Power estimation of the t test for detecting differential gene expression

There exist now a number of statistical methods for detecting differential gene expression in experiments with microarray data. In trials under two conditions, a version of the two-sample t statistic is usually used. However, the problem of estimating the power for these tests has so far been insufficiently studied. In this paper, we propose a method to calculate the power of the robust t test for detecting differential gene expression in experiments with twins. We discuss also the results of the implementation of this method to simulated data.     

X inactivation as a source of behavioural differences in monozygotic female twins.

Although members of monozygotic twin pairs are identical in genome sequence, they may differ in patterns of gene expression. One early and irreversible process affecting gene expression, which can create differences within pairs of female monozygotic twins, is X inactivation - one twin can express mainly paternally-received genes on the X chromosome while the other twin expresses mainly maternally-received genes. It follows that non-identical X chromosome expression may cause female monozygotic twins to correlate less strongly than male monozygotic twins on complex behavioural traits affected by X-linked loci. We tested this hypothesis using data from around 4000 same-sex twin pairs on 9 social, behavioural and cognitive measures at ages 2, 3 and 4. Consistent with our hypothesis, monozygotic males were generally more similar than monozygotic females. Three of four significant differences were in traits showing higher correlations in males than females, and these traits - prosocial behaviour, peer problems, and verbal ability - have all been proposed previously in the literature as being influenced by genes on the X chromosome. Interestingly, dizygotic twins showed the reverse pattern of correlations for similar variables, which is also consistent with the X inactivation hypothesis; taken together, then, our monozygotic and dizygotic results suggest the presence of quantitative trait loci on the X chromosome.     

Sex differences and non-additivity in the effects of genes on personality.

New large-sample data show that non-additive genetic effects, probably epistatic interactions between loci, and sex-limited gene expression are significant features of the genetic architecture of human personality as measured by questionnaire scales of extraversion and neuroticism. Three large data sets--new data on large samples (n = 20,554) of US twins, their spouses, parents, siblings and children, correlations for Australian twins (n = 7,532), and previously published twin data from Finland (n = 14,288)--are subjected to an integrated analysis to test alternative hypotheses about the genetic causes of family resemblance in personality. When allowance is made for differences in reliability of the scales, the combined data are consistent with the same model for variation. There are significant amounts of genetic non-additivity for both dimensions of personality. The evidence favours additive x additive epistatic interactions rather than dominance. In the case of neuroticism, there is especially strong evidence of sex differences in genetic architecture favouring a greater relative contribution of non-additive genetic effects in males. The data confirm previous claims to find no major contribution of the shared environment of twins and siblings to these dimensions of personality. Correlations between spouses are zero, and the correlations for very large samples of siblings and non-identical twins do not differ significantly.     

Additive,Epistatic, and Environmental Effects Through the Lens of Expression Variability QTL in a Twin Cohort

The expression of a gene can vary across individuals in the general population, as well as between monozygotic twins. This variable expression is assumed to be due to the influence of both genetic and nongenetic factors. Yet little evidence supporting this assumption has been obtained from empirical data. In this study, we used expression data from a large twin cohort to investigate the influences of genetic and nongenetic factors on variable gene expression. We focused on a set of expression variability QTL (evQTL)—i.e., genetic loci associated with the variance, as opposed to the mean, of gene expression. We identified evQTL for 99, 56, and 79 genes in lymphoblastoid cell lines, skin, and fat, respectively. The differences in gene expression, measured by the relative mean difference (RMD), tended to be larger between pairs of dizygotic (DZ) twins than between pairs of monozygotic (MZ) twins, showing that genetic background influenced the expression variability. Furthermore, a more profound RMD was observed between pairs of MZ twins whose genotypes were associated with greater expression variability than the RMD found between pairs of MZ twins whose genotypes were associated with smaller expression variability. This suggests that nongenetic (e.g., environmental) factors contribute to the variable expression. Lastly, we demonstrated that the formation of evQTL is likely due to partial linkages between eQTL SNPs that are additively associated with the mean of gene expression; in most cases, no epistatic effect is involved. Our findings have implications for understanding divergent sources of gene expression variability.     

Generalised Anxiety Disorder – A Twin Study of Genetic Architecture,Genome-Wide Association and Differential Gene Expression

Generalised Anxiety Disorder (GAD) is a common anxiety-related diagnosis, affecting approximately 5% of the adult population. One characteristic of GAD is a high degree of anxiety sensitivity (AS), a personality trait which describes the fear of arousal-related sensations. Here we present a genome-wide association study of AS using a cohort of 730 MZ and DZ female twins. The GWAS showed a significant association for a variant within the RBFOX1 gene. A heritability analysis of the same cohort also confirmed a significant genetic component with h2 of 0.42. Additionally, a subset of the cohort (25 MZ twins discordant for AS) was studied for evidence of differential expression using RNA-seq data. Significant differential expression of two exons with the ITM2B gene within the discordant MZ subset was observed, a finding that was replicated in an independent cohort. While previous research has shown that anxiety has a high comorbidity with a variety of psychiatric and neurodegenerative disorders, our analysis suggests a novel etiology specific to AS.     

DNA methylation and gene expression changes in monozygotic twins discordant for psoriasis: identification of epigenetically dysregulated genes

Monozygotic (MZ) twins do not show complete concordance for many complex diseases; for example, discordance rates for autoimmune diseases are 20%-80%. MZ discordance indicates a role for epigenetic or environmental factors in disease. We used MZ twins discordant for psoriasis to search for genome-wide differences in DNA methylation and gene expression in CD4(+) and CD8(+) cells using Illumina's HumanMethylation27 and HT-12 expression assays, respectively. Analysis of these data revealed no differentially methylated or expressed genes between co-twins when analyzed separately, although we observed a substantial amount of small differences. However, combined analysis of DNA methylation and gene expression identified genes where differences in DNA methylation between unaffected and affected twins were correlated with differences in gene expression. Several of the top-ranked genes according to significance of the correlation in CD4(+) cells are known to be associated with psoriasis. Further, gene ontology (GO) analysis revealed enrichment of biological processes associated with the immune response and clustering of genes in a biological pathway comprising cytokines and chemokines. These data suggest that DNA methylation is involved in an epigenetic dysregulation of biological pathways involved in the pathogenesis of psoriasis. This is the first study based on data from MZ twins discordant for psoriasis to detect epigenetic alterations that potentially contribute to development of the disease.     

Monozygotic twins reveal germline contribution to allelic expression differences

Variation in the level of gene expression is a major determinant of a cell's function and characteristics. Common allelic variants of genes can be expressed at different levels and thus contribute to phenotypic diversity. We have measured allelic expression differences at heterozygous loci in monozygotic twins and in unrelated individuals. We show that the extent of differential allelic expression is highly similar within monozygotic twin pairs for many loci, implying that allelic differences in gene expression are under genetic control. We also show that even subtle departures from equal allelic expression are often genetically determined.     

Estimation of false discovery proportion under general dependence

MOTIVATION: Wide-scale correlations between genes are commonly observed in gene expression data, due to both biological and technical reasons. These correlations increase the variability of the standard estimate of the false discovery rate (FDR). We highlight the false discovery proportion (FDP, instead of the FDR) as the suitable quantity for assessing differential expression in microarray data, demonstrate the deleterious effects of correlation on FDP estimation and propose an improved estimation method that accounts for the correlations. METHODS: We analyse the variation pattern of the distribution of test statistics under permutation using the singular value decomposition. The results suggest a latent FDR model that accounts for the effects of correlation, and is statistically closer to the FDP. We develop a procedure for estimating the latent FDR (ELF) based on a Poisson regression model. RESULTS: For simulated data based on the correlation structure of real datasets, we find that ELF performs substantially better than the standard FDR approach in estimating the FDP. We illustrate the use of ELF in the analysis of breast cancer and lymphoma data. AVAILABILITY: R code to perform ELF is available in http://www.meb.ki.se/~yudpaw.     

Environmental effects on gene expression phenotype have regional biases in the human genome

Phenotypic discordance between monozygotic twins, such as a difference in disease susceptibility, implicates the role of the environment in determining phenotype. To assess genomewide environmental effects on "gene expression phenotype," we employed a published microarray data set for twins. We found that variations in expression phenotypes between monozygotic twins have biases in their chromosomal locations. They also showed a strong inverse correlation with gene density. Genomic regions of low gene density were environmentally sensitive, containing genes involved in response to external signals, cell differentiation, and development, etc. Genetic factors were found to make no contribution to the observed regional biases, stressing the role of epigenetics. We propose that epigenetic modifications might occur more frequently in heterochromatic, gene-poor regions in response to environmental signals while gene-rich regions tend to remain in an active chromatin configuration for the constitutive expression of underlying genes.     

Camera: a competitive gene set test accounting for inter-gene correlation

Competitive gene set tests are commonly used in molecular pathway analysis to test for enrichment of a particular gene annotation category amongst the differential expression results from a microarray experiment. Existing gene set tests that rely on gene permutation are shown here to be extremely sensitive to inter-gene correlation. Several data sets are analyzed to show that inter-gene correlation is non-ignorable even for experiments on homogeneous cell populations using genetically identical model organisms. A new gene set test procedure (CAMERA) is proposed based on the idea of estimating the inter-gene correlation from the data, and using it to adjust the gene set test statistic. An efficient procedure is developed for estimating the inter-gene correlation and characterizing its precision. CAMERA is shown to control the type I error rate correctly regardless of inter-gene correlations, yet retains excellent power for detecting genuine differential expression. Analysis of breast cancer data shows that CAMERA recovers known relationships between tumor subtypes in very convincing terms. CAMERA can be used to analyze specified sets or as a pathway analysis tool using a database of molecular signatures.     

A multiorganism based method for Bayesian gene network estimation

The primary goal of this article is to infer genetic interactions based on gene expression data. A new method for multiorganism Bayesian gene network estimation is presented based on multitask learning. When the input datasets are sparse, as is the case in microarray gene expression data, it becomes difficult to separate random correlations from true correlations that would lead to actual edges when modeling the gene interactions as a Bayesian network. Multitask learning takes advantage of the similarity between related tasks, in order to construct a more accurate model of the underlying relationships represented by the Bayesian networks. The proposed method is tested on synthetic data to illustrate its validity. Then it is iteratively applied on real gene expression data to learn the genetic regulatory networks of two organisms with homologous genes.     

Gender diagnosticity and androgen receptor gene CAG repeat sequence.

The gender diagnosticity (GD) approach of Lippa (1995) was used to evaluate the relationship of within-sex differences in psychological masculinity-femininity to a genetic characteristic, the length of a repeated CAG sequence in the X-linked androgen receptor (AR) gene. Previously assessed adult samples in Australia and Sweden were used for this purpose. A weak relationship (correlations in the range .11 to .14) was obtained in both countries. Additional data from adolescent twins from Australia (12-, 14-, 16-year-olds) did not confirm such a relationship at those ages, especially for males. The fact that this sample consisted of twins permitted two kinds of within-pair comparisons: (1) Did the dizygotic twin who had the longer AR sequence have the higher GD score? (2) Was one twin's GD score more highly correlated with the other twin's AR score in MZ than in DZ pairs? The answer in both cases was negative. Clarification of these relationships will require large samples and measurements at additional ages.     

基于人类信号网络和表达谱数据挖掘动脉粥样硬化相关模块

目的:动脉粥样硬化是一种高致死率的慢性炎症疾病,其发生和发展的机制尚不明确。本文基于人类信号网络和基因表达谱数据对动脉粥样硬化相关模块进行挖掘,以探究其在疾病发生发展中的作用机制。方法:结合人类信号网络和基因表达谱数据,设计显著差异模块筛选策略,通过功能分析,挖掘动脉粥样硬化相关模块,对动脉粥样硬化的致病机制进行研究。结果:基于网络模块的平均表达值改变量,采用两种随机方法,进行显著差异模块筛选,最终获得8个动脉粥样硬化相关的显著差异模块。结论:应用本文提出的整合筛选策略,能识别与动脉粥样硬化相关的模块,获得潜在的致病基因,并从外周血的基因表达改变来探究动脉粥样硬化致病机制,这对动脉粥样硬化的诊断、治疗以及发生发展机制的研究具有重要意义。     

Multidimensional analysis and functional assignment of DNA-microarray transcription profiles of genes involved in adipogenesis

The results of two independent DNA-microarray experiments concerning adipogenesis in the murine preadipocyte 3T3-L1 cell line, which covered the first two days after the induction of differentiation, were analyzed using the multidimensional scaling (MDS) method. In both data arrays, the first three scaling components accounted for 73.5–73.8% of the total dispersion. This result implies that both arrays of the gene expression profiles are in fact three-dimensional and each component reflects a definite principal process involved in one of the three early stages of adipogenesis: (i) determination of the fibroblast-like stem cells, (ii) clonal expansion of adipoblasts, and (iii) preadipocyte conversion into a mature adipocyte phenotype. Each profile of the gene expression is characterized by coefficients of correlation with the first three scaling components. The functional annotation in terms of the Gene Ontology database profiles (sorted according to the correlations with each component) generally corresponds to a regular change of elementary biological processes during the three early stages of adipogenesis. Analysis of correlations with the principal scaling components for the genes previously classified as subject to differential expression in the course of adipogenesis in mice suggests a complicated role of these genes in early adipogenesis (in some cases, described in the literature). The MDS analysis of the gene expression profiles and the analysis of correlations between these profiles and the main scaling components provides a deeper insight into the fine role of these genes and makes possible the search for new biomarkers of various differentiation stages.

     

Lipoprotein(a) in women twins: heritability and relationship to apolipoprotein(a) phenotypes.

Lp(a) is a unique lipoprotein consisting of an LDL-like particle and a characteristic protein, apo(a). Increased levels of Lp(a) constitute a risk factor for coronary heart disease. Variation in the size of the apo(a) protein is a phenotype controlled by the apo(a) gene on chromosome 6 and is related to Lp(a) plasma levels. Based on 169 MZ and 125 DZ adult female twin pairs, this study's purpose was to estimate the proportion of the variation in Lp(a) levels that is due to genetic influences and to determine the extent to which the apo(a) locus explains this heritability. Lp(a) levels were significantly more similar in MZ twins than in DZ twins: mean co-twin differences were 3.9 +/- 5.7 mg/dl and 16.0 +/- 19.9 mg/dl (P less than .001), respectively. Intraclass correlations were .94 in MZ twins and .32 in DZ twins, resulting in a heritability estimate of .94 (P less than .001). Heritability was then calculated using only co-twins with the same apo(a) phenotype: the heritability estimate decreased to .45 but was still highly significant (P less than .001). Therefore, on the basis of heritability analysis of women twins, Lp(a) levels are almost entirely genetically controlled. Variation at the apo(a) locus contributes to this heritability, although other genetic factors could be involved.