GenoExp: a web tool for predicting gene expression levels from single nucleotide polymorphisms

Understanding the effect of single nucleotide polymorphisms (SNPs) on the expression level of genes is an important goal. We recently published a study in which we devised a multi-SNP predictive model for gene expression in Lymphoblastoid cell lines (LCL), and showed that it can robustly predict the expression of a small number of genes in test individuals. Here, we validate the generality of our models by predicting expression profiles for genes in LCL in an independent study, and extend the pool of predictable genes for which we are able to explain more than 25% of their expression variability to 232 genes across 14 different cell types. As the number of people who obtained their SNP profiles through companies such as 23andMe is rising rapidly, we developed GenoExp, a web-based tool in which users can upload their individual SNP data and obtain predicted expression levels for the set of predictable genes across the 14 different cell types. Our tool thus allows users with biological knowledge to study the possible effects that their set of SNPs might have on these genes and predict their cell-specific expression levels relative to the population average.

AVAILABILITY AND IMPLEMENTATION

GenoExp is freely available at http://genie.weizmann.ac.il/pubs/GenoExp/.

Mapping the inner workings of the microbiome: genomic- and metagenomic-based study of metabolism and metabolic interactions in the human microbiome

The human gut microbiome is a major contributor to human metabolism and health, yet the metabolic processes that are carried out by various community members, the way these members interact with each other and with the host, and the impact of such interactions on the overall metabolic machinery of the microbiome have not yet been mapped. Here, we discuss recent efforts to study the metabolic inner workings of this complex ecosystem. We will specifically highlight two interrelated lines of work, the first aiming to deconvolve the microbiome and to characterize the metabolic capacity of various microbiome species and the second aiming to utilize computational modeling to infer and study metabolic interactions between these species.

Probing the effect of promoters on noise in gene expression using thousands of designed sequences

Genetically identical cells exhibit large variability (noise) in gene expression, with important consequences for cellular function. Although the amount of noise decreases with and is thus partly determined by the mean expression level, the extent to which different promoter sequences can deviate away from this trend is not fully known. Here, we present a high-throughput method for measuring promoter-driven noise for thousands of designed synthetic promoters in parallel. We use it to investigate how promoters encode different noise levels and find that the noise levels of promoters with similar mean expression levels can vary more than one order of magnitude, with nucleosome-disfavoring sequences resulting in lower noise and more transcription factor binding sites resulting in higher noise. We propose a kinetic model of gene expression that takes into account the nonspecific DNA binding and one-dimensional sliding along the DNA, which occurs when transcription factors search for their target sites. We show that this assumption can improve the prediction of the mean-independent component of expression noise for our designed promoter sequences, suggesting that a transcription factor target search may affect gene expression noise. Consistent with our findings in designed promoters, we find that binding-site multiplicity in native promoters is associated with higher expression noise. Overall, our results demonstrate that small changes in promoter DNA sequence can tune noise levels in a manner that is predictable and partly decoupled from effects on the mean expression levels. These insights may assist in designing promoters with desired noise levels.

Multifaceted activities of type I interferon are revealed by a receptor antagonist

Type I interferons (IFNs), including various IFN-α isoforms and IFN-β, are a family of homologous, multifunctional cytokines. IFNs activate different cellular responses by binding to a common receptor that consists of two subunits, IFNAR1 and IFNAR2. In addition to stimulating antiviral responses, they also inhibit cell proliferation and modulate other immune responses. We characterized various IFNs, including a mutant IFN-α2 (IFN-1ant) that bound tightly to IFNAR2 but had markedly reduced binding to IFNAR1. Whereas IFN-1ant stimulated antiviral activity in a range of cell lines, it failed to elicit immunomodulatory and antiproliferative activities. The antiviral activities of the various IFNs tested depended on a set of IFN-sensitive genes (the "robust" genes) that were controlled by canonical IFN response elements and responded at low concentrations of IFNs. Conversely, these elements were not found in the promoters of genes required for the antiproliferative responses of IFNs (the "tunable" genes). The extent of expression of tunable genes was cell type-specific and correlated with the magnitude of the antiproliferative effects of the various IFNs. Although IFN-1ant induced the expression of robust genes similarly in five different cell lines, its antiviral activity was virus- and cell type-specific. Our findings suggest that IFN-1ant may be a therapeutic candidate for the treatment of specific viral infections without inducing the immunomodulatory and antiproliferative functions of wild-type IFN.

Landscape and variation of RNA secondary structure across the human transcriptome

In parallel to the genetic code for protein synthesis, a second layer of information is embedded in all RNA transcripts in the form of RNA structure. RNA structure influences practically every step in the gene expression program¹. Yet the nature of most RNA structures or effects of sequence variation on structure are not known. Here we report the initial landscape and variation of RNA secondary structures (RSS) in a human family Trio, providing a comprehensive RSS map of human coding and noncoding RNAs. We identify unique RSS signatures that demarcate open reading frames, splicing junctions, and define authentic microRNA binding sites. Comparison of native deproteinized RNA isolated from cells versus refolded purified RNA suggests that the majority of the RSS information is encoded within RNA sequence. Over 1900 transcribed single nucleotide variants (~15% of all transcribed SNVs) alter local RNA structure. We discover simple sequence and spacing rules that determine the ability of point mutations to impact RSS. Selective depletion of RiboSNitches versus structurally synonymous variants at precise locations suggests selection for specific RNA shapes at thousands of sites, including 3’UTRs, binding sites of miRNAs and RNA binding proteins genome-wide. These results highlight the potentially broad contribution of RNA structure and its variation to gene regulation.

Predicting disease risk using bootstrap ranking and classification algorithms

Genome-wide association studies (GWAS) are widely used to search for genetic loci that underlie human disease. Another goal is to predict disease risk for different individuals given their genetic sequence. Such predictions could either be used as a “black box” in order to promote changes in life-style and screening for early diagnosis, or as a model that can be studied to better understand the mechanism of the disease. Current methods for risk prediction typically rank single nucleotide polymorphisms (SNPs) by the p-value of their association with the disease, and use the top-associated SNPs as input to a classification algorithm. However, the predictive power of such methods is relatively poor. To improve the predictive power, we devised BootRank, which uses bootstrapping in order to obtain a robust prioritization of SNPs for use in predictive models. We show that BootRank improves the ability to predict disease risk of unseen individuals in the Wellcome Trust Case Control Consortium (WTCCC) data and results in a more robust set of SNPs and a larger number of enriched pathways being associated with the different diseases. Finally, we show that combining BootRank with seven different classification algorithms improves performance compared to previous studies that used the WTCCC data. Notably, diseases for which BootRank results in the largest improvements were recently shown to have more heritability than previously thought, likely due to contributions from variants with low minimum allele frequency (MAF), suggesting that BootRank can be beneficial in cases where SNPs affecting the disease are poorly tagged or have low MAF. Overall, our results show that improving disease risk prediction from genotypic information may be a tangible goal, with potential implications for personalized disease screening and treatment.

Genome-wide association studies are widely used to search for genetic loci that underlie human disease. Another goal is to predict disease risk for different individuals given their genetic sequence. Such predictions could either be used as a “black box” in order to promote changes in life-style and screening for early diagnosis, or as a model that can be studied to better understand the mechanism of the disease. Current methods for risk prediction have relatively poor performance, with one possible explanation being the fact they rely on a noisy ranking of genetic variants given to them as input. To improve the predictive power, we devised BootRank, a ranking method less sensitive to noise. We show that BootRank improves the ability to predict disease risk of unseen individuals in the Wellcome Trust Case Control Consortium (WTCCC) data, and that combining BootRank with different classification algorithms improves performance compared to previous studies that used these data. Overall, our results show that improving disease risk prediction from genotypic information may be a tangible goal, with potential implications for personalized disease screening and treatment.

Robust prediction of expression differences among human individuals using only genotype information

Many genetic variants that are significantly correlated to gene expression changes across human individuals have been identified, but the ability of these variants to predict expression of unseen individuals has rarely been evaluated. Here, we devise an algorithm that, given training expression and genotype data for a set of individuals, predicts the expression of genes of unseen test individuals given only their genotype in the local genomic vicinity of the predicted gene. Notably, the resulting predictions are remarkably robust in that they agree well between the training and test sets, even when the training and test sets consist of individuals from distinct populations. Thus, although the overall number of genes that can be predicted is relatively small, as expected from our choice to ignore effects such as environmental factors and trans sequence variation, the robust nature of the predictions means that the identity and quantitative degree to which genes can be predicted is known in advance. We also present an extension that incorporates heterogeneous types of genomic annotations to differentially weigh the importance of the various genetic variants, and we show that assigning higher weights to variants with particular annotations such as proximity to genes and high regional G/C content can further improve the predictions. Finally, genes that are successfully predicted have, on average, higher expression and more variability across individuals, providing insight into the characteristics of the types of genes that can be predicted from their cis genetic variation.

Variation in gene expression across different individuals has been found to play a role in susceptibility to different diseases. In addition, many genetic variants that are linked to changes in expression have been found to date. However, their joint ability to accurately predict these changes is not well understood and has rarely been evaluated. Here, we devise a method that uses multiple genetic variants to explain the variation in expression of genes across individuals. One important aspect of our method is its robustness, in that our predictions agree well between training and test sets. Thus, although the number of genes that could be explained is relatively small, the identity and quantitative degree to which genes can be predicted is known in advance. We also present an extension to our method that integrates different genomic annotations such as location of the genetic variant or its context to differentially weigh the genetic variants in our model and improve predictions. Finally, genes that are successfully predicted have, on average, higher expression and more variability across individuals, providing insight into the characteristics of the types of genes that can be predicted by our method.

Sex differences in risk factors for coronary artery disease and stroke in men and women aged 45-65 years

BACKGROUND

Coronary artery disease (CAD) and stroke both result from atherosclerosis. Risk factor profiles for CAD and stroke have been reported to differ between middle-aged men and women.

OBJECTIVE

To compare, for men and women aged 45-65 years, between risk factor profiles for CAD and stroke.

METHODS

This is a retrospective study based on the medical records of 179 women and 270 men diagnosed with CAD, and 114 women and 190 men diagnosed with stroke, hospitalised in one of two medical centres in Jerusalem. We assessed and compared the number of metabolic risk factors (diabetes, hypertension and hypercholesterolaemia) presenting among men and women between the CAD and stroke groups.

RESULTS

Among patients with CAD, significantly more women than men presented with diabetes, hypertension and hypercholesterolaemia. In contrast, no statistically significant differences were observed between genders in the prevalence of diabetes, hypertension and hypercholesterolaemia among the stroke patients. Hypertension was more prevalent in both men and women among stroke patients than CAD patients. In the stroke group, 29.1% of the women compared with 14.2% of the men presented with the three metabolic risk factors investigated.

CONCLUSIONS

In a middle-aged population, CAD risk factor profiles differed between genders while stroke risk factor profiles did not.

FC21-06 - Schizophrenia after the 1941 farhud (a pogrom in IRAQ); a study of 6,781 IRAQI-born men and women observed as parents in the jerusalem cohort

Introduction

Maternal stress during pregnancy has been associated with schizophrenia in some settings with evidence suggesting that male and female offspring might be affected differently.

Objectives/aims

In an outbreak of violence in 1941, at least 180 Jews died in Baghdad; rapes, beatings and property damage injured hundreds more in other towns. We questioned whether this stress might be reflected in any change in schizophrenia incidence in people born in 1941.

Methods

We studied admissions to psychiatric hospitals in 39,606 men and 41,208 women, parents of the population-based Jerusalem cohort. We used proportional hazards models to estimate relative incidence of schizophrenia in various groups, over time. Subjects were followed from age in 1950 or immigration, till age at first hospital admission, death or 2004. Schizophrenia was defined by discharge diagnosis, ICD-10 = F20–F29 at any hospital event. Models controlled for secular and cyclic time trends.

Results

For all years combined, the 3,679 male immigrants from Iraq showed a schizophrenia incidence similar to other men (relative risk RR = 0.92, 95% confidence limits 0.67–1.26). But for 141 men born in Iraq in 1941, RR = 4.12 (1.67–10.2, p = .0021 based on 6 cases), compared with men from Iraq born in all other years. Among other men, RR for 1941 births was 1.21 (0.78–1.88,). Women from Iraq showed no significant findings.

Conclusions

These findings enlarge on long-term consequences of ethnic violence. They raise intriguing questions about the relative resilience of the sexes, but should be interpreted cautiously, given that all subjects in this cohort had been able to reproduce.

Age, sex and first treatment of schizophrenia in a population cohort

OBJECTIVE

Schizophrenia affects men more than women, but this may not be true at all ages. This study examines the incidence of first hospitalization for treatment of schizophrenia in each sex over different ages.

METHODS

We compared the incidence of first admission for treatment in a cohort of 46,388 males and 43,680 females followed from birth until ages 29-41, using life tables and proportional hazards methods.

RESULTS

Life table estimates of cumulative incidence by age 40 were 1.44% in males and 0.86% in females. For over all ages the relative risk (RR) in males was 1.6 (95% confidence limits=1.4-1.8) compared with females. Before age 17 there was no significant difference between the sexes (RR=0.86, 0.56-1.3). Excess risk in males was observed only from age 17 (RR=1.7, 1.4-1.9). There was no evidence of the incidence in females catching up with that in males, during the 30s.

CONCLUSION

In this population, there was a significant change, over age, in the relative incidence of first hospitalization for schizophrenia between the sexes; the excess incidence in males first developed at age 17.

Long noncoding RNA as modular scaffold of histone modification complexes

Long intergenic noncoding RNAs (lincRNAs) regulate chromatin states and epigenetic inheritance. Here, we show that the lincRNA HOTAIR serves as a scaffold for at least two distinct histone modification complexes. A 5' domain of HOTAIR binds polycomb repressive complex 2 (PRC2), whereas a 3' domain of HOTAIR binds the LSD1/CoREST/REST complex. The ability to tether two distinct complexes enables RNA-mediated assembly of PRC2 and LSD1 and coordinates targeting of PRC2 and LSD1 to chromatin for coupled histone H3 lysine 27 methylation and lysine 4 demethylation. Our results suggest that lincRNAs may serve as scaffolds by providing binding surfaces to assemble select histone modification enzymes, thereby specifying the pattern of histone modifications on target genes.

EGF decreases the abundance of microRNAs that restrain oncogenic transcription factors

Epidermal growth factor (EGF) stimulates cells by launching gene expression programs that are frequently deregulated in cancer. MicroRNAs, which attenuate gene expression by binding complementary regions in messenger RNAs, are broadly implicated in cancer. Using genome-wide approaches, we showed that EGF stimulation initiates a coordinated transcriptional program of microRNAs and transcription factors. The earliest event involved a decrease in the abundance of a subset of 23 microRNAs. This step permitted rapid induction of oncogenic transcription factors, such as c-FOS, encoded by immediate early genes. In line with roles as suppressors of EGF receptor (EGFR) signaling, we report that the abundance of this early subset of microRNAs is decreased in breast and in brain tumors driven by the EGFR or the closely related HER2. These findings identify specific microRNAs as attenuators of growth factor signaling and oncogenesis.

The effect of a handwashing intervention on preschool educator beliefs, attitudes, knowledge and self-efficacy

This paper describes the effect of a preschool hygiene intervention program on psychosocial measures of educators regarding handwashing and communicable pediatric disease. A cluster-randomized trial, with randomization at the level of the preschool, was run in 40 Jerusalem preschool classrooms. Eighty preschool educators participated. The program used a multipronged approach which included elements aimed at staff, children, parents, school nurses and the classroom environment. Frontal lectures by medical, epidemiological and educational experts, along with printed materials and experiential learning, were provided to staff. Responses from a validated survey instrument were used to build four scales for each respondent regarding beliefs, attitudes, self-efficacy and knowledge. The scales were built on a Likert-type 1-7 scale (1 = minimum, 7 = maximum). The effect of the intervention was tested using mixed model analysis of variance. Response was received from 92.5% of educators. Educators believed that handwashing could affect health (mean = 5.5, SD = 1.1), had high levels of self-efficacy (mean = 6.1, SD = 0.9) and had positive attitudes toward handwashing (mean = 5.7, SD = 1.2). Knowledge was affected by the intervention (intervention: mean = 6.2, SD = 0.7; control: mean = 5.8, SD = 0.8). The combination of positive attitudes toward handwashing among educators and the program's effectiveness in imparting knowledge helped to create a sustained social norm of handwashing among many children in disparate locations.

Co-occurrence of risk factors for cardiovascular disease by social class: 1958 British birth cohort

AIM

To establish whether social differences in multiple risk factors for cardiovascular disease are due to a greater strength of association (higher correlation) between risk factors in less advantaged groups.

METHODS

Co-occurrence of five risk factors (smoking, hypertension, low high-density lipoprotein cholesterol, obesity, diabetes) in 3614 British 45-year-old men and 3560 women in the manual and non-manual social groups.

RESULTS

4.0% of women in manual groups had >or=3 risk factors compared with 1.7% in non-manual groups: 6.2% and 3.4% respectively for men. There was a higher than expected percentage of the population, overall, with >or=3 risk factors assuming independence between risk factors; correspondingly, there was a slightly lower than expected proportion with one factor. However, patterns of observed to expected ratios were consistent in manual and non-manual groups and did not differ by the number of risk factors.

CONCLUSIONS

Higher prevalence of multiple risk factors in manual groups was due to the higher prevalence of individual factors rather than a greater tendency of those with an individual risk factor to have additional risks. Strategies to reduce multiple risk factors in less advantaged groups would help to lessen their health burden.

Examining cultural, psychosocial, community and behavioural factors in relationship to socioeconomic inequalities in limiting longstanding illness among the Arab minority in Israel

BACKGROUND

Few studies have examined the explanatory pathways to social inequalities in health within ethnic minorities. The current study examined the relative contributions of specific pathways explaining the associations between socioeconomic status (SES) and limiting longstanding illness (LLI) among the Arab minority in Israel.

METHODS

A cross-sectional study of a random sample of 902 individuals aged 30-70 selected in a multistage sampling procedure. SES was measured by education, land ownership and relative family income. Five-stage logistic regressions assessed the attenuations in the odds of LLI among those with lower SES compared to higher SES after including relevant groups of explanatory factors: psychosocial, behavioural and community, and their integration.

RESULTS

Rates of LLI were significantly higher in participants with lower SES. Inclusion of groups of explanatory variables attenuated all SES-LLI associations in a similar pattern: psychosocial factors played a main explanatory role, yielding 15-40% attenuation in odds ratios (OR). The contribution of community indicators was modest (10-21%); health behaviours had a marginal contribution (6-7%). Cultural factors were not associated with SES or LLI. The integrative model contributed up to 49% reduction in the OR.

CONCLUSIONS

The significant associations between SES and LLI suggest that formative policy to reduce SES-LLI disparities should emphasise creating opportunities for economic development to improve SES, which was the main predictor of inequalities. Combining strategies of community capacity building and reinforcement of individual inner resources might be complementary. Such conclusions might apply to other minorities in a similar context, for which future studies are required.

Cancer risk after exposure to treatments for ovulation induction

Uncertainty continues as to whether treatments for ovulation induction are associated with increased risk of cancer. The authors conducted a long-term population-based historical cohort study of parous women. A total of 15,030 women in the Jerusalem Perinatal Study who gave birth in 1974-1976 participated in a postpartum survey. Cancer incidence through 2004 was analyzed using Cox's proportional hazards models, controlling for age and other covariates. Women who used drugs to induce ovulation (n = 567) had increased risks of cancer at any site (multivariate hazard ratio (HR) = 1.36, 95% confidence interval (CI): 1.06, 1.74). An increased risk of uterine cancer was found among women treated with ovulation-inducing agents (HR = 3.39, 95% CI: 1.28, 8.97), specifically clomiphene (HR = 4.56, 95% CI: 1.56, 13.34). No association was noted between use of ovulation-inducing agents and ovarian cancer (age-adjusted HR = 0.61, 95% CI: 0.08, 4.42). Ovulation induction was associated with a borderline-significant increased risk of breast cancer (multivariate HR = 1.42, 95% CI: 0.99, 2.05). Increased risks were also observed for malignant melanoma and non-Hodgkin lymphoma. These associations appeared stronger among women who waited more than 1 year to conceive. Additional follow-up studies assessing these associations by drug type, dosage, and duration are needed.

Twin pregnancy and the risk of schizophrenia

BACKGROUND

Twins are exposed to intrauterine environments that differ significantly from those of singletons. These diverse environments might alter the risk for schizophrenia in twins and make it difficult to generalize from findings in twins when studying the risk of schizophrenia in the general population. Previous studies report contradictory findings on the risk for schizophrenia in twins.

METHODS

We studied the incidence of schizophrenia spectrum disorders, ascertained from Israel's National Psychiatric Registry, in a cohort of 2124 twins and 87,955 singletons. These offspring were followed from their birth in 1964-76 in the Jerusalem Perinatal study. Cox proportional hazards methods were used to compare outcomes over 28-41 years, adjusting for ages of parents.

RESULTS

Twins showed a relative risk [RR] of .84 relative to singletons, with a 95% confidence interval [CI] of (.51-1.4). RRs and CIs for males and females were .68 [.34-1.4] and 1.1 [.55-2.2] respectively. Twins in male-male, female-female or opposite-sex sets showed no significant variation in RRs; furthermore, first- or second-born twins did not differ significantly from each other. Siblings of twins had the same risk of schizophrenia as siblings of singletons.

CONCLUSION

Twins have the same risk for schizophrenia as the general population.