Analyzing illumina gene expression microarray data from different tissues: methodological aspects of data analysis in the metaxpress consortium

Interpretable multimodal machine learning (IMML) framework reveals pathological signatures of distal sensorimotor polyneuropathy

BACKGROUND

Distal sensorimotor polyneuropathy (DSPN) is a common neurological disorder in elderly adults and people with obesity, prediabetes and diabetes and is associated with high morbidity and premature mortality. DSPN is a multifactorial disease and not fully understood yet.

METHODS

Here, we developed the Interpretable Multimodal Machine Learning (IMML) framework for predicting DSPN prevalence and incidence based on sparse multimodal data. Exploiting IMMLs interpretability further empowered biomarker identification. We leveraged the population-based KORA F4/FF4 cohort including 1091 participants and their deep multimodal characterisation, i.e. clinical data, genomics, methylomics, transcriptomics, proteomics, inflammatory proteins and metabolomics.

RESULTS

Clinical data alone is sufficient to stratify individuals with and without DSPN (AUROC = 0.752), whilst predicting DSPN incidence 6.5 ± 0.2 years later strongly benefits from clinical data complemented with two or more molecular modalities (improved ΔAUROC > 0.1, achieved AUROC of 0.714). Important and interpretable features of incident DSPN prediction include up-regulation of proinflammatory cytokines, down-regulation of SUMOylation pathway and essential fatty acids, thus yielding novel insights in the disease pathophysiology.

CONCLUSIONS

These may become biomarkers for incident DSPN, guide prevention strategies and serve as proof of concept for the utility of IMML in studying complex diseases.

RIOK2 transcriptionally regulates TRiC and dyskerin complexes to prevent telomere shortening

Telomere shortening is a prominent hallmark of aging and is emerging as a characteristic feature of Myelodysplastic Syndromes (MDS) and Idiopathic Pulmonary Fibrosis (IPF). Optimal telomerase activity prevents progressive shortening of telomeres that triggers DNA damage responses. However, the upstream regulation of telomerase holoenzyme components remains poorly defined. Here, we identify RIOK2, a master regulator of human blood cell development, as a critical transcription factor for telomere maintenance. Mechanistically, loss of RIOK2 or its DNA-binding/transactivation properties downregulates mRNA expression of both TRiC and dyskerin complex subunits that impairs telomerase activity, thereby causing telomere shortening. We further show that RIOK2 expression is diminished in aged individuals and IPF patients, and it strongly correlates with shortened telomeres in MDS patient-derived bone marrow cells. Importantly, ectopic expression of RIOK2 alleviates telomere shortening in IPF patient-derived primary lung fibroblasts. Hence, increasing RIOK2 levels prevents telomere shortening, thus offering therapeutic strategies for telomere biology disorders.

Distinct genetic liability profiles define clinically relevant patient strata across common diseases

Stratified medicine holds great promise to tailor treatment to the needs of individual patients. While genetics holds great potential to aid patient stratification, it remains a major challenge to operationalize complex genetic risk factor profiles to deconstruct clinical heterogeneity. Contemporary approaches to this problem rely on polygenic risk scores (PRS), which provide only limited clinical utility and lack a clear biological foundation. To overcome these limitations, we develop the CASTom-iGEx approach to stratify individuals based on the aggregated impact of their genetic risk factor profiles on tissue specific gene expression levels. The paradigmatic application of this approach to coronary artery disease or schizophrenia patient cohorts identified diverse strata or biotypes. These biotypes are characterized by distinct endophenotype profiles as well as clinical parameters and are fundamentally distinct from PRS based groupings. In stark contrast to the latter, the CASTom-iGEx strategy discovers biologically meaningful and clinically actionable patient subgroups, where complex genetic liabilities are not randomly distributed across individuals but rather converge onto distinct disease relevant biological processes. These results support the notion of different patient biotypes characterized by partially distinct pathomechanisms. Thus, the universally applicable approach presented here has the potential to constitute an important component of future personalized medicine paradigms.

Methylation Patterns of the FKBP5 Gene in Association with Childhood Maltreatment and Depressive Disorders

Childhood maltreatment is an important risk factor for adult depression and has been associated with changes in the hypothalamic pituitary adrenal (HPA) axis, including cortisol secretion and methylation of the FKBP5 gene. Furthermore, associations between depression and HPA changes have been reported. This study investigated the associations of whole-blood FKBP5 mRNA levels, serum cortisol levels, childhood maltreatment, and depressive symptoms with the whole-blood methylation status (assessed via target bisulfite sequencing) of 105 CpGs at the FKBP5 locus using data from the general population-based Study of Health in Pomerania (SHIP) (N = 203). Both direct and interaction effects with the rs1360780 single-nucleotide polymorphism were investigated. Nominally significant associations of main effects on methylation of a single CpG site were observed at intron 3, intron 7, and the 3'-end of the gene. Additionally, methylation at two clusters at the 3'-end and intron 7 were nominally associated with childhood maltreatment × rs1360780 and depressive symptoms × rs1360780, respectively. The results add to the understanding of molecular mechanisms underlying the emergence of depression and could aid the development of personalised depression therapy and drug development.

Epigenome-Wide Association Study Reveals CpG Sites Associated with Thyroid Function and Regulatory Effects on KLF9

Background: Thyroid hormones play a key role in differentiation and metabolism and are known regulators of gene expression through both genomic and epigenetic processes including DNA methylation. The aim of this study was to examine associations between thyroid hormones and DNA methylation. Methods: We carried out a fixed-effect meta-analysis of epigenome-wide association study (EWAS) of blood DNA methylation sites from 8 cohorts from the ThyroidOmics Consortium, incorporating up to 7073 participants of both European and African ancestry, implementing a discovery and replication stage. Statistical analyses were conducted using normalized beta CpG values as dependent and log-transformed thyrotropin (TSH), free thyroxine, and free triiodothyronine levels, respectively, as independent variable in a linear model. The replicated findings were correlated with gene expression levels in whole blood and tested for causal influence of TSH and free thyroxine by two-sample Mendelian randomization (MR). Results: Epigenome-wide significant associations (p-value <1.1E-7) of three CpGs for free thyroxine, five for free triiodothyronine, and two for TSH concentrations were discovered and replicated (combined p-values = 1.5E-9 to 4.3E-28). The associations included CpG sites annotated to KLF9 (cg00049440) and DOT1L (cg04173586) that overlap with all three traits, consistent with hypothalamic-pituitary-thyroid axis physiology. Significant associations were also found for CpGs in FKBP5 for free thyroxine, and at CSNK1D/LINCO1970 and LRRC8D for free triiodothyronine. MR analyses supported a causal effect of thyroid status on DNA methylation of KLF9. DNA methylation of cg00049440 in KLF9 was inversely correlated with KLF9 gene expression in blood. The CpG at CSNK1D/LINC01970 overlapped with thyroid hormone receptor alpha binding peaks in liver cells. The total additive heritability of the methylation levels of the six significant CpG sites was between 25% and 57%. Significant methylation QTLs were identified for CpGs at KLF9, FKBP5, LRRC8D, and CSNK1D/LINC01970. Conclusions: We report novel associations between TSH, thyroid hormones, and blood-based DNA methylation. This study advances our understanding of thyroid hormone action particularly related to KLF9 and serves as a proof-of-concept that integrations of EWAS with other -omics data can provide a valuable tool for unraveling thyroid hormone signaling in humans by complementing and feeding classical in vitro and animal studies.

The interplay between genetic variation and gene expression of the glucocorticoid receptor gene NR3C1 and blood cortisol levels on verbal memory and hippocampal volumes

The hypothalamus-pituitary-adrenal axis is the main physiological stress response system and regulating the release of cortisol. The two corticoid receptors encoded by the genes NR3C1 and NR3C2 are the main players in regulating the physiological response to cortisol. This biological system has been linked to neurocognitive processes and memory, yet the mechanisms remain largely unclear. In two independent general population studies (SHIP, total sample size > 5500), we aim to diseantangle the effects of genetic variation, gene expression and cortisol on verbal memory and memory associated brain volume. Especially for NR3C1 results exhibited a consistent pattern of direct an interactive effects. All three biological layers, genetic variation (rs56149945), gene expression for NR3C1 and cortisol levels, were directly associated with verbal memory. Interactions between these components showed significant effects on verbal memory as well as hippocampal volume. For NR3C2 such a complex association pattern could not be observed. Our analyses revealed that different components of the stress response system are acting together on different aspects of cognition. Complex phenotypes, such as cognition and memory function are regulated by a complex interplay between different genetic and epigenetic features. We promote the glucocorticoid receptor NR3C1 as a main target to focus in the context of verbal memory and provided a mechanistic concept of the interaction between various biological layers spanning NR3C1 function and its effects on memory. Especially the NR3C1 transcript seemed to be a key element in this complex system.

TREML2 Gene Expression and Its Missense Variant rs3747742 Associate with White Matter Hyperintensity Volume and Alzheimer's Disease-Related Brain Atrophy in the General Population

Although the common pathology of Alzheimer's disease (AD) and white matter hyperintensities (WMH) is disputed, the gene TREML2 has been implicated in both conditions: its whole-blood gene expression was associated with WMH volume and its missense variant rs3747742 with AD risk. We re-examined those associations within one comprehensive dataset of the general population, additionally searched for cross-relations and illuminated the role of the apolipoprotein E (APOE) ε4 status in the associations. For our linear regression and linear mixed effect models, we used 1949 participants from the Study of Health in Pomerania (Germany). AD was assessed using a continuous pre-symptomatic MRI-based score evaluating a participant's AD-related brain atrophy. In our study, increased whole-blood TREML2 gene expression was significantly associated with reduced WMH volume but not with the AD score. Conversely, rs3747742-C was significantly associated with a reduced AD score but not with WMH volume. The APOE status did not influence the associations. In sum, TREML2 robustly associated with WMH volume and AD-related brain atrophy on different molecular levels. Our results thus underpin TREML2's role in neurodegeneration, might point to its involvement in AD and WMH via different biological mechanisms, and highlight TREML2 as a worthwhile target for disentangling the two pathologies.

Limited evidence for blood eQTLs in human sexual dimorphism

BACKGROUND

The genetic underpinning of sexual dimorphism is very poorly understood. The prevalence of many diseases differs between men and women, which could be in part caused by sex-specific genetic effects. Nevertheless, only a few published genome-wide association studies (GWAS) were performed separately in each sex. The reported enrichment of expression quantitative trait loci (eQTLs) among GWAS-associated SNPs suggests a potential role of sex-specific eQTLs in the sex-specific genetic mechanism underlying complex traits.

METHODS

To explore this scenario, we combined sex-specific whole blood RNA-seq eQTL data from 3447 European individuals included in BIOS Consortium and GWAS data from UK Biobank. Next, to test the presence of sex-biased causal effect of gene expression on complex traits, we performed sex-specific transcriptome-wide Mendelian randomization (TWMR) analyses on the two most sexually dimorphic traits, waist-to-hip ratio (WHR) and testosterone levels. Finally, we performed power analysis to calculate the GWAS sample size needed to observe sex-specific trait associations driven by sex-biased eQTLs.

RESULTS

Among 9 million SNP-gene pairs showing sex-combined associations, we found 18 genes with significant sex-biased cis-eQTLs (FDR 5%). Our phenome-wide association study of the 18 top sex-biased eQTLs on >700 traits unraveled that these eQTLs do not systematically translate into detectable sex-biased trait-associations. In addition, we observed that sex-specific causal effects of gene expression on complex traits are not driven by sex-specific eQTLs. Power analyses using real eQTL- and causal-effect sizes showed that millions of samples would be necessary to observe sex-biased trait associations that are fully driven by sex-biased cis-eQTLs. Compensatory effects may further hamper their detection.

CONCLUSIONS

Our results suggest that sex-specific eQTLs in whole blood do not translate to detectable sex-specific trait associations of complex diseases, and vice versa that the observed sex-specific trait associations cannot be explained by sex-specific eQTLs.

Associations of carotid intima media thickness with gene expression in whole blood and genetically predicted gene expression across 48 tissues

Carotid intima media thickness (cIMT) is a biomarker of subclinical atherosclerosis and a predictor of future cardiovascular events. Identifying associations between gene expression levels and cIMT may provide insight to atherosclerosis etiology. Here, we use two approaches to identify associations between mRNA levels and cIMT: differential gene expression analysis in whole blood and S-PrediXcan. We used microarrays to measure genome-wide whole blood mRNA levels of 5647 European individuals from four studies. We examined the association of mRNA levels with cIMT adjusted for various potential confounders. Significant associations were tested for replication in three studies totaling 3943 participants. Next, we applied S-PrediXcan to summary statistics from a cIMT genome-wide association study (GWAS) of 71 128 individuals to estimate the association between genetically determined mRNA levels and cIMT and replicated these analyses using S-PrediXcan on an independent GWAS on cIMT that included 22 179 individuals from the UK Biobank. mRNA levels of TNFAIP3, CEBPD and METRNL were inversely associated with cIMT, but these associations were not significant in the replication analysis. S-PrediXcan identified associations between cIMT and genetically determined mRNA levels for 36 genes, of which six were significant in the replication analysis, including TLN2, which had not been previously reported for cIMT. There was weak correlation between our results using differential gene expression analysis and S-PrediXcan. Differential expression analysis and S-PrediXcan represent complementary approaches for the discovery of associations between phenotypes and gene expression. Using these approaches, we prioritize TNFAIP3, CEBPD, METRNL and TLN2 as new candidate genes whose differential expression might modulate cIMT.

AMDBNorm: an approach based on distribution adjustment to eliminate batch effects of gene expression data

Batch effects explain a large part of the noise when merging gene expression data. Removing irrelevant variations introduced by batch effects plays an important role in gene expression studies. To obtain reliable differential analysis results, it is necessary to remove the variation caused by technical conditions between different batches while preserving biological variation. Usually, merging data directly with batch effects leads to a sharp rise in false positives. Although some methods of batch correction have been developed, they have some drawbacks. In this study, we develop a new algorithm, adjustment mean distribution-based normalization (AMDBNorm), which is based on a probability distribution to correct batch effects while preserving biological variation. AMDBNorm solves the defects of the existing batch correction methods. We compared several popular methods of batch correction with AMDBNorm using two real gene expression datasets with batch effects and analyzed the results of batch correction from the visual and quantitative perspectives. To ensure the biological variation was well protected, the effects of the batch correction methods were verified by hierarchical cluster analysis. The results showed that the AMDBNorm algorithm could remove batch effects of gene expression data effectively and retain more biological variation than other methods. Our approach provides the researchers with reliable data support in the study of differential gene expression analysis and prognostic biomarker selection.

Meta-analyses identify DNA methylation associated with kidney function and damage

Chronic kidney disease is a major public health burden. Elevated urinary albumin-to-creatinine ratio is a measure of kidney damage, and used to diagnose and stage chronic kidney disease. To extend the knowledge on regulatory mechanisms related to kidney function and disease, we conducted a blood-based epigenome-wide association study for estimated glomerular filtration rate (n = 33,605) and urinary albumin-to-creatinine ratio (n = 15,068) and detected 69 and seven CpG sites where DNA methylation was associated with the respective trait. The majority of these findings showed directionally consistent associations with the respective clinical outcomes chronic kidney disease and moderately increased albuminuria. Associations of DNA methylation with kidney function, such as CpGs at JAZF1, PELI1 and CHD2 were validated in kidney tissue. Methylation at PHRF1, LDB2, CSRNP1 and IRF5 indicated causal effects on kidney function. Enrichment analyses revealed pathways related to hemostasis and blood cell migration for estimated glomerular filtration rate, and immune cell activation and response for urinary albumin-to-creatinineratio-associated CpGs.

Differentially expressed genes reflect disease-induced rather than disease-causing changes in the transcriptome

Comparing transcript levels between healthy and diseased individuals allows the identification of differentially expressed genes, which may be causes, consequences or mere correlates of the disease under scrutiny. We propose a method to decompose the observational correlation between gene expression and phenotypes driven by confounders, forward- and reverse causal effects. The bi-directional causal effects between gene expression and complex traits are obtained by Mendelian Randomization integrating summary-level data from GWAS and whole-blood eQTLs. Applying this approach to complex traits reveals that forward effects have negligible contribution. For example, BMI- and triglycerides-gene expression correlation coefficients robustly correlate with trait-to-expression causal effects (rBMI = 0.11, PBMI = 2.0 × 10-51 and rTG = 0.13, PTG = 1.1 × 10-68), but not detectably with expression-to-trait effects. Our results demonstrate that studies comparing the transcriptome of diseased and healthy subjects are more prone to reveal disease-induced gene expression changes rather than disease causing ones.

Mycolactone toxin induces an inflammatory response by targeting the IL-1β pathway: Mechanistic insight into Buruli ulcer pathophysiology

Mycolactone, a lipid-like toxin, is the major virulence factor of Mycobacterium ulcerans, the etiological agent of Buruli ulcer. Its involvement in lesion development has been widely described in early stages of the disease, through its cytotoxic and immunosuppressive activities, but less is known about later stages. Here, we revisit the role of mycolactone in disease outcome and provide the first demonstration of the pro-inflammatory potential of this toxin. We found that the mycolactone-containing mycobacterial extracellular vesicles produced by M. ulcerans induced the production of IL-1β, a potent pro-inflammatory cytokine, in a TLR2-dependent manner, targeting NLRP3/1 inflammasomes. We show our data to be relevant in a physiological context. The in vivo injection of these mycolactone-containing vesicles induced a strong local inflammatory response and tissue damage, which were prevented by corticosteroids. Finally, several soluble pro-inflammatory factors, including IL-1β, were detected in infected tissues from mice and Buruli ulcer patients. Our results revisit Buruli ulcer pathophysiology by providing new insight, thus paving the way for the development of new therapeutic strategies taking the pro-inflammatory potential of mycolactone into account.

Buruli ulcer is a neglected tropical disease occurring mainly in poor rural areas of West and Central Africa. This cutaneous disease is caused by Mycobacterium ulcerans, a bacterium belonging to the same family as M. tuberculosis and M. leprae. The skin lesions are caused by a cytotoxic toxin named mycolactone, also known to act as an immunosuppressor and an anti-inflammatory molecule. However, Buruli ulcer lesions are characterized by a chronic cutaneous inflammation with a recruitment of cellular immune cells trying to counteract M. ulcerans. Our work allows for a reconcilitation of previous observations. We found by in vitro experiment on macrophages that the mycolactone-containing mycobacterial extracellular vesicles produced by M. ulcerans induced the production of IL-1β, a potent pro-inflammatory molecule, while other pro-inflammatory soluble factors are inhibited. We also detected IL-1β protein in a mouse model of M. ulcerans infection as well as in biopsies of Buruli ulcer patients. The pro-inflammatory potential of mycolacone has to be taken into account to understand the full pathophysiology of Buruli ulcer.

Smoking-related changes in DNA methylation and gene expression are associated with cardio-metabolic traits

BACKGROUND

Tobacco smoking is a well-known modifiable risk factor for many chronic diseases, including cardiovascular disease (CVD). One of the proposed underlying mechanism linking smoking to disease is via epigenetic modifications, which could affect the expression of disease-associated genes. Here, we conducted a three-way association study to identify the relationship between smoking-related changes in DNA methylation and gene expression and their associations with cardio-metabolic traits.

RESULTS

We selected 2549 CpG sites and 443 gene expression probes associated with current versus never smokers, from the largest epigenome-wide association study and transcriptome-wide association study to date. We examined three-way associations, including CpG versus gene expression, cardio-metabolic trait versus CpG, and cardio-metabolic trait versus gene expression, in the Rotterdam study. Subsequently, we replicated our findings in The Cooperative Health Research in the Region of Augsburg (KORA) study. After correction for multiple testing, we identified both cis- and trans-expression quantitative trait methylation (eQTM) associations in blood. Specifically, we found 1224 smoking-related CpGs associated with at least one of the 443 gene expression probes, and 200 smoking-related gene expression probes to be associated with at least one of the 2549 CpGs. Out of these, 109 CpGs and 27 genes were associated with at least one cardio-metabolic trait in the Rotterdam Study. We were able to replicate the associations with cardio-metabolic traits of 26 CpGs and 19 genes in the KORA study. Furthermore, we identified a three-way association of triglycerides with two CpGs and two genes (GZMA; CLDND1), and BMI with six CpGs and two genes (PID1; LRRN3). Finally, our results revealed the mediation effect of cg03636183 (F2RL3), cg06096336 (PSMD1), cg13708645 (KDM2B), and cg17287155 (AHRR) within the association between smoking and LRRN3 expression.

CONCLUSIONS

Our study indicates that smoking-related changes in DNA methylation and gene expression are associated with cardio-metabolic risk factors. These findings may provide additional insights into the molecular mechanisms linking smoking to the development of CVD.

Blood DNA methylation sites predict death risk in a longitudinal study of 12, 300 individuals

DNA methylation has fundamental roles in gene programming and aging that may help predict mortality. However, no large-scale study has investigated whether site-specific DNA methylation predicts all-cause mortality. We used the Illumina-HumanMethylation450-BeadChip to identify blood DNA methylation sites associated with all-cause mortality for 12, 300 participants in 12 Cohorts of the Heart and Aging Research in Genetic Epidemiology (CHARGE) Consortium. Over an average 10-year follow-up, there were 2,561 deaths across the cohorts. Nine sites mapping to three intergenic and six gene-specific regions were associated with mortality (P < 9.3x10-7) independently of age and other mortality predictors. Six sites (cg14866069, cg23666362, cg20045320, cg07839457, cg07677157, cg09615688)-mapping respectively to BMPR1B, MIR1973, IFITM3, NLRC5, and two intergenic regions-were associated with reduced mortality risk. The remaining three sites (cg17086398, cg12619262, cg18424841)-mapping respectively to SERINC2, CHST12, and an intergenic region-were associated with increased mortality risk. DNA methylation at each site predicted 5%-15% of all deaths. We also assessed the causal association of those sites to age-related chronic diseases by using Mendelian randomization, identifying weak causal relationship between cg18424841 and cg09615688 with coronary heart disease. Of the nine sites, three (cg20045320, cg07839457, cg07677157) were associated with lower incidence of heart disease risk and two (cg20045320, cg07839457) with smoking and inflammation in prior CHARGE analyses. Methylation of cg20045320, cg07839457, and cg17086398 was associated with decreased expression of nearby genes (IFITM3, IRF, NLRC5, MT1, MT2, MARCKSL1) linked to immune responses and cardiometabolic diseases. These sites may serve as useful clinical tools for mortality risk assessment and preventative care.

Single-center versus multi-center data sets for molecular prognostic modeling: a simulation study

Background

Prognostic models based on high-dimensional omics data generated from clinical patient samples, such as tumor tissues or biopsies, are increasingly used for prognosis of radio-therapeutic success. The model development process requires two independent discovery and validation data sets. Each of them may contain samples collected in a single center or a collection of samples from multiple centers. Multi-center data tend to be more heterogeneous than single-center data but are less affected by potential site-specific biases. Optimal use of limited data resources for discovery and validation with respect to the expected success of a study requires dispassionate, objective decision-making. In this work, we addressed the impact of the choice of single-center and multi-center data as discovery and validation data sets, and assessed how this impact depends on the three data characteristics signal strength, number of informative features and sample size.

Methods

We set up a simulation study to quantify the predictive performance of a model trained and validated on different combinations of in silico single-center and multi-center data. The standard bioinformatical analysis workflow of batch correction, feature selection and parameter estimation was emulated. For the determination of model quality, four measures were used: false discovery rate, prediction error, chance of successful validation (significant correlation of predicted and true validation data outcome) and model calibration.

Results

In agreement with literature about generalizability of signatures, prognostic models fitted to multi-center data consistently outperformed their single-center counterparts when the prediction error was the quality criterion of interest. However, for low signal strengths and small sample sizes, single-center discovery sets showed superior performance with respect to false discovery rate and chance of successful validation.

Conclusions

With regard to decision making, this simulation study underlines the importance of study aims being defined precisely a priori. Minimization of the prediction error requires multi-center discovery data, whereas single-center data are preferable with respect to false discovery rate and chance of successful validation when the expected signal or sample size is low. In contrast, the choice of validation data solely affects the quality of the estimator of the prediction error, which was more precise on multi-center validation data.

Epigenetic Link Between Statin Therapy and Type 2 Diabetes

OBJECTIVE

To investigate the role of epigenetics in statins' diabetogenic effect comparing DNA methylation (DNAm) between statin users and nonusers in an epigenome-wide association study in blood.

RESEARCH DESIGN AND METHODS

Five cohort studies' participants (n = 8,270) were classified as statin users when they were on statin therapy at the time of DNAm assessment with Illumina 450K or EPIC array or noncurrent users otherwise. Associations of DNAm with various outcomes like incident type 2 diabetes, plasma glucose, insulin, and insulin resistance (HOMA of insulin resistance [HOMA-IR]) as well as with gene expression were investigated.

RESULTS

Discovery (n = 6,820) and replication (n = 1,450) phases associated five DNAm sites with statin use: cg17901584 (1.12 × 10-25 [DHCR24]), cg10177197 (3.94 × 10-08 [DHCR24]), cg06500161 (2.67 × 10-23 [ABCG1]), cg27243685 (6.01 × 10-09 [ABCG1]), and cg05119988 (7.26 × 10-12 [SC4MOL]). Two sites were associated with at least one glycemic trait or type 2 diabetes. Higher cg06500161 methylation was associated with higher fasting glucose, insulin, HOMA-IR, and type 2 diabetes (odds ratio 1.34 [95% CI 1.22, 1.47]). Mediation analyses suggested that ABCG1 methylation partially mediates the effect of statins on high insulin and HOMA-IR. Gene expression analyses showed that statin exposure and ABCG1 methylation were associated with ABCG1 downregulation, suggesting epigenetic regulation of ABCG1 expression. Further, outcomes insulin and HOMA-IR were significantly associated with ABCG1 expression.

CONCLUSIONS

This study sheds light on potential mechanisms linking statins with type 2 diabetes risk, providing evidence on DNAm partially mediating statins' effects on insulin traits. Further efforts shall disentangle the molecular mechanisms through which statins may induce DNAm changes, potentially leading to ABCG1 epigenetic regulation.

DeepWAS: Multivariate genotype-phenotype associations by directly integrating regulatory information using deep learning

Genome-wide association studies (GWAS) identify genetic variants associated with traits or diseases. GWAS never directly link variants to regulatory mechanisms. Instead, the functional annotation of variants is typically inferred by post hoc analyses. A specific class of deep learning-based methods allows for the prediction of regulatory effects per variant on several cell type-specific chromatin features. We here describe "DeepWAS", a new approach that integrates these regulatory effect predictions of single variants into a multivariate GWAS setting. Thereby, single variants associated with a trait or disease are directly coupled to their impact on a chromatin feature in a cell type. Up to 61 regulatory SNPs, called dSNPs, were associated with multiple sclerosis (MS, 4,888 cases and 10,395 controls), major depressive disorder (MDD, 1,475 cases and 2,144 controls), and height (5,974 individuals). These variants were mainly non-coding and reached at least nominal significance in classical GWAS. The prediction accuracy was higher for DeepWAS than for classical GWAS models for 91% of the genome-wide significant, MS-specific dSNPs. DSNPs were enriched in public or cohort-matched expression and methylation quantitative trait loci and we demonstrated the potential of DeepWAS to generate testable functional hypotheses based on genotype data alone. DeepWAS is available at https://github.com/cellmapslab/DeepWAS.

A meta-analysis of two high-risk prospective cohort studies reveals autism-specific transcriptional changes to chromatin, autoimmune, and environmental response genes in umbilical cord blood

Background

Autism spectrum disorder (ASD) is a neurodevelopmental disorder that affects more than 1% of children in the USA. ASD risk is thought to arise from both genetic and environmental factors, with the perinatal period as a critical window. Understanding early transcriptional changes in ASD would assist in clarifying disease pathogenesis and identifying biomarkers. However, little is known about umbilical cord blood gene expression profiles in babies later diagnosed with ASD compared to non-typically developing and non-ASD (Non-TD) or typically developing (TD) children.

Methods

Genome-wide transcript levels were measured by Affymetrix Human Gene 2.0 array in RNA from cord blood samples from both the Markers of Autism Risk in Babies-Learning Early Signs (MARBLES) and the Early Autism Risk Longitudinal Investigation (EARLI) high-risk pregnancy cohorts that enroll younger siblings of a child previously diagnosed with ASD. Younger siblings were diagnosed based on assessments at 36 months, and 59 ASD, 92 Non-TD, and 120 TD subjects were included. Using both differential expression analysis and weighted gene correlation network analysis, gene expression between ASD and TD, and between Non-TD and TD, was compared within each study and via meta-analysis.

Results

While cord blood gene expression differences comparing either ASD or Non-TD to TD did not reach genome-wide significance, 172 genes were nominally differentially expressed between ASD and TD cord blood (log₂(fold change) > 0.1, p < 0.01). These genes were significantly enriched for functions in xenobiotic metabolism, chromatin regulation, and systemic lupus erythematosus (FDR q < 0.05). In contrast, 66 genes were nominally differentially expressed between Non-TD and TD, including 8 genes that were also differentially expressed in ASD. Gene coexpression modules were significantly correlated with demographic factors and cell type proportions.

Limitations

ASD-associated gene expression differences identified in this study are subtle, as cord blood is not the main affected tissue, it is composed of many cell types, and ASD is a heterogeneous disorder.

Conclusions

This is the first study to identify gene expression differences in cord blood specific to ASD through a meta-analysis across two prospective pregnancy cohorts. The enriched gene pathways support involvement of environmental, immune, and epigenetic mechanisms in ASD etiology.

Discovery of mitochondrial DNA variants associated with genome-wide blood cell gene expression: a population-based mtDNA sequencing study

The effect of mitochondrial DNA (mtDNA) variation on peripheral blood transcriptomics in health and disease is not fully known. Sex-specific mitochondrially controlled gene expression patterns have been shown in Drosophila melanogaster but in humans, evidence is lacking. Functional variation in mtDNA may also have a role in the development of type 2 diabetes and its precursor state, i.e. prediabetes. We examined the associations between mitochondrial single-nucleotide polymorphisms (mtSNPs) and peripheral blood transcriptomics with a focus on sex- and prediabetes-specific effects. The genome-wide blood cell expression data of 19 637 probes, 199 deep-sequenced mtSNPs and nine haplogroups of 955 individuals from a population-based Young Finns Study cohort were used. Significant associations were identified with linear regression and analysis of covariance. The effects of sex and prediabetes on the associations between gene expression and mtSNPs were studied using random-effect meta-analysis. Our analysis identified 53 significant expression probe-mtSNP associations after Bonferroni correction, involving 7 genes and 31 mtSNPs. Eight probe-mtSNP signals remained independent after conditional analysis. In addition, five genes showed differential expression between haplogroups. The meta-analysis did not show any significant differences in linear model effect sizes between males and females but identified the association between the OASL gene and mtSNP C16294T to show prediabetes-specific effects. This study pinpoints new independent mtSNPs associated with peripheral blood transcriptomics and replicates six previously reported associations, providing further evidence of the mitochondrial genetic control of blood cell gene expression. In addition, we present evidence that prediabetes might lead to perturbations in mitochondrial control.

Methylation of the FKBP5 gene in association with FKBP5 genotypes, childhood maltreatment and depression

DNA methylation of the FKBP5 gene is assumed to alter FKBP5 expression and hence the synthesis of the FK506 binding protein 51, a central element of a genomic negative feedback loop for glucocorticoid receptor signaling. The present study aimed to replicate and extend previously reported influences of FKBP5 genotypes, childhood maltreatment and depression on methylation levels of five CpG sites in intron 7 of the FKBP5 gene in a large population-based sample. Besides the single nucleotide polymorphism (SNP) rs1360780, associations of the FKBP5 methylation with 22 other, unlinked FKBP5 SNPs as well as associations between FKBP5 methylation levels and transcription levels were investigated. Using whole-blood methylation of 3965 subjects of the Study of Health in Pomerania (SHIP) reduced methylation levels in TT allele carriers of rs1360780 (OR = 0.975, p = .005) and currently depressed subjects (OR = 0.995, p = 0.005) were found. Further, an impact of two yet undescribed SNPs (rs6910300, rs7771727) on methylation levels was observed. However, main and interactive effects for childhood maltreatment and lifetime major depressive disorder observed in previous studies could not be replicated. Finally, FKBP5 methylation levels were not related to FKBP5 transcription levels in whole blood. Thus, the present study verified the associations of FKBP5 genotypes and state depression on the FKBP5 methylation levels of five CpG sites in intron 7. However, FKBP5 methylation of these five CpG sites could not be validated as a valuable clinical biomarker for biological long-term effects of childhood maltreatment or lifetime depression.

Comparison of statistical methods and the use of quality control samples for batch effect correction in human transcriptome data

Batch effects are technical sources of variation introduced by the necessity of conducting gene expression analyses on different dates due to the large number of biological samples in population-based studies. The aim of this study is to evaluate the performances of linear mixed models (LMM) and Combat in batch effect removal. We also assessed the utility of adding quality control samples in the study design as technical replicates. In order to do so, we simulated gene expression data by adding “treatment” and batch effects to a real gene expression dataset. The performances of LMM and Combat, with and without quality control samples, are assessed in terms of sensitivity and specificity while correcting for the batch effect using a wide range of effect sizes, statistical noise, sample sizes and level of balanced/unbalanced designs. The simulations showed small differences among LMM and Combat. LMM identifies stronger relationships between big effect sizes and gene expression than Combat, while Combat identifies in general more true and false positives than LMM. However, these small differences can still be relevant depending on the research goal. When any of these methods are applied, quality control samples did not reduce the batch effect, showing no added value for including them in the study design.

Fatty liver is associated with blood pathways of inflammatory response, immune system activation and prothrombotic state in Young Finns Study

Fatty liver (FL) disease is the most common type of chronic liver disease. We hypothesized that liver’s response to the process where large droplets of triglyceride fat accumulate in liver cells is reflected also in gene pathway expression in blood. Peripheral blood genome wide gene expression analysis and ultrasonic imaging of liver were performed for 1,650 participants (316 individuals with FL and 1,334 controls) of the Young Finns Study. Gene set enrichment analysis (GSEA) was performed for the expression data. Fourteen gene sets were upregulated (false discovery rate, FDR < 0.05) in subjects with FL. These pathways related to extracellular matrix (ECM) turnover, immune response regulation, prothrombotic state and neural tissues. After adjustment for known risk factors and biomarkers of FL, we found i) integrin A4B1 signaling, ii) leukocyte transendothelial migration, iii) CD40/CD40L and iv) netrin-1 signaling pathways to be upregulated in individuals with FL (nominal p < 0.05). From these all but not ii) remained significantly upregulated when analyzing only subjects without history of heavy alcohol use. In conclusion, FL was associated with blood gene sets of ECM turnover, inflammatory response, immune system activation and prothrombotic state. These may form a systemic link between FL and the development of cardiovascular diseases.

Scalable Generation of Mesenchymal Stem Cells from Human Embryonic Stem Cells in 3D

Human embryonic stem cell (hESC) derived mesenchymal stem cells (EMSC) are efficacious in treating a series of autoimmune, inflammatory, and degenerative diseases in animal models. However, all the EMSC derivation methods reported so far rely on two-dimensional (2D) culture systems, which are inefficient, costive and difficult for large-scale production. HESC, as an unlimited source, can be successively propagated in spheroids. Here, we demonstrate that hESC spheroids can directly differentiate into MSC spheroids (EMSC_Sp) within 20 days in one vessel without passaging and the system is scalable to any desired size. EMSC_Sp can further differentiate into osteocytes and chondrocytes in spheres or demineralized bone matrix. EMSC_Sp also retains immune-modulatory effects in vitro and therapeutic effects on two mouse models of colitis after dissociation. Compared to EMSC differentiated in monolayer, EMSC_Sp-derived cells have faster proliferation and higher yield and develop less apoptosis and slower senescence. Thus, the 3D differentiation system allows simple, cost-effective, and scalable production of high-quality EMSC and subsequently bone and cartilage tissues for therapeutic application.

Automated Slide Scanning and Segmentation in Fluorescently-labeled Tissues Using a Widefield High-content Analysis System

Automated slide scanning and segmentation of fluorescently-labeled tissues is the most efficient way to analyze whole slides or large tissue sections. Unfortunately, many researchers spend large amounts of time and resources developing and optimizing workflows that are only relevant to their own experiments. In this article, we describe a protocol that can be used by those with access to a widefield high-content analysis system (WHCAS) to image any slide-mounted tissue, with options for customization within pre-built modules found in the associated software. Not originally intended for slide scanning, the steps detailed in this article make it possible to acquire slide scanning images in the WHCAS which can be imported into the associated software. In this example, the automated segmentation of brain tumor slides is demonstrated, but the automated segmentation of any fluorescently-labeled nuclear or cytoplasmic marker is possible. Furthermore, there are a variety of other quantitative software modules including assays for protein localization/translocation, cellular proliferation/viability/apoptosis, and angiogenesis that can be run. This technique will save researchers time and effort and create an automated protocol for slide analysis.

Blood RNA expression profiles undergo major changes during the seventh decade

Genome-wide alterations in RNA expression profiles are age-associated. Yet the rate and temporal pattern of those alterations are poorly understood. We investigated temporal changes in RNA expression profiles in blood from population-based studies using a quadratic regression model. Comparative analysis between two independent studies was carried out after sample-weighting that downsized differences in sample distribution over age between the datasets. We show that age-associated expression profiles are clustered into two major inclinations and transcriptional alternations occur predominantly from the seventh decade onwards. The age-associated genes in blood are enriched in functional groups of the translational machinery and the immune system. The results are highly consistent between the two population-based studies indicating that our analysis overcomes potential confounders in population-based studies. We suggest that the critical age when major transcriptional alterations occur could help understanding aging and disease risk during adulthood.

A whole-blood transcriptome meta-analysis identifies gene expression signatures of cigarette smoking

Cigarette smoking is a leading modifiable cause of death worldwide. We hypothesized that cigarette smoking induces extensive transcriptomic changes that lead to target-organ damage and smoking-related diseases. We performed a meta-analysis of transcriptome-wide gene expression using whole blood-derived RNA from 10,233 participants of European ancestry in six cohorts (including 1421 current and 3955 former smokers) to identify associations between smoking and altered gene expression levels. At a false discovery rate (FDR) <0.1, we identified 1270 differentially expressed genes in current vs. never smokers, and 39 genes in former vs. never smokers. Expression levels of 12 genes remained elevated up to 30 years after smoking cessation, suggesting that the molecular consequence of smoking may persist for decades. Gene ontology analysis revealed enrichment of smoking-related genes for activation of platelets and lymphocytes, immune response, and apoptosis. Many of the top smoking-related differentially expressed genes, including LRRN3 and GPR15, have DNA methylation loci in promoter regions that were recently reported to be hypomethylated among smokers. By linking differential gene expression with smoking-related disease phenotypes, we demonstrated that stroke and pulmonary function show enrichment for smoking-related gene expression signatures. Mediation analysis revealed the expression of several genes (e.g. ALAS2) to be putative mediators of the associations between smoking and inflammatory biomarkers (IL6 and C-reactive protein levels). Our transcriptomic study provides potential insights into the effects of cigarette smoking on gene expression in whole blood and their relations to smoking-related diseases. The results of such analyses may highlight attractive targets for treating or preventing smoking-related health effects.

Evidence for Stress-like Alterations in the HPA-Axis in Women Taking Oral Contraceptives

Using oral contraceptives has been implicated in the aetiology of stress-related disorders like depression. Here, we followed the hypothesis that oral contraceptives deregulate the HPA-axis by elevating circulating cortisol levels. We report for a sample of 233 pre-menopausal women increased circulating cortisol levels in those using oral contraceptives. For women taking oral contraceptives, we observed alterations in circulating phospholipid levels and elevated triglycerides and found evidence for increased glucocorticoid signalling as the transcript levels of the glucocorticoid-regulated genes DDIT4 and FKBP5 were increased in whole blood. The effects were statistically mediated by cortisol. The associations of oral contraceptives with higher FKBP5 mRNA and altered phospholipid levels were modified by rs1360780, a genetic variance implicated in psychiatric diseases. Accordingly, the methylation pattern of FKBP5 intron 7 was altered in women taking oral contraceptives depending on the rs1360780 genotype. Moreover, oral contraceptives modified the association of circulating cortisol with depressive symptoms, potentially explaining conflicting results in the literature. Finally, women taking oral contraceptives displayed smaller hippocampal volumes than non-using women. In conclusion, the integrative analyses of different types of physiological data provided converging evidence indicating that oral contraceptives may cause effects analogous to chronic psychological stressors regarding the regulation of the HPA axis.

Transcriptome-Wide Analysis Identifies Novel Associations With Blood Pressure

Hypertension represents a major cardiovascular risk factor. The pathophysiology of increased blood pressure (BP) is not yet completely understood. Transcriptome profiling offers possibilities to uncover genetics effects on BP. Based on 2 populations including 2549 individuals, a meta-analyses of monocytic transcriptome-wide profiles were performed to identify transcripts associated with BP. Replication was performed in 2 independent studies of whole-blood transcriptome data including 1990 individuals. For identified candidate genes, a direct link between long-term changes in BP and gene expression over time and by treatment with BP-lowering therapy was assessed. The predictive value of protein levels encoded by candidate genes for subsequent cardiovascular disease was investigated. Eight transcripts (CRIP1, MYADM, TIPARP, TSC22D3, CEBPA, F12, LMNA, and TPPP3) were identified jointly accounting for up to 13% (95% confidence interval, 8.7-16.2) of BP variability. Changes in CRIP1, MYADM, TIPARP, LMNA, TSC22D3, CEBPA, and TPPP3 expression associated with BP changes-among these, CRIP1 gene expression was additionally correlated to measures of cardiac hypertrophy. Assessment of circulating CRIP1 (cystein-rich protein 1) levels as biomarkers showed a strong association with increased risk for incident stroke (hazard ratio, 1.06; 95% confidence interval, 1.03-1.09; P=5.0×10-5). Our comprehensive analysis of global gene expression highlights 8 novel transcripts significantly associated with BP, providing a link between gene expression and BP. Translational approaches further established evidence for the potential use of CRIP1 as emerging disease-related biomarker.

Deciphering the genetic control of gene expression following Mycobacterium leprae antigen stimulation

Leprosy is a human infectious disease caused by Mycobacterium leprae. A strong host genetic contribution to leprosy susceptibility is well established. However, the modulation of the transcriptional response to infection and the mechanism(s) of disease control are poorly understood. To address this gap in knowledge of leprosy pathogenicity, we conducted a genome-wide search for expression quantitative trait loci (eQTL) that are associated with transcript variation before and after stimulation with M. leprae sonicate in whole blood cells. We show that M. leprae antigen stimulation mainly triggered the upregulation of immune related genes and that a substantial proportion of the differential gene expression is genetically controlled. Indeed, using stringent criteria, we identified 318 genes displaying cis-eQTL at an FDR of 0.01, including 66 genes displaying response-eQTL (reQTL), i.e. cis-eQTL that showed significant evidence for interaction with the M. leprae stimulus. Such reQTL correspond to regulatory variations that affect the interaction between human whole blood cells and M. leprae sonicate and, thus, likely between the human host and M. leprae bacilli. We found that reQTL were significantly enriched among binding sites of transcription factors that are activated in response to infection, and that they were enriched among single nucleotide polymorphisms (SNPs) associated with susceptibility to leprosy per se and Type-I Reaction, and seven of them have been targeted by recent positive selection. Our study suggested that natural selection shaped our genomic diversity to face pathogen exposure including M. leprae infection.

Each year, 200,000 new leprosy cases are reported worldwide. While there is unambiguous evidence for a role of host genetics in leprosy pathogenesis, the mechanisms by which the human host fights the infection are poorly understood. Here, we highlight the search for naturally occurring genetic variations that modulate gene expression levels following exposure to sonicate of Mycobacterium leprae, the bacterium causing the disease. Because M. leprae is not cultivable and the genuine immune cells involved in the host response during infection are still unknown, we performed a genome-wide search for such genetic variations after stimulation of whole-blood from leprosy patients with M. leprae sonicate. This design allowed to provide a general framework for the genetic control of host responses to M. leprae and outlined the contribution of host genetics to leprosy pathogenesis. Among the M. leprae-dependent genetic regulators of gene expression levels there was an enrichment of variants (i) associated with leprosy, (ii) located in transcription factor binding sites and (iii) targeted by recent positive selection.

Recapitulating and Correcting Marfan Syndrome in a Cellular Model

Marfan syndrome (MFS) is a connective tissue disorder caused by mutations in FBN1 gene, which encodes a key extracellular matrix protein FIBRILLIN-1. The haplosufficiency of FBN1 has been implicated in pathogenesis of MFS with manifestations primarily in cardiovascular, muscular, and ocular tissues. Due to limitations in animal models to study the late-onset diseases, human pluripotent stem cells (PSCs) offer a homogeneic tool for dissection of cellular and molecular pathogenic mechanism for MFS in vitro. Here, we first derived induced PSCs (iPSCs) from a MFS patient with a FBN1 mutation and corrected the mutation, thereby generating an isogenic "gain-of-function" control cells for the parental MFS iPSCs. Reversely, we knocked out FBN1 in both alleles in a wild-type (WT) human embryonic stem cell (ESC) line, which served as a loss-of-function model for MFS with the WT cells as an isogenic control. Mesenchymal stem cells derived from both FBN1-mutant iPSCs and -ESCs demonstrated reduced osteogenic differentiation and microfibril formation. We further demonstrated that vascular smooth muscle cells derived from FBN1-mutant iPSCs showed less sensitivity to carbachol as demonstrated by contractility and Ca²⁺ influx assay, compared to the isogenic controls cells. These findings were further supported by transcriptomic anaylsis of the cells. Therefore, this study based on both gain- and loss-of-function approaches confirmed the pathogenetic role of FBN1 mutations in these MFS-related phenotypic changes.

Epigenome-wide association study of body mass index, and the adverse outcomes of adiposity

Overweight and obesity affect ~1.5 billion people worldwide, and are major risk factors for type-2 diabetes (T2D), cardiovascular disease and related metabolic and inflammatory disturbances.1,2 Although the mechanisms linking adiposity to its clinical sequelae are poorly understood, recent studies suggest that adiposity may influence DNA methylation,3–6 a key regulator of gene expression and molecular phenotype.7 Here we use epigenome-wide association to show that body mass index (BMI, a key measure of adiposity) is associated with widespread changes in DNA methylation (187 genetic loci at P<1x10^-7, range P=9.2x10^-8 to 6.0x10^-46; N=10,261 samples). Genetic association analyses demonstrate that the alterations in DNA methylation are predominantly the consequence of adiposity, rather than the cause. We find the methylation loci are enriched for functional genomic features in multiple tissues (P<0.05), and show that sentinel methylation markers identify gene expression signatures at 38 loci (P<9.0x10^-6, range P=5.5x10^-6 to 6.1x10^-35, N=1,785 samples). The methylation loci identified highlight genes involved in lipid and lipoprotein metabolism, substrate transport, and inflammatory pathways. Finally, we show that the disturbances in DNA methylation predict future type-2 diabetes (relative risk per 1SD increase in Methylation Risk Score: 2.3 [2.07-2.56]; P=1.1x10^-54). Our results provide new insights into the biologic pathways influenced by adiposity, and may enable development of new strategies for prediction and prevention of type-2 diabetes and other adverse clinical consequences of obesity.

Genomewide meta-analysis identifies loci associated with IGF-I and IGFBP-3 levels with impact on age-related traits

The growth hormone/insulin‐like growth factor (IGF) axis can be manipulated in animal models to promote longevity, and IGF‐related proteins including IGF‐I and IGF‐binding protein‐3 (IGFBP‐3) have also been implicated in risk of human diseases including cardiovascular diseases, diabetes, and cancer. Through genomewide association study of up to 30 884 adults of European ancestry from 21 studies, we confirmed and extended the list of previously identified loci associated with circulating IGF‐I and IGFBP‐3 concentrations (IGF1, IGFBP3,GCKR,TNS3, GHSR, FOXO3, ASXL2, NUBP2/IGFALS, SORCS2, and CELSR2). Significant sex interactions, which were characterized by different genotype–phenotype associations between men and women, were found only for associations of IGFBP‐3 concentrations with SNPs at the loci IGFBP3 and SORCS2. Analyses of SNPs, gene expression, and protein levels suggested that interplay between IGFBP3 and genes within the NUBP2 locus (IGFALS and HAGH) may affect circulating IGF‐I and IGFBP‐3 concentrations. The IGF‐I‐decreasing allele of SNP rs934073, which is an eQTL of ASXL2, was associated with lower adiposity and higher likelihood of survival beyond 90 years. The known longevity‐associated variant rs2153960 (FOXO3) was observed to be a genomewide significant SNP for IGF‐I concentrations. Bioinformatics analysis suggested enrichment of putative regulatory elements among these IGF‐I‐ and IGFBP‐3‐associated loci, particularly of rs646776 at CELSR2. In conclusion, this study identified several loci associated with circulating IGF‐I and IGFBP‐3 concentrations and provides clues to the potential role of the IGF axis in mediating effects of known (FOXO3) and novel (ASXL2) longevity‐associated loci.

Comparison of pre-processing methods for multiplex bead-based immunoassays

BACKGROUND

High throughput protein expression studies can be performed using bead-based protein immunoassays, such as the Luminex® xMAP® technology. Technical variability is inherent to these experiments and may lead to systematic bias and reduced power. To reduce technical variability, data pre-processing is performed. However, no recommendations exist for the pre-processing of Luminex® xMAP® data.

RESULTS

We compared 37 different data pre-processing combinations of transformation and normalization methods in 42 samples on 384 analytes obtained from a multiplex immunoassay based on the Luminex® xMAP® technology. We evaluated the performance of each pre-processing approach with 6 different performance criteria. Three performance criteria were plots. All plots were evaluated by 15 independent and blinded readers. Four different combinations of transformation and normalization methods performed well as pre-processing procedure for this bead-based protein immunoassay.

CONCLUSIONS

The following combinations of transformation and normalization were suitable for pre-processing Luminex® xMAP® data in this study: weighted Box-Cox followed by quantile or robust spline normalization (rsn), asinh transformation followed by loess normalization and Box-Cox followed by rsn.

Endophenotype Network Models: Common Core of Complex Diseases

Historically, human diseases have been differentiated and categorized based on the organ system in which they primarily manifest. Recently, an alternative view is emerging that emphasizes that different diseases often have common underlying mechanisms and shared intermediate pathophenotypes, or endo(pheno)types. Within this framework, a specific disease’s expression is a consequence of the interplay between the relevant endophenotypes and their local, organ-based environment. Important examples of such endophenotypes are inflammation, fibrosis, and thrombosis and their essential roles in many developing diseases. In this study, we construct endophenotype network models and explore their relation to different diseases in general and to cardiovascular diseases in particular. We identify the local neighborhoods (module) within the interconnected map of molecular components, i.e., the subnetworks of the human interactome that represent the inflammasome, thrombosome, and fibrosome. We find that these neighborhoods are highly overlapping and significantly enriched with disease-associated genes. In particular they are also enriched with differentially expressed genes linked to cardiovascular disease (risk). Finally, using proteomic data, we explore how macrophage activation contributes to our understanding of inflammatory processes and responses. The results of our analysis show that inflammatory responses initiate from within the cross-talk of the three identified endophenotypic modules.

Removing Batch Effects from Longitudinal Gene Expression - Quantile Normalization Plus ComBat as Best Approach for Microarray Transcriptome Data

Technical variation plays an important role in microarray-based gene expression studies, and batch effects explain a large proportion of this noise. It is therefore mandatory to eliminate technical variation while maintaining biological variability. Several strategies have been proposed for the removal of batch effects, although they have not been evaluated in large-scale longitudinal gene expression data. In this study, we aimed at identifying a suitable method for batch effect removal in a large study of microarray-based longitudinal gene expression. Monocytic gene expression was measured in 1092 participants of the Gutenberg Health Study at baseline and 5-year follow up. Replicates of selected samples were measured at both time points to identify technical variability. Deming regression, Passing-Bablok regression, linear mixed models, non-linear models as well as ReplicateRUV and ComBat were applied to eliminate batch effects between replicates. In a second step, quantile normalization prior to batch effect correction was performed for each method. Technical variation between batches was evaluated by principal component analysis. Associations between body mass index and transcriptomes were calculated before and after batch removal. Results from association analyses were compared to evaluate maintenance of biological variability. Quantile normalization, separately performed in each batch, combined with ComBat successfully reduced batch effects and maintained biological variability. ReplicateRUV performed perfectly in the replicate data subset of the study, but failed when applied to all samples. All other methods did not substantially reduce batch effects in the replicate data subset. Quantile normalization plus ComBat appears to be a valuable approach for batch correction in longitudinal gene expression data.

Tobacco smoking is associated with DNA methylation of diabetes susceptibility genes

AIMS/HYPOTHESIS

Tobacco smoking, a risk factor for diabetes, is an established modifier of DNA methylation. We hypothesised that tobacco smoking modifies DNA methylation of genes previously identified for diabetes.

METHODS

We annotated CpG sites available on the Illumina Human Methylation 450K array to diabetes genes previously identified by genome-wide association studies (GWAS), and investigated them for an association with smoking by comparing current to never smokers. The discovery study consisted of 630 individuals (Bonferroni-corrected p = 1.4 × 10(-5)), and we sought replication in an independent sample of 674 individuals. The replicated sites were tested for association with nearby genetic variants and gene expression and fasting glucose and insulin levels.

RESULTS

We annotated 3,620 CpG sites to the genes identified in the GWAS on type 2 diabetes. Comparing current smokers to never smokers, we found 12 differentially methylated CpG sites, of which five replicated: cg23161492 within ANPEP (p = 1.3 × 10(-12)); cg26963277 (p = 1.2 × 10(-9)), cg01744331 (p = 8.0 × 10(-6)) and cg16556677 (p = 1.2 × 10(-5)) within KCNQ1 and cg03450842 (p = 3.1 × 10(-8)) within ZMIZ1. The effect of smoking on DNA methylation at the replicated CpG sites attenuated after smoking cessation. Increased DNA methylation at cg23161492 was associated with decreased gene expression levels of ANPEP (p = 8.9 × 10(-5)). rs231356-T, which was associated with hypomethylation of cg26963277 (KCNQ1), was associated with a higher odds of diabetes (OR 1.06, p = 1.3 × 10(-5)). Additionally, hypomethylation of cg26963277 was associated with lower fasting insulin levels (p = 0.04).

CONCLUSIONS/INTERPRETATION

Tobacco smoking is associated with differential DNA methylation of the diabetes risk genes ANPEP, KCNQ1 and ZMIZ1. Our study highlights potential biological mechanisms connecting tobacco smoking to excess risk of type 2 diabetes.

Antiphospholipid antibodies in a large population-based cohort: genome-wide associations and effects on monocyte gene expression

The antiphospholipid syndrome (APS) is characterised by venous and/or arterial thrombosis and pregnancy morbidity in women combined with the persistent presence of antiphospholipid antibodies (aPL). We aimed to identify genetic factors associated with the presence of aPL in a population based cohort. Furthermore, we wanted to clarify if the presence of aPL affects gene expression in circulating monocytes. Titres of IgG and IgM against cardiolipin, β2glycoprotein 1 (anti-β2GPI), and IgG against domain 1 of β2GPI (anti-domain 1) were determined in approx. 5,000 individuals from the Gutenberg Health Study (GHS) a population based cohort of German descent. Genotyping was conducted on Affymetrix Genome-Wide Human SNP 6.0 arrays. Monocyte gene expression was determined in a subgroup of 1,279 individuals by using the Illumina HT-12 v3 BeadChip. Gene expression data were confirmed in vitro and ex vivo by qRT-PCR. Genome wide analysis revealed significant associations of anti-β2GPI IgG and APOH on chromosome 17, which had been previously identified by candidate gene approaches, and of anti-domain1 and MACROD2 on chromosome 20 which has been listed in a previous GWAS as a suggestive locus associated with the occurrence of anti-β2GPI antibodies. Expression analysis confirmed increased expression of TNFα in monocytes and identified and confirmed neuron navigator 3 (NAV3) as a novel gene induced by aPL. In conclusion, MACROD2 represents a novel genetic locus associated with aPL. Furthermore, we show that aPL induce the expression of NAV3 in monocytes and endothelial cells. This will stimulate further research into the role of these genes in the APS.

Association between DNA Methylation in Whole Blood and Measures of Glucose Metabolism: KORA F4 Study

Epigenetic regulation has been postulated to affect glucose metabolism, insulin sensitivity and the risk of type 2 diabetes. Therefore, we performed an epigenome-wide association study for measures of glucose metabolism in whole blood samples of the population-based Cooperative Health Research in the Region of Augsburg F4 study using the Illumina HumanMethylation 450 BeadChip. We identified a total of 31 CpG sites where methylation level was associated with measures of glucose metabolism after adjustment for age, sex, smoking, and estimated white blood cell proportions and correction for multiple testing using the Benjamini-Hochberg (B-H) method (four for fasting glucose, seven for fasting insulin, 25 for homeostasis model assessment-insulin resistance [HOMA-IR]; B-H-adjusted p-values between 9.2x10(-5) and 0.047). In addition, DNA methylation at cg06500161 (annotated to ABCG1) was associated with all the aforementioned phenotypes and 2-hour glucose (B-H-adjusted p-values between 9.2x10(-5) and 3.0x10(-3)). Methylation status of additional three CpG sites showed an association with fasting insulin only after additional adjustment for body mass index (BMI) (B-H-adjusted p-values = 0.047). Overall, effect strengths were reduced by around 30% after additional adjustment for BMI, suggesting that this variable has an influence on the investigated phenotypes. Furthermore, we found significant associations between methylation status of 21 of the aforementioned CpG sites and 2-hour insulin in a subset of samples with seven significant associations persisting after additional adjustment for BMI. In a subset of 533 participants, methylation of the CpG site cg06500161 (ABCG1) was inversely associated with ABCG1 gene expression (B-H-adjusted p-value = 1.5x10(-9)). Additionally, we observed an enrichment of the top 1,000 CpG sites for diabetes-related canonical pathways using Ingenuity Pathway Analysis. In conclusion, our study indicates that DNA methylation and diabetes-related traits are associated and that these associations are partially BMI-dependent. Furthermore, the interaction of ABCG1 with glucose metabolism is modulated by epigenetic processes.

Gene transcripts associated with muscle strength: a CHARGE meta-analysis of 7,781 persons

Lower muscle strength in midlife predicts disability and mortality in later life. Blood-borne factors, including growth differentiation factor 11 (GDF11), have been linked to muscle regeneration in animal models. We aimed to identify gene transcripts associated with muscle strength in adults. Meta-analysis of whole blood gene expression (overall 17,534 unique genes measured by microarray) and hand-grip strength in four independent cohorts (n = 7,781, ages: 20-104 yr, weighted mean = 56), adjusted for age, sex, height, weight, and leukocyte subtypes. Separate analyses were performed in subsets (older/younger than 60, men/women). Expression levels of 221 genes were associated with strength after adjustment for cofactors and for multiple statistical testing, including ALAS2 (rate-limiting enzyme in heme synthesis), PRF1 (perforin, a cytotoxic protein associated with inflammation), IGF1R, and IGF2BP2 (both insulin like growth factor related). We identified statistical enrichment for hemoglobin biosynthesis, innate immune activation, and the stress response. Ten genes were associated only in younger individuals, four in men only and one in women only. For example, PIK3R2 (a negative regulator of PI3K/AKT growth pathway) was negatively associated with muscle strength in younger (<60 yr) individuals but not older (≥ 60 yr). We also show that 115 genes (52%) have not previously been linked to muscle in NCBI PubMed abstracts. This first large-scale transcriptome study of muscle strength in human adults confirmed associations with known pathways and provides new evidence for over half of the genes identified. There may be age- and sex-specific gene expression signatures in blood for muscle strength.

Analyzing Illumina Gene Expression Microarray Data Obtained From Human Whole Blood Cell and Blood Monocyte Samples

Microarray profiling of gene expression is widely applied to studies in molecular biology and functional genomics. Experimental and technical variations make not only the statistical analysis of single studies but also meta-analyses of different studies very challenging. Here, we describe the analytical steps required to substantially reduce the variations of gene expression data without affecting true effect sizes. A software pipeline has been established using gene expression data from a total of 3358 whole blood cell and blood monocyte samples, all from three German population-based cohorts, measured on the Illumina HumanHT-12 v3 BeadChip array. In summary, adjustment for a few selected technical factors greatly improved reliability of gene expression analyses. Such adjustments are particularly required for meta-analyses of different studies.

Activated immune-inflammatory pathways are associated with long-standing depressive symptoms: Evidence from gene-set enrichment analyses in the Young Finns Study

We used genome wide expression (GWE) data of circulating blood cells and pathway analysis to investigate the inflammatory and other molecular pathways that may be associated with long-standing depressive symptoms. Participants were 607 women and 316 men (mean age 42 years) from the Young Finns Study who participated in three consecutive study phases in 2001, 2007 and 2012. Using Gene-set enrichment analyses (GSEA) we focused our analyses to pathways (available in MSigDB database) that are likely to affect immunological and inflammatory processes. GSEA were performed for blood cell GWE data in 2012. Depressive symptoms were assessed using a modified 21-item Beck Depression Inventory in each of the three study phases. Participants who scored in the top quartile of depressive symptoms in each of the three measurement points (n = 191) differed from other participants (n = 732) in several gene-set pathways related to inflammatory processes or immune-inflammatory signaling including interleukin (IL-1) pathway, and pathways related to various immuno-inflammatory processes, such as toll-like, the NEF protein, the nuclear factor kB, the kinase AKT and the mature B cell antigen receptor pathway (false discovery rates, FDRs<0.12). The results provide novel genome wide molecular evidence that support the association between chronic depressive symptoms and altered immune-inflammatory regulation.

Characterization of whole-genome autosomal differences of DNA methylation between men and women

Background

Disease risk and incidence between males and females reveal differences, and sex is an important component of any investigation of the determinants of phenotypes or disease etiology. Further striking differences between men and women are known, for instance, at the metabolic level. The extent to which men and women vary at the level of the epigenome, however, is not well documented. DNA methylation is the best known epigenetic mechanism to date.

Results

In order to shed light on epigenetic differences, we compared autosomal DNA methylation levels between men and women in blood in a large prospective European cohort of 1799 subjects, and replicated our findings in three independent European cohorts. We identified and validated 1184 CpG sites to be differentially methylated between men and women and observed that these CpG sites were distributed across all autosomes. We showed that some of the differentially methylated loci also exhibit differential gene expression between men and women. Finally, we found that the differentially methylated loci are enriched among imprinted genes, and that their genomic location in the genome is concentrated in CpG island shores.

Conclusion

Our epigenome-wide association study indicates that differences between men and women are so substantial that they should be considered in design and analyses of future studies.

Electronic supplementary material

The online version of this article (doi:10.1186/s13072-015-0035-3) contains supplementary material, which is available to authorized users.

Extensive alterations of the whole-blood transcriptome are associated with body mass index: results of an mRNA profiling study involving two large population-based cohorts

Background

Obesity, defined as pathologically increased body mass index (BMI), is strongly related to an increased risk for numerous common cardiovascular and metabolic diseases. It is particularly associated with insulin resistance, hyperglycemia, and systemic oxidative stress and represents the most important risk factor for type 2 diabetes (T2D). However, the pathophysiological mechanisms underlying these associations are still not completely understood. Therefore, in order to identify potentially disease-relevant BMI-associated gene expression signatures, a transcriptome-wide association study (TWAS) on BMI was performed.

Methods

Whole-blood mRNA levels determined by array-based transcriptional profiling were correlated with BMI in two large independent population-based cohort studies (KORA F4 and SHIP-TREND) comprising a total of 1977 individuals.

Results

Extensive alterations of the whole-blood transcriptome were associated with BMI: More than 3500 transcripts exhibited significant positive or negative BMI-correlation. Three major whole-blood gene expression signatures associated with increased BMI were identified. The three signatures suggested: i) a ratio shift from mature erythrocytes towards reticulocytes, ii) decreased expression of several genes essentially involved in the transmission and amplification of the insulin signal, and iii) reduced expression of several key genes involved in the defence against reactive oxygen species (ROS).

Conclusions

Whereas the first signature confirms published results, the other two provide possible mechanistic explanations for well-known epidemiological findings under conditions of increased BMI, namely attenuated insulin signaling and increased oxidative stress. The putatively causative BMI-dependent down-regulation of the expression of numerous genes on the mRNA level represents a novel finding. BMI-associated negative transcriptional regulation of insulin signaling and oxidative stress management provide new insights into the pathogenesis of metabolic syndrome and T2D.

Electronic supplementary material

The online version of this article (doi:10.1186/s12920-015-0141-x) contains supplementary material, which is available to authorized users.

Re-Annotator: Annotation Pipeline for Microarray Probe Sequences

Microarray technologies are established approaches for high throughput gene expression, methylation and genotyping analysis. An accurate mapping of the array probes is essential to generate reliable biological findings. However, manufacturers of the microarray platforms typically provide incomplete and outdated annotation tables, which often rely on older genome and transcriptome versions that differ substantially from up-to-date sequence databases. Here, we present the Re-Annotator, a re-annotation pipeline for microarray probe sequences. It is primarily designed for gene expression microarrays but can also be adapted to other types of microarrays. The Re-Annotator uses a custom-built mRNA reference database to identify the positions of gene expression array probe sequences. We applied Re-Annotator to the Illumina Human-HT12 v4 microarray platform and found that about one quarter (25%) of the probes differed from the manufacturer's annotation. In further computational experiments on experimental gene expression data, we compared Re-Annotator to another probe re-annotation tool, ReMOAT, and found that Re-Annotator provided an improved re-annotation of microarray probes. A thorough re-annotation of probe information is crucial to any microarray analysis. The Re-Annotator pipeline is freely available at http://sourceforge.net/projects/reannotator along with re-annotated files for Illumina microarrays HumanHT-12 v3/v4 and MouseRef-8 v2.

CX3CR1 is dysregulated in blood and brain from schizophrenia patients

The molecular mechanisms underlying schizophrenia remain largely unknown. Although schizophrenia is a mental disorder, there is increasing evidence to indicate that inflammatory processes driven by diverse environmental factors play a significant role in its development. With gene expression studies having been conducted across a variety of sample types, e.g., blood and postmortem brain, it is possible to investigate convergent signatures that may reveal interactions between the immune and nervous systems in schizophrenia pathophysiology. We conducted two meta-analyses of schizophrenia microarray gene expression data (N=474) and non-psychiatric control (N=485) data from postmortem brain and blood. Then, we assessed whether significantly dysregulated genes in schizophrenia could be shared between blood and brain. To validate our findings, we selected a top gene candidate and analyzed its expression by RT-qPCR in a cohort of schizophrenia subjects stabilized by atypical antipsychotic monotherapy (N=29) and matched controls (N=31). Meta-analyses highlighted inflammation as the major biological process associated with schizophrenia and that the chemokine receptor CX3CR1 was significantly down-regulated in schizophrenia. This differential expression was also confirmed in our validation cohort. Given both the recent data demonstrating selective CX3CR1 expression in subsets of neuroimmune cells, as well as behavioral and neuropathological observations of CX3CR1 deficiency in mouse models, our results of reduced CX3CR1 expression adds further support for a role played by monocyte/microglia in the neurodevelopment of schizophrenia.

Molecular Characterization of the NLRC4 Expression in Relation to Interleukin-18 Levels

BACKGROUND

Interleukin-18 (IL-18) is a pleiotropic cytokine centrally involved in the cytokine cascade with complex immunomodulatory functions in innate and acquired immunity. Circulating IL-18 concentrations are associated with type 2 diabetes mellitus, cardiovascular events, and diverse inflammatory and autoimmune disorders.

METHODS AND RESULTS

To identify causal variants affecting circulating IL-18 concentrations, we applied various omics and molecular biology approaches. By genome-wide association study, we confirmed association of IL-18 levels with a single nucleotide polymorphism in the untranslated exon 2 of the inflammasome component NLRC4 (NLR family, caspase recruitment domain-containing 4) gene on chromosome 2 (rs385076; P=2.4 × 10(-45)). Subsequent molecular analyses by gene expression analysis and reporter gene assays indicated an effect of rs385076 on NLRC4 expression and differential isoform usage by modulating binding of the transcription factor PU.1.

CONCLUSIONS

Our study provides evidence for the functional causality of single nucleotide polymorphism rs385076 within the NLRC4 gene in relation to IL-18 activation.

The Human Blood Metabolome-Transcriptome Interface

Biological systems consist of multiple organizational levels all densely interacting with each other to ensure function and flexibility of the system. Simultaneous analysis of cross-sectional multi-omics data from large population studies is a powerful tool to comprehensively characterize the underlying molecular mechanisms on a physiological scale. In this study, we systematically analyzed the relationship between fasting serum metabolomics and whole blood transcriptomics data from 712 individuals of the German KORA F4 cohort. Correlation-based analysis identified 1,109 significant associations between 522 transcripts and 114 metabolites summarized in an integrated network, the 'human blood metabolome-transcriptome interface' (BMTI). Bidirectional causality analysis using Mendelian randomization did not yield any statistically significant causal associations between transcripts and metabolites. A knowledge-based interpretation and integration with a genome-scale human metabolic reconstruction revealed systematic signatures of signaling, transport and metabolic processes, i.e. metabolic reactions mainly belonging to lipid, energy and amino acid metabolism. Moreover, the construction of a network based on functional categories illustrated the cross-talk between the biological layers at a pathway level. Using a transcription factor binding site enrichment analysis, this pathway cross-talk was further confirmed at a regulatory level. Finally, we demonstrated how the constructed networks can be used to gain novel insights into molecular mechanisms associated to intermediate clinical traits. Overall, our results demonstrate the utility of a multi-omics integrative approach to understand the molecular mechanisms underlying both normal physiology and disease.

Associations between Serum Sex Hormone Concentrations and Whole Blood Gene Expression Profiles in the General Population

Background

Despite observational evidence from epidemiological and clinical studies associating sex hormones with various cardiometabolic risk factors or diseases, pathophysiological explanations are sparse to date. To reveal putative functional insights, we analyzed associations between sex hormone levels and whole blood gene expression profiles.

Methods

We used data of 991 individuals from the population-based Study of Health in Pomerania (SHIP-TREND) with whole blood gene expression levels determined by array-based transcriptional profiling and serum concentrations of total testosterone (TT), sex hormone-binding globulin (SHBG), free testosterone (free T), dehydroepiandrosterone sulfate (DHEAS), androstenedione (AD), estradiol (E2), and estrone (E1) measured by liquid chromatography-tandem mass spectrometry (LC-MS/MS) and immunoassay. Associations between sex hormone concentrations and gene expression profiles were analyzed using sex-specific regression models adjusted for age, body mass index, and technical covariables.

Results

In men, positive correlations were detected between AD and DDIT4 mRNA levels, as well as between SHBG and the mRNA levels of RPIA, RIOK3, GYPB, BPGM, and RAB2B. No additional significant associations were observed.

Conclusions

Besides the associations between AD and DDIT4 expression and SHBG and the transcript levels of RPIA, RIOK3, GYPB, BPGM, and RAB2B, the present study did not indicate any association between sex hormone concentrations and whole blood gene expression profiles in men and women from the general population.

Modulation of insulin degrading enzyme activity and liver cell proliferation

Diabetes mellitus type 2 (T2DM), insulin therapy, and hyperinsulinemia are independent risk factors of liver cancer. Recently, the use of a novel inhibitor of insulin degrading enzyme (IDE) was proposed as a new therapeutic strategy in T2DM. However, IDE inhibition might stimulate liver cell proliferation via increased intracellular insulin concentration. The aim of this study was to characterize effects of inhibition of IDE activity in HepG2 hepatoma cells and to analyze liver specific expression of IDE in subjects with T2DM. HepG2 cells were treated with 10 nM insulin for 24 h with or without inhibition of IDE activity using IDE RNAi, and cell transcriptome and proliferation rate were analyzed. Human liver samples (n = 22) were used for the gene expression profiling by microarrays. In HepG2 cells, IDE knockdown changed expression of genes involved in cell cycle and apoptosis pathways. Proliferation rate was lower in IDE knockdown cells than in controls. Microarray analysis revealed the decrease of hepatic IDE expression in subjects with T2DM accompanied by the downregulation of the p53-dependent genes FAS and CCNG2, but not by the upregulation of proliferation markers MKI67, MCM2 and PCNA. Similar results were found in the liver microarray dataset from GEO Profiles database. In conclusion, IDE expression is decreased in liver of subjects with T2DM which is accompanied by the dysregulation of p53 pathway. Prolonged use of IDE inhibitors for T2DM treatment should be carefully tested in animal studies regarding its potential effect on hepatic tumorigenesis.