, 19 floral organ number-related hub genes with 121 cis-eQTLs were obtained by weighted gene co-expression network analysis, among which five hub genes belonging to the ABCE genes of the MADS-box family and their spatial-temporal co-expression and regulatory network were constructed. These results not only help our understanding of the genetic basis of floral organ number variation during domestication, but also pave the way to studying the quantitative genetics and evolution of flower organ number and their regulatory network within populations.

Invited review: Good practices in genome-wide association studies to identify candidate sequence variants in dairy cattle

Genotype data from dairy cattle selection programs have greatly facilitated GWAS to identify variants related to economic traits. Results can enhance the accuracy of genomic prediction, analyze more complex models that go beyond additive effects, elucidate the genetic architecture of a trait, and finally, decipher the underlying biology of traits. The entire process, comprising data generation, quality control, statistical analyses, interpretation of association results, and linking results to biology should be designed and executed to minimize the generation of false-positive and false-negative associations and misleading links to biological processes. This review aims to provide general guidelines for data analysis that address data quality control, association tests, adjustment for population stratification, and significance evaluation to improve the reliability of conclusions. We also provide guidance on post-GWAS strategy and the interpretation of results. These guidelines are tailored to dairy cattle, which are characterized by long-range linkage disequilibrium, large half-sib families, and routinely collected phenotypes, requiring different approaches than those applied in human GWAS. We discuss common limitations and challenges that have been overlooked in the analysis and interpretation of GWAS to identify candidate sequence variants in dairy cattle.

Prioritization of genes associated with type 2 diabetes mellitus for functional studies

Existing therapies for type 2 diabetes mellitus (T2DM) show limited efficacy or have adverse effects. Numerous genetic variants associated with T2DM have been identified, but progress in translating these findings into potential drug targets has been limited. Here, we describe the tools and platforms available to identify effector genes from T2DM-associated coding and non-coding variants and prioritize them for functional studies. We discuss QSER1 and SLC12A8 as examples of genes that have been identified as possible T2DM candidate genes using these tools and platforms. We suggest further approaches, including the use of sequencing data with increased sample size and ethnic diversity, single-cell omics data for analyses, glycaemic trait associations to predict gene function and, potentially, human induced pluripotent stem cell 'village' cultures, to strengthen current gene functionalization workflows. Effective prioritization of T2DM-associated genes for experimental validation could expedite our understanding of the genetic mechanisms responsible for T2DM to facilitate the use of precision medicine in its treatment.

DeepCORE: An interpretable multi-view deep neural network model to detect co-operative regulatory elements

Gene transcription is an essential process involved in all aspects of cellular functions with significant impact on biological traits and diseases. This process is tightly regulated by multiple elements that co-operate to jointly modulate the transcription levels of target genes. To decipher the complicated regulatory network, we present a novel multi-view attention-based deep neural network that models the relationship between genetic, epigenetic, and transcriptional patterns and identifies co-operative regulatory elements (COREs). We applied this new method, named DeepCORE, to predict transcriptomes in 25 different cell lines, which outperformed the state-of-the-art algorithms. Furthermore, DeepCORE translates the attention values embedded in the neural network into interpretable information, including locations of putative regulatory elements and their correlations, which collectively implies COREs. These COREs are significantly enriched with known promoters and enhancers. Novel regulatory elements discovered by DeepCORE showed epigenetic signatures consistent with the status of histone modification marks.

Integrative Genomic Analysis of m6a-SNPs Identifies Potential Functional Variants Associated with Alzheimer's Disease

Alzheimer's disease (AD) is a neurodegenerative disorder that affects 35 million people worldwide. However, no potential therapeutics currently are available for AD because of the multiple factors involved in it, such as regulatory factors with their candidate genes, factors associated with the expression levels of its corresponding genes, and many others. To date, 29 novel loci from GWAS have been reported for AD by the Psychiatric Genomics Consortium (PGC2). Nevertheless, the main challenge of the post-GWAS era, namely to detect significant variants of the target disease, has not been conducted for AD. N6-methyladenosine (m6a) is reported as the most prevalent mRNA modification that exists in eukaryotes and that influences mRNA nuclear export, translation, splicing, and the stability of mRNA. Furthermore, studies have also reported m6a's association with neurogenesis and brain development. We carried out an integrative genomic analysis of AD variants from GWAS and m6a-SNPs from m6AVAR to identify the effects of m6a-SNPs on AD and identified the significant variants using the statistically significance value (p-value <0.05). The cis-regularity variants with their corresponding genes and their influence on gene expression in the gene expression profiles of AD patients were determined, and showed 1458 potential m6a-SNPs (based on p-value <0.05) associated with AD. eQTL analysis showed that 258 m6a-SNPs had cis-eQTL signals that overlapped with six significant differentially expressed genes based on p-value <0.05 in two datasets of AD gene expression profiles. A follow-up study to elucidate the impact of our identified m6a-SNPs in the experimental study would validate our findings for AD, which would contribute to the etiology of AD.

Candidate Genes and Gene Networks Change with Age in Japanese Black Cattle by Blood Transcriptome Analysis

Age is an important physiological factor that affects the metabolism and immune function of beef cattle. While there have been many studies using the blood transcriptome to study the effects of age on gene expression, few have been reported on beef cattle. To this end, we used the blood transcriptomes of Japanese black cattle at different ages as the study subjects and screened 1055, 345, and 1058 differential expressed genes (DEGs) in the calf vs. adult, adult vs. old, and calf vs. old comparison groups, respectively. The weighted co-expression network consisted of 1731 genes. Finally, blue, brown, and yellow age-specific modules were obtained, in which genes were enriched in signaling pathways related to growth and development and immune metabolic dysfunction, respectively. Protein-protein interaction (PPI) analysis showed gene interactions in each specific module, and 20 of the highest connectivity genes were chosen as potential hub genes. Finally, we identified 495, 244, and 1007 genes by exon-wide selection signature (EWSS) analysis of different comparison groups. Combining the results of hub genes, we found that VWF, PARVB, PRKCA, and TGFB1I1 could be used as candidate genes for growth and development stages of beef cattle. CORO2B and SDK1 could be used as candidate marker genes associated with aging. In conclusion, by comparing the blood transcriptome of calves, adult cattle, and old cattle, the candidate genes related to immunity and metabolism affected by age were identified, and the gene co-expression network of different age stages was constructed. It provides a data basis for exploring the growth, development, and aging of beef cattle.

Selective effects of a short transient environmental fluctuation on a natural population

Natural populations experience continuous and often transient changes of environmental conditions. These in turn may result in fluctuating selection pressures leading to variable demographic and evolutionary population responses. Rapid adaptation as short-term response to a sudden environmental change has in several cases been attributed to polygenic traits, but the underlying genomic dynamics and architecture are poorly understood. In this study, we took advantage of a natural experiment in an insect population of the non-biting midge Chironomus riparius by monitoring genome-wide allele frequencies before and after a cold snap event. Whole genome pooled sequencing of time series samples revealed 10 selected haplotypes carrying ancient polymorphisms, partially with signatures of balancing selection. By constantly cold exposing genetically variable individuals in the laboratory, we could demonstrate with whole genome resequencing (i) that among the survivors, the same alleles rose in frequency as in the wild, and (ii) that the identified variants additively predicted fitness (survival time) of its bearers. Finally, by simultaneously sequencing the genome and the transcriptome of cold exposed individuals we could tentatively link some of the selected SNPs to the cis- and trans-regulation of genes and pathways known to be involved in cold response of insects, such as cytochrome P450 and fatty acid metabolism. Altogether, our results shed light on the strength and speed of selection in natural populations and the genomic architecture of its underlying polygenic trait. Population genomic time series data thus appear as promising tool for measuring the selective tracking of fluctuating selection in natural populations.

Identification of transcriptional regulatory variants in pig duodenum, liver, and muscle tissues

BACKGROUND

In humans and livestock species, genome-wide association studies (GWAS) have been applied to study the association between variants distributed across the genome and a phenotype of interest. To discover genetic polymorphisms affecting the duodenum, liver, and muscle transcriptomes of 300 pigs from 3 different breeds (Duroc, Landrace, and Large White), we performed expression GWAS between 25,315,878 polymorphisms and the expression of 13,891 genes in duodenum, 12,748 genes in liver, and 11,617 genes in muscle.

RESULTS

More than 9.68 × 1011 association tests were performed, yielding 14,096,080 significantly associated variants, which were grouped in 26,414 expression quantitative trait locus (eQTL) regions. Over 56% of the variants were within 1 Mb of their associated gene. In addition to the 100-kb region upstream of the transcription start site, we identified the importance of the 100-kb region downstream of the 3'UTR for gene regulation, as most of the cis-regulatory variants were located within these 2 regions. We also observed 39,874 hotspot regulatory polymorphisms associated with the expression of 10 or more genes that could modify the protein structure or the expression of a regulator gene. In addition, 2 motifs (5'-GATCCNGYGTTGCYG-3' and a poly(A) sequence) were enriched across the 3 tissues within the neighboring sequences of the most significant single-nucleotide polymorphisms in each cis-eQTL region.

CONCLUSIONS

The 14 million significant associations obtained in this study are publicly available and have enabled the identification of expression-associated cis-, trans-, and hotspot regulatory variants within and across tissues, thus shedding light on the molecular mechanisms of regulatory variations that shape end-trait phenotypes.

Integrating extrusion complex-associated pattern to predict cell type-specific long-range chromatin loops

The chromatin loop plays a critical role in the study of gene expression and disease. Supervised learning-based algorithms to predict the chromatin loops require large priori information to satisfy the model construction, while the prediction sensitivity of unsupervised learning-based algorithms is still unsatisfactory. Therefore, we propose an unsupervised algorithm, Ecomap-loop. It takes advantage of extrusion complex-associated patterns, including CTCF, RAD21, and SMC enrichments, as well as the orientation distribution of CTCF motif of loops to build feature matrices; then the eigen decomposition model is employed to obtain the cell type-specific loops. We compare the performance of Ecomap-loop with the state-of-the-art unsupervised algorithm using Hi-C, ChIA-PET, expression quantitative trait locus (eQTL), and CRISPR interference (CRISPRi) screen data; the results show that Ecomap-loop achieves the best in four cell types. In addition, the functional analysis reveals the ability of Ecomap-loop to predict active functionality-related and cell type-specific loops.

Detection of potential functional variants based on systems-biology: the case of feed efficiency in beef cattle

BACKGROUND

Potential functional variants (PFVs) can be defined as genetic variants responsible for a given phenotype. Ultimately, these are the best DNA markers for animal breeding and selection, especially for polygenic and complex phenotypes. Herein, we described the identification of PFVs for complex phenotypes (in this case, Feed Efficiency in beef cattle) using a systems-biology driven approach based on RNA-seq data from physiologically relevant organs.

RESULTS

The systems-biology coupled with deep molecular phenotyping by RNA-seq of liver, muscle, hypothalamus, pituitary, and adrenal glands of animals with high and low feed efficiency (FE) measured by residual feed intake (RFI) identified 2,000,936 uniquely variants. Among them, 9986 variants were significantly associated with FE and only 78 had a high impact on protein expression and were considered as PFVs. A set of 169 significant uniquely variants were expressed in all five organs, however, only 27 variants had a moderate impact and none of them a had high impact on protein expression. These results provide evidence of tissue-specific effects of high-impact PFVs. The PFVs were enriched (FDR < 0.05) for processing and presentation of MHC Class I and II mediated antigens, which are an important part of the adaptive immune response. The experimental validation of these PFVs was demonstrated by the increased prediction accuracy for RFI using the weighted G matrix (ssGBLUP+wG; Acc = 0.10 and b = 0.48) obtained in the ssGWAS in comparison to the unweighted G matrix (ssGBLUP; Acc = 0.29 and b = 1.10).

CONCLUSION

Here we identified PFVs for FE in beef cattle using a strategy based on systems-biology and deep molecular phenotyping. This approach has great potential to be used in genetic prediction programs, especially for polygenic phenotypes.

Network reconstruction for trans acting genetic loci using multi-omics data and prior information

BACKGROUND

Molecular measurements of the genome, the transcriptome, and the epigenome, often termed multi-omics data, provide an in-depth view on biological systems and their integration is crucial for gaining insights in complex regulatory processes. These data can be used to explain disease related genetic variants by linking them to intermediate molecular traits (quantitative trait loci, QTL). Molecular networks regulating cellular processes leave footprints in QTL results as so-called trans-QTL hotspots. Reconstructing these networks is a complex endeavor and use of biological prior information can improve network inference. However, previous efforts were limited in the types of priors used or have only been applied to model systems. In this study, we reconstruct the regulatory networks underlying trans-QTL hotspots using human cohort data and data-driven prior information.

METHODS

We devised a new strategy to integrate QTL with human population scale multi-omics data. State-of-the art network inference methods including BDgraph and glasso were applied to these data. Comprehensive prior information to guide network inference was manually curated from large-scale biological databases. The inference approach was extensively benchmarked using simulated data and cross-cohort replication analyses. Best performing methods were subsequently applied to real-world human cohort data.

RESULTS

Our benchmarks showed that prior-based strategies outperform methods without prior information in simulated data and show better replication across datasets. Application of our approach to human cohort data highlighted two novel regulatory networks related to schizophrenia and lean body mass for which we generated novel functional hypotheses.

CONCLUSIONS

We demonstrate that existing biological knowledge can improve the integrative analysis of networks underlying trans associations and generate novel hypotheses about regulatory mechanisms.

Population-level transcriptomes reveal gene expression and splicing underlying cadmium accumulation in barley

Genetic variation is an important determinant of gene transcription, which in turn contributes to functional and phenotypic diversity. Identification of the genetic variants controlling gene expression and alternative splicing in crops responding to cadmium (Cd), an important issue for food safety and human health, is of great value to improve our understanding of Cd accumulation-related genes. Here we report an in-depth survey of population-level transcriptome variation of barley (Hordeum vulgare) core accessions under Cd exposure. We reveal marked transcriptomic changes in response to Cd exposure, and these are largely independent of tissues. A genome-wide association study (GWAS) revealed 59 498 expression quantitative trait loci (eQTLs) and 23 854 splicing quantitative trait loci (sQTLs), leading to a complex network that covers 66.6% of the expressed genes, including 68 metal transporter genes. On average, 41.0% of sQTLs overlapped with eQTLs across different tissues, indicating that these two dimensions of transcript variation are largely independent. Moreover, we found that 34.5% of GWAS QTLs that underlie 10 Cd accumulation traits in barley are co-localized with eQTLs and sQTLs, which could imply a mechanistic role of different genetic variants affecting gene expression and alternative splicing in these traits. This study highlights the role of distal and proximal genetic effects on gene expression, splicing, and phenotypic plasticity. We anticipate that our results on the genetic control of expression and splicing underlying Cd accumulation provide a bridge to better understand genetic variation and phenotypic diversity to elucidate the mechanisms underlying Cd accumulation in plants.

Mathematical model for the relationship between single-cell and bulk gene expression to clarify the interpretation of bulk gene expression data

BACKGROUND

Differential expression analysis is a standard approach in molecular biology. For example, genes whose expression levels differ between diseased and non-diseased samples are considered to be associated with that disease. On the other hand, differential variability analysis focuses on the differences of the variances of gene expression between sample groups. Although differential variability is also known to capture biological information, its interpretation remains unclear and controversial. Recent single-cell analyses have revealed that differences between sample groups can affect gene expression in a cellular subset-specific manner or by altering the proportion of a particular cellular subset. The aim of this study is to clarify the interpretation of mean and variance of bulk gene expression data.

METHOD

We developed a mathematical model in which the bulk gene expression value is proportional to the mean value of the single-cell gene expression profile. Based on this model, we performed theoretical, simulated and real single-cell RNA-seq data analyses.

RESULT AND CONCLUSION

We identified how differences in single-cell gene expression profiles affect the differences in the mean and the variance of bulk gene expression. It is shown that differential expression analysis of bulk expression data can overlook significant changes in gene expression at the single-cell level. Further, differential variability analysis capture the complex feature affected by different gene expression shifts for each subset, changes in the proportions of cellular subsets, and variation in single-cell distribution parameters among samples.

Interpretable generative deep learning: an illustration with single cell gene expression data

Deep generative models can learn the underlying structure, such as pathways or gene programs, from omics data. We provide an introduction as well as an overview of such techniques, specifically illustrating their use with single-cell gene expression data. For example, the low dimensional latent representations offered by various approaches, such as variational auto-encoders, are useful to get a better understanding of the relations between observed gene expressions and experimental factors or phenotypes. Furthermore, by providing a generative model for the latent and observed variables, deep generative models can generate synthetic observations, which allow us to assess the uncertainty in the learned representations. While deep generative models are useful to learn the structure of high-dimensional omics data by efficiently capturing non-linear dependencies between genes, they are sometimes difficult to interpret due to their neural network building blocks. More precisely, to understand the relationship between learned latent variables and observed variables, e.g., gene transcript abundances and external phenotypes, is difficult. Therefore, we also illustrate current approaches that allow us to infer the relationship between learned latent variables and observed variables as well as external phenotypes. Thereby, we render deep learning approaches more interpretable. In an application with single-cell gene expression data, we demonstrate the utility of the discussed methods.

eQTL analysis of laying hens divergently selected for feather pecking identifies KLF14 as a potential key regulator for this behavioral disorder

Feather pecking in chickens is a damaging behavior, seriously impacting animal welfare and leading to economic losses. Feather pecking is a complex trait, which is partly under genetic control. Different hypotheses have been proposed to explain the etiology of feather pecking and notably, several studies have identified similarities between feather pecking and human mental disorders such as obsessive-compulsive disorder and schizophrenia. This study uses transcriptomic and phenotypic data from 167 chickens to map expression quantitative trait loci and to identify regulatory genes with a significant effect on this behavioral disorder using an association weight matrix approach. From 70 of the analyzed differentially expressed genes, 11,790 genome wide significantly associated variants were detected, of which 23 showed multiple associations (≥15). These were located in proximity to a number of genes, which are transcription regulators involved in chromatin binding, nucleic acid metabolism, protein translation and putative regulatory RNAs. The association weight matrix identified 36 genes and the two transcription factors: SP6 (synonym: KLF14) and ENSGALG00000042129 (synonym: CHTOP) as the most significant, with an enrichment of KLF14 binding sites being detectable in 40 differentially expressed genes. This indicates that differential expression between animals showing high and low levels of feather pecking was significantly associated with a genetic variant in proximity to KLF14. This multiallelic variant was located 652 bp downstream of KLF14 and is a deletion of 1-3 bp. We propose that a deletion downstream of the transcription factor KLF14 has a negative impact on the level of T cells in the developing brain of high feather pecking chickens, which leads to developmental and behavioral abnormalities. The lack of CD4 T cells and gamma-Aminobutyric acid (GABA) receptors are important factors for the increased propensity of laying hens to perform feather pecking. As such, KLF14 is a clear candidate regulator for the expression of genes involved in the pathogenic development. By further elucidating the regulatory pathways involved in feather pecking we hope to take significant steps forward in explaining and understanding other mental disorders, not just in chickens.

Identifying cancer tissue-of-origin by a novel machine learning method based on expression quantitative trait loci

Cancer of unknown primary (CUP) refers to cancer with primary lesion unidentifiable by regular pathological and clinical diagnostic methods. This kind of cancer is extremely difficult to treat, and patients with CUP usually have a very short survival time. Recent studies have suggested that cancer treatment targeting primary lesion will significantly improve the survival of CUP patients. Thus, it is critical to develop accurate yet fast methods to infer the tissue-of-origin (TOO) of CUP. In the past years, there are a few computational methods to infer TOO based on single omics data like gene expression, methylation, somatic mutation, and so on. However, the metastasis of tumor involves the interaction of multiple levels of biological molecules. In this study, we developed a novel computational method to predict TOO of CUP patients by explicitly integrating expression quantitative trait loci (eQTL) into an XGBoost classification model. We trained our model with The Cancer Genome Atlas (TCGA) data involving over 7,000 samples across 20 types of solid tumors. In the 10-fold cross-validation, the prediction accuracy of the model with eQTL was over 0.96, better than that without eQTL. In addition, we also tested our model in an independent data downloaded from Gene Expression Omnibus (GEO) consisting of 87 samples across 4 cancer types. The model also achieved an f1-score of 0.7-1 depending on different cancer types. In summary, eQTL was an important information in inferring cancer TOO and the model might be applied in clinical routine test for CUP patients in the future.

eQTLs are key players in the integration of genomic and transcriptomic data for phenotype prediction

Background

Multi-omics represent a promising link between phenotypes and genome variation. Few studies yet address their integration to understand genetic architecture and improve predictability.

Results

Our study used 241 poplar genotypes, phenotyped in two common gardens, with xylem and cambium RNA sequenced at one site, yielding large phenotypic, genomic (SNP), and transcriptomic datasets. Prediction models for each trait were built separately for SNPs and transcripts, and compared to a third model integrated by concatenation of both omics. The advantage of integration varied across traits and, to understand such differences, an eQTL analysis was performed to characterize the interplay between the genome and transcriptome and classify the predicting features into cis or trans relationships. A strong, significant negative correlation was found between the change in predictability and the change in predictor ranking for trans eQTLs for traits evaluated in the site of transcriptomic sampling.

Conclusions

Consequently, beneficial integration happens when the redundancy of predictors is decreased, likely leaving the stage to other less prominent but complementary predictors. An additional gene ontology (GO) enrichment analysis appeared to corroborate such statistical output. To our knowledge, this is a novel finding delineating a promising method to explore data integration.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-022-08690-7.

WGCNA Analysis Identifies the Hub Genes Related to Heat Stress in Seedling of Rice (Oryza sativa L.)

Frequent high temperature weather affects the growth and development of rice, resulting in the decline of seed–setting rate, deterioration of rice quality and reduction of yield. Although some high temperature tolerance genes have been cloned, there is still little success in solving the effects of high temperature stress in rice (Oryza sativa L.). Based on the transcriptional data of seven time points, the weighted correlation network analysis (WGCNA) method was used to construct a co–expression network of differentially expressed genes (DEGs) between the rice genotypes IR64 (tolerant to heat stress) and Koshihikari (susceptible to heat stress). There were four modules in both genotypes that were highly correlated with the time points after heat stress in the seedling. We further identified candidate hub genes through clustering and analysis of protein interaction network with known–core genes. The results showed that the ribosome and protein processing in the endoplasmic reticulum were the common pathways in response to heat stress between the two genotypes. The changes of starch and sucrose metabolism and the biosynthesis of secondary metabolites pathways are possible reasons for the sensitivity to heat stress for Koshihikari. Our findings provide an important reference for the understanding of high temperature response mechanisms and the cultivation of high temperature resistant materials.

Transcriptomic analysis provides insights into molecular mechanisms of thermal physiology

Physiological trait variation underlies health, responses to global climate change, and ecological performance. Yet, most physiological traits are complex, and we have little understanding of the genes and genomic architectures that define their variation. To provide insight into the genetic architecture of physiological processes, we related physiological traits to heart and brain mRNA expression using a weighted gene co-expression network analysis. mRNA expression was used to explain variation in six physiological traits (whole animal metabolism (WAM), critical thermal maximum (CT_max), and four substrate specific cardiac metabolic rates (CaM)) under 12 °C and 28 °C acclimation conditions. Notably, the physiological trait variations among the three geographically close (within 15 km) and genetically similar F. heteroclitus populations are similar to those found among 77 aquatic species spanning 15–20° of latitude (~ 2,000 km). These large physiological trait variations among genetically similar individuals provide a powerful approach to determine the relationship between mRNA expression and heritable fitness related traits unconfounded by interspecific differences. Expression patterns explained up to 82% of metabolic trait variation and were enriched for multiple signaling pathways known to impact metabolic and thermal tolerance (e.g., AMPK, PPAR, mTOR, FoxO, and MAPK) but also contained several unexpected pathways (e.g., apoptosis, cellular senescence), suggesting that physiological trait variation is affected by many diverse genes.

Powerful and robust inference of complex phenotypes' causal genes with dependent expression quantitative loci by a median-based Mendelian randomization

Isolating the causal genes from numerous genetic association signals in genome-wide association studies (GWASs) of complex phenotypes remains an open and challenging question. In the present study, we proposed a statistical approach, the effective-median-based Mendelian randomization (MR) framework, for inferring the causal genes of complex phenotypes with the GWAS summary statistics (named EMIC). The effective-median method solved the high false-positive issue in the existing MR methods due to either correlation among instrumental variables or noises in approximated linkage disequilibrium (LD). EMIC can further perform a pleiotropy fine-mapping analysis to remove possible false-positive estimates. With the usage of multiple cis-expression quantitative trait loci (eQTLs), EMIC was also more powerful than the alternative methods for the causal gene inference in the simulated datasets. Furthermore, EMIC rediscovered many known causal genes of complex phenotypes (schizophrenia, bipolar disorder, and total cholesterol) and reported many new and promising candidate causal genes. In sum, this study provided an efficient solution to discriminate the candidate causal genes from vast amounts of GWAS signals with eQTLs. EMIC has been implemented in our integrative software platform KGGSEE.

Mitochondria as environments for the nuclear genome in Drosophila: mitonuclear G×G×E

Mitochondria evolved from a union of microbial cells belonging to distinct lineages that were likely anaerobic. The evolution of eukaryotes required a massive reorganization of the 2 genomes and eventual adaptation to aerobic environments. The nutrients and oxygen that sustain eukaryotic metabolism today are processed in mitochondria through coordinated expression of 37 mitochondrial genes and over 1000 nuclear genes. This puts mitochondria at the nexus of gene-by-gene (G×G) and gene-by-environment (G×E) interactions that sustain life. Here we use a Drosophila model of mitonuclear genetic interactions to explore the notion that mitochondria are environments for the nuclear genome, and vice versa. We construct factorial combinations of mtDNA and nuclear chromosomes to test for epistatic interactions (G×G), and expose these mitonuclear genotypes to altered dietary environments to examine G×E interactions. We use development time and genome-wide RNAseq analyses to assess the relative contributions of mtDNA, nuclear chromosomes, and environmental effects on these traits (mitonuclear G×G×E). We show that the nuclear transcriptional response to alternative mitochondrial "environments" (G×G) has significant overlap with the transcriptional response of mitonuclear genotypes to altered dietary environments. These analyses point to specific transcription factors (e.g., giant) that mediated these interactions, and identified coexpressed modules of genes that may account for the overlap in differentially expressed genes. Roughly 20% of the transcriptome includes G×G genes that are concordant with G×E genes, suggesting that mitonuclear interactions are part of an organism's environment.

Comparison of Gene Expression Between Resistant and Susceptible Families Against VPAHPND and Identification of Biomarkers Used for Resistance Evaluation in Litopenaeus vannamei

Acute hepatopancreatic necrosis disease (AHPND) has caused a heavy loss to shrimp aquaculture since its outbreak. Vibrio parahaemolyticus (VP_AHPND) is regarded as one of the main pathogens that caused AHPND in the Pacific white shrimp Litopenaeus vannamei. In order to learn more about the mechanism of resistance to AHPND, the resistant and susceptible shrimp families were obtained through genetic breeding, and comparative transcriptome approach was used to analyze the gene expression patterns between resistant and susceptible families. A total of 95 families were subjected to VP_AHPND challenge test, and significant variations in the resistance of these families were observed. Three pairs of resistant and susceptible families were selected for transcriptome sequencing. A total of 489 differentially expressed genes (DEGs) that presented in at least two pairwise comparisons were screened, including 196 DEGs highly expressed in the susceptible families and 293 DEGs in the resistant families. Among these DEGs, 16 genes demonstrated significant difference in all three pairwise comparisons. Gene set enrichment analysis (GSEA) of all 27,331 expressed genes indicated that some energy metabolism processes were enriched in the resistant families, while signal transduction and immune system were enriched in the susceptible families. A total of 32 DEGs were further confirmed in the offspring of the detected families, among which 19 genes were successfully verified. The identified genes in this study will be useful for clarifying the genetic mechanism of shrimp resistance against Vibrio and will further provide molecular markers for evaluating the disease resistance of shrimp in the breeding program.

Selection on Accessible Chromatin Regions in Capsella grandiflora

Accurate estimates of genome-wide rates and fitness effects of new mutations are essential for an improved understanding of molecular evolutionary processes. Although eukaryotic genomes generally contain a large noncoding fraction, functional noncoding regions and fitness effects of mutations in such regions are still incompletely characterized. A promising approach to characterize functional noncoding regions relies on identifying accessible chromatin regions (ACRs) tightly associated with regulatory DNA. Here, we applied this approach to identify and estimate selection on ACRs in Capsella grandiflora, a crucifer species ideal for population genomic quantification of selection due to its favorable population demography. We describe a population-wide ACR distribution based on ATAC-seq data for leaf samples of 16 individuals from a natural population. We use population genomic methods to estimate fitness effects and proportions of positively selected fixations (α) in ACRs and find that intergenic ACRs harbor a considerable fraction of weakly deleterious new mutations, as well as a significantly higher proportion of strongly deleterious mutations than comparable inaccessible intergenic regions. ACRs are enriched for expression quantitative trait loci (eQTL) and depleted of transposable element insertions, as expected if intergenic ACRs are under selection because they harbor regulatory regions. By integrating empirical identification of intergenic ACRs with analyses of eQTL and population genomic analyses of selection, we demonstrate that intergenic regulatory regions are an important source of nearly neutral mutations. These results improve our understanding of selection on noncoding regions and the role of nearly neutral mutations for evolutionary processes in outcrossing Brassicaceae species.

Systematic Exploration in Tissue-Pathway Associations of Complex Traits Using Comprehensive eQTLs Catalog

The collection of expression quantitative trait loci (eQTLs) is an important resource to study complex traits through understanding where and how transcriptional regulations are controlled by genetic variations in the non-coding regions of the genome. Previous studies have focused on associating eQTLs with traits to identify the roles of trait-related eQTLs and their corresponding target genes involved in trait determination. Since most genes function as a part of pathways in a systematic manner, it is crucial to explore the pathways’ involvements in complex traits to test potentially novel hypotheses and to reveal underlying mechanisms of disease pathogenesis. In this study, we expanded and applied loci2path software to perform large-scale eQTLs enrichment [i.e., eQTLs’ target genes (eGenes) enrichment] analysis at pathway level to identify the tissue-specific enriched pathways within trait-related genomic intervals. By utilizing 13,791,909 eQTLs cataloged in the Genotype-Tissue Expression (GTEx) V8 data for 49 tissue types, 2,893 pathway sets reported from MSigDB, and query regions derived from the Phenotype-Genotype Integrator (PheGenI) catalog, we identified intriguing biological pathways that are likely to be involved in ten traits [Alzheimer’s disease (AD), body mass index, Parkinson’s disease (PD), schizophrenia, amyotrophic lateral sclerosis, non-small cell lung cancer (NSCLC), stroke, blood pressure, autism spectrum disorder, and myocardial infarction]. Furthermore, we extracted the most significant pathways for AD, such as BioCarta D4-GDI pathway and WikiPathways sulfation biotransformation reaction and viral acute myocarditis pathways, to study specific genes within pathways. Our data presented new hypotheses in AD pathogenesis supported by previous studies, like the increased level of caspase-3 in the amygdala that cleaves GDP dissociation inhibitor and binds to beta-amyloid, leading to increased apoptosis and neuronal loss. Our findings also revealed potential pathogenesis mechanisms for PD, schizophrenia, NSCLC, blood pressure, autism spectrum disorder, and myocardial infarction, which were consistent with past studies. Our results indicated that loci2path′s eQTLs enrichment test was valuable in unveiling novel biological mechanisms of complex traits. The discovered mechanisms of disease pathogenesis and traits require further in-depth analysis and experimental validation.

Inheritance of gene expression throughout fruit development in chili pepper

Gene expression is the primary molecular phenotype and can be estimated in specific organs or tissues at particular times. Here we analyzed genome-wide inheritance of gene expression in fruits of chili pepper (Capsicum annuum L.) in reciprocal crosses between a domesticated and a wild accession, estimating this parameter during fruit development. We defined a general hierarchical schema to classify gene expression inheritance which can be employed for any quantitative trait. We found that inheritance of gene expression is affected by both, the time of fruit development as well as the direction of the cross, and propose that such variations could be common in many developmental processes. We conclude that classification of inheritance patterns is important to have a better understanding of the mechanisms underlying gene expression regulation, and demonstrate that sets of genes with specific inheritance pattern at particular times of fruit development are enriched in different biological processes, molecular functions and cell components. All curated data and functions for analysis and visualization are publicly available as an R package.

Integrative-omics for discovery of network-level disease biomarkers: a case study in Alzheimer's disease

A large number of genetic variations have been identified to be associated with Alzheimer's disease (AD) and related quantitative traits. However, majority of existing studies focused on single types of omics data, lacking the power of generating a community including multi-omic markers and their functional connections. Because of this, the immense value of multi-omics data on AD has attracted much attention. Leveraging genomic, transcriptomic and proteomic data, and their backbone network through functional relations, we proposed a modularity-constrained logistic regression model to mine the association between disease status and a group of functionally connected multi-omic features, i.e. single-nucleotide polymorphisms (SNPs), genes and proteins. This new model was applied to the real data collected from the frontal cortex tissue in the Religious Orders Study and Memory and Aging Project cohort. Compared with other state-of-art methods, it provided overall the best prediction performance during cross-validation. This new method helped identify a group of densely connected SNPs, genes and proteins predictive of AD status. These SNPs are mostly expression quantitative trait loci in the frontal region. Brain-wide gene expression profile of these genes and proteins were highly correlated with the brain activation map of 'vision', a brain function partly controlled by frontal cortex. These genes and proteins were also found to be associated with the amyloid deposition, cortical volume and average thickness of frontal regions. Taken together, these results suggested a potential pathway underlying the development of AD from SNPs to gene expression, protein expression and ultimately brain functional and structural changes.

Privacy-preserving genotype imputation in a trusted execution environment

Genotype imputation is an essential tool in genomics research, whereby missing genotypes are inferred using reference genomes to enhance downstream analyses. Recently, public imputation servers have allowed researchers to leverage large-scale genomic data resources for imputation. However, privacy concerns about uploading one's genetic data to a server limit the utility of these services. We introduce a secure hardware-based solution for privacy-preserving genotype imputation, which keeps the input genomes private by processing them within Intel SGX's trusted execution environment. Our solution features SMac, an efficient and secure imputation algorithm designed for Intel SGX, which employs a state-of-the-art imputation strategy also utilized by existing imputation servers. SMac achieves imputation accuracy equivalent to existing tools and provides protection against known side-channel attacks on SGX while maintaining scalability. We also show the necessity of our enhanced security by identifying vulnerabilities in existing imputation software. Our work represents a step toward privacy-preserving genomic analysis services.

Leveraging three-dimensional chromatin architecture for effective reconstruction of enhancer-target gene regulatory interactions

A growing amount of evidence in literature suggests that germline sequence variants and somatic mutations in non-coding distal regulatory elements may be crucial for defining disease risk and prognostic stratification of patients, in genetic disorders as well as in cancer. Their functional interpretation is challenging because genome-wide enhancer-target gene (ETG) pairing is an open problem in genomics. The solutions proposed so far do not account for the hierarchy of structural domains which define chromatin three-dimensional (3D) architecture. Here we introduce a change of perspective based on the definition of multi-scale structural chromatin domains, integrated in a statistical framework to define ETG pairs. In this work (i) we develop a computational and statistical framework to reconstruct a comprehensive map of ETG pairs leveraging functional genomics data; (ii) we demonstrate that the incorporation of chromatin 3D architecture information improves ETG pairing accuracy and (iii) we use multiple experimental datasets to extensively benchmark our method against previous solutions for the genome-wide reconstruction of ETG pairs. This solution will facilitate the annotation and interpretation of sequence variants in distal non-coding regulatory elements. We expect this to be especially helpful in clinically oriented applications of whole genome sequencing in cancer and undiagnosed genetic diseases research.

Associations between pancreatic expression quantitative traits and risk of pancreatic ductal adenocarcinoma

Pancreatic ductal adenocarcinoma (PDAC) is among the most lethal cancers. Its poor prognosis is predominantly due to the fact that most patients remain asymptomatic until the disease reaches an advanced stage, alongside the lack of early markers and screening strategies. A better understanding of PDAC risk factors is essential for the identification of groups at high risk in the population. Genome-wide association studies (GWAS) have been a powerful tool for detecting genetic variants associated with complex traits, including pancreatic cancer. By exploiting functional and GWAS data, we investigated the associations between polymorphisms affecting gene function in the pancreas (expression quantitative trait loci, eQTLs) and PDAC risk. In a two-phase approach, we analysed 13 713 PDAC cases and 43 784 controls and identified a genome-wide significant association between the A allele of the rs2035875 polymorphism and increased PDAC risk (P = 7.14 × 10-10). This allele is known to be associated with increased expression in the pancreas of the keratin genes KRT8 and KRT18, whose increased levels have been reported to correlate with various tumour cell characteristics. Additionally, the A allele of the rs789744 variant was associated with decreased risk of developing PDAC (P = 3.56 × 10-6). This single nucleotide polymorphism is situated in the SRGAP1 gene and the A allele is associated with higher expression of the gene, which in turn inactivates the cyclin-dependent protein 42 (CDC42) gene expression, thus decreasing the risk of PDAC. In conclusion, we present here a functional-based novel PDAC risk locus and an additional strong candidate supported by significant associations and plausible biological mechanisms.