Epistasis: what it means, what it doesn't mean, and statistical methods to detect it in humans

Accurately annotate compound effects of genetic variants using a context-sensitive framework

In genomics, effectively identifying the biological effects of genetic variants is crucial. Current methods handle each variant independently, assuming that each variant acts in a context-free manner. However, variants within the same gene may interfere with each other, producing combinational (compound) rather than individual effects. In this work, we introduce COPE, a gene-centric variant annotation tool that integrates the entire sequential context in evaluating the functional effects of intra-genic variants. Applying COPE to the 1000 Genomes dataset, we identified numerous cases of multiple-variant compound effects that frequently led to false-positive and false-negative loss-of-function calls by conventional variant-centric tools. Specifically, 64 disease-causing mutations were identified to be rescued in a specific genomic context, thus potentially contributing to the buffering effects for highly penetrant deleterious mutations. COPE is freely available for academic use at http://cope.cbi.pku.edu.cn.

Effect of Population Size and Mutation Rate on the Evolution of RNA Sequences on an Adaptive Landscape Determined by RNA Folding

The dynamics of populations evolving on an adaptive landscape depends on multiple factors, including the structure of the landscape, the rate of mutations, and effective population size. Existing theoretical work often makes ad hoc and simplifying assumptions about landscape structure, whereas experimental work can vary important parameters only to a limited extent. We here overcome some of these limitations by simulating the adaptive evolution of RNA molecules, whose fitness is determined by the thermodynamics of RNA secondary structure folding. We study the influence of mutation rates and population sizes on final mean population fitness, on the substitution rates of mutations, and on population diversity. We show that evolutionary dynamics cannot be understood as a function of mutation rate µ, population size N, or population mutation rate Nµ alone. For example, at a given mutation rate, clonal interference prevents the fixation of beneficial mutations as population size increases, but larger populations still arrive at a higher mean fitness. In addition, at the highest population mutation rates we study, mean final fitness increases with population size, because small populations are driven to low fitness by the relatively higher incidence of mutations they experience. Our observations show that mutation rate and population size can interact in complex ways to influence the adaptive dynamics of a population on a biophysically motivated fitness landscape.

Pathway-based discovery of genetic interactions in breast cancer

Breast cancer is the second largest cause of cancer death among U.S. women and the leading cause of cancer death among women worldwide. Genome-wide association studies (GWAS) have identified several genetic variants associated with susceptibility to breast cancer, but these still explain less than half of the estimated genetic contribution to the disease. Combinations of variants (i.e. genetic interactions) may play an important role in breast cancer susceptibility. However, due to a lack of statistical power, the current tests for genetic interactions from GWAS data mainly leverage prior knowledge to focus on small sets of genes or SNPs that are known to have an association with breast cancer. Thus, many genetic interactions, particularly among novel variants, remain understudied. Reverse-genetic interaction screens in model organisms have shown that genetic interactions frequently cluster into highly structured motifs, where members of the same pathway share similar patterns of genetic interactions. Based on this key observation, we recently developed a method called BridGE to search for such structured motifs in genetic networks derived from GWAS studies and identify pathway-level genetic interactions in human populations. We applied BridGE to six independent breast cancer cohorts and identified significant pathway-level interactions in five cohorts. Joint analysis across all five cohorts revealed a high confidence consensus set of genetic interactions with support in multiple cohorts. The discovered interactions implicated the glutathione conjugation, vitamin D receptor, purine metabolism, mitotic prometaphase, and steroid hormone biosynthesis pathways as major modifiers of breast cancer risk. Notably, while many of the pathways identified by BridGE show clear relevance to breast cancer, variants in these pathways had not been previously discovered by traditional single variant association tests, or single pathway enrichment analysis that does not consider SNP-SNP interactions.

Dynamics behind affinity maturation of an anti-HCMV antibody family influencing antigen binding

The investigation of antibody affinity maturation and its effects on antigen binding is important with respect to understanding the regulation of the immune response. To shed light on this crucial process, we analyzed two Igs neutralizing the human cytomegalovirus: the primary germline antibody M2J1 and its related mature antibody 8F9. Both antibodies target the AD-2S1 epitope of the gB envelope protein and are considered to establish similar interactions with the cognate antigen. We used molecular dynamics simulations to understand the effect of mutations on the antibody-antigen interactions. The results provide a qualitative explanation for the increased 8F9 peptide affinity compared with that of M2J1. The emerging atomistic-detailed description of these complexes reveals the molecular effects of the somatic hypermutations occurring during affinity maturation.

Linear mixed model for heritability estimation that explicitly addresses environmental variation

The linear mixed model (LMM) is now routinely used to estimate heritability. Unfortunately, as we demonstrate, LMM estimates of heritability can be inflated when using a standard model. To help reduce this inflation, we used a more general LMM with two random effects-one based on genomic variants and one based on easily measured spatial location as a proxy for environmental effects. We investigated this approach with simulated data and with data from a Uganda cohort of 4,778 individuals for 34 phenotypes including anthropometric indices, blood factors, glycemic control, blood pressure, lipid tests, and liver function tests. For the genomic random effect, we used identity-by-descent estimates from accurately phased genome-wide data. For the environmental random effect, we constructed a covariance matrix based on a Gaussian radial basis function. Across the simulated and Ugandan data, narrow-sense heritability estimates were lower using the more general model. Thus, our approach addresses, in part, the issue of "missing heritability" in the sense that much of the heritability previously thought to be missing was fictional. Software is available at https://github.com/MicrosoftGenomics/FaST-LMM.

Further investigations of the W-test for pairwise epistasis testing

Background: In a recent paper, a novel W-test for pairwise epistasis testing was proposed that appeared, in computer simulations, to have higher power than competing alternatives. Application to genome-wide bipolar data detected significant epistasis between SNPs in genes of relevant biological function. Network analysis indicated that the implicated genes formed two separate interaction networks, each containing genes highly related to autism and neurodegenerative disorders.

Methods: Here we investigate further the properties and performance of the W-test via theoretical evaluation, computer simulations and application to real data.

Results: We demonstrate that, for common variants, the W-test is closely related to several existing tests of association allowing for interaction, including logistic regression on 8 degrees of freedom, although logistic regression can show inflated type I error for low minor allele frequencies,  whereas the W-test shows good/conservative type I error control. Although in some situations the W-test can show higher power, logistic regression is not limited to tests on 8 degrees of freedom but can instead be tailored to impose greater structure on the assumed alternative hypothesis, offering a power advantage when the imposed structure matches the true structure.

Conclusions: The W-test is a potentially useful method for testing for association - without necessarily implying interaction - between genetic variants disease, particularly when one or more of the genetic variants are rare. For common variants, the advantages of the W-test are less clear, and, indeed, there are situations where existing methods perform better. In our investigations, we further uncover a number of problems with the practical implementation and application of the W-test (to bipolar disorder) previously described, apparently due to inadequate use of standard data quality-control procedures. This observation leads us to urge caution in interpretation of the previously-presented results, most of which we consider are highly likely to be artefacts.

Distribution of Silicified Microstructures, Regulation of Cinnamyl Alcohol Dehydrogenase and Lodging Resistance in Silicon and Paclobutrazol Mediated Oryza sativa

Lodging is a phenomenon that affects most of the cereal crops including rice, Oryza sativa. This is due to the fragile nature of herbaceous plants whose stems are non-woody, thus affecting its ability to grow upright. Silicon (Si), a beneficial nutrient is often used to toughen and protect plants from biotic and abiotic stresses. Deposition of Si in plant tissues enhances the rigidity and stiffness of the plant as a whole. Silicified cells provide the much needed strength to the culm to resist breaking. Lignin plays important roles in cell wall structural integrity, stem strength, transport, mechanical support, and plant pathogen defense. The aim of this study is to resolve effects of Si on formation of microstructure and regulation of cinnamyl alcohol dehydrogenase (CAD), a key gene responsible for lignin biosynthesis. Besides evaluating silicon, paclobutrazol (PBZ) a plant growth retartdant that reduces internode elongation is also incorporated in this study. Hardness, brittleness and stiffness were improved in presence of silicon thus reducing lodging. Scanning electron micrographs with the aid of energy dispersive x-ray (EDX) was used to map silicon distribution. Presence of trichomes, silica cells, and silica bodies were detected in silicon treated plants. Transcripts of CAD gene was also upregulated in these plants. Besides, phloroglucinol staining showed presence of lignified vascular bundles and sclerenchyma band. In conclusion, silicon treated rice plants showed an increase in lignin content, silicon content, and formation of silicified microstructures.

Epistasis in genomic and survival data of cancer patients

Cancer aggressiveness and its effect on patient survival depends on mutations in the tumor genome. Epistatic interactions between the mutated genes may guide the choice of anticancer therapy and set predictive factors of its success. Inhibitors targeting synthetic lethal partners of genes mutated in tumors are already utilized for efficient and specific treatment in the clinic. The space of possible epistatic interactions, however, is overwhelming, and computational methods are needed to limit the experimental effort of validating the interactions for therapy and characterizing their biomarkers. Here, we introduce SurvLRT, a statistical likelihood ratio test for identifying epistatic gene pairs and triplets from cancer patient genomic and survival data. Compared to established approaches, SurvLRT performed favorable in predicting known, experimentally verified synthetic lethal partners of PARP1 from TCGA data. Our approach is the first to test for epistasis between triplets of genes to identify biomarkers of synthetic lethality-based therapy. SurvLRT proved successful in identifying the known gene TP53BP1 as the biomarker of success of the therapy targeting PARP in BRCA1 deficient tumors. Search for other biomarkers for the same interaction revealed a region whose deletion was a more significant biomarker than deletion of TP53BP1. With the ability to detect not only pairwise but twelve different types of triple epistasis, applicability of SurvLRT goes beyond cancer therapy, to the level of characterization of shapes of fitness landscapes.

Genomic alterations in tumors affect the fitness of tumor cells, controlling how well they replicate and survive compared to other cells. The landscape of tumor fitness is shaped by epistasis. Epistasis occurs when the contribution of gene alterations to the total fitness is non-linear. The type of epistatic genetic interactions with great potential for cancer therapy is synthetic lethality. Inhibitors targeting synthetic lethal partners of genes mutated in tumors can selectively kill tumor and not normal cells. Therapy based on synthetic lethality is, however, context dependent, and it is crucial to identify its biomarkers. Unfortunately, the space of possible interactions and their biomarkers is overwhelming for experimental validation. Computational pre-selection methods are required to limit the experimental effort. Here, we introduce a statistical approach called SurvLRT, for the identification of epistatic gene pairs and triplets based on patient genomic and survival data. First, we show that using SurvLRT, we can deliver synthetic lethal interactions of pairs of genes that are specific to cancer. Second, we demonstrate the applicability of SurvLRT to identify biomarkers for synthetic lethality, such as mutational status of other genes that can alleviate the synthetic effect.

On the mechanistic nature of epistasis in a canonical cis-regulatory element

Understanding the relation between genotype and phenotype remains a major challenge. The difficulty of predicting individual mutation effects, and particularly the interactions between them, has prevented the development of a comprehensive theory that links genotypic changes to their phenotypic effects. We show that a general thermodynamic framework for gene regulation, based on a biophysical understanding of protein-DNA binding, accurately predicts the sign of epistasis in a canonical cis-regulatory element consisting of overlapping RNA polymerase and repressor binding sites. Sign and magnitude of individual mutation effects are sufficient to predict the sign of epistasis and its environmental dependence. Thus, the thermodynamic model offers the correct null prediction for epistasis between mutations across DNA-binding sites. Our results indicate that a predictive theory for the effects of cis-regulatory mutations is possible from first principles, as long as the essential molecular mechanisms and the constraints these impose on a biological system are accounted for.

DOI:http://dx.doi.org/10.7554/eLife.25192.001

Mutations are changes to DNA that provide the raw material upon which evolution can act. Therefore, to understand evolution, we need to know the effects of mutations, and how those mutations interact with each other (a phenomenon referred to as epistasis). So far, few mathematical models allow scientists to predict the effects of mutations, and even fewer are able to predict epistasis.

Biological systems are complex and consist of many proteins and other molecules. Genes are the sections of DNA that provide the instructions needed to produce these molecules, and some genes encode proteins that can bind to DNA to control whether other genes are switched on or off. Lagator, Paixão et al. have now used mathematical models and experiments to understand how the environment inside the cells of a bacterium known as E. coli, specifically the amount of particular proteins, affects epistasis.

These mathematical models are able to predict interactions between mutations in the most abundant class of DNA-binding sites in proteins. This approach found that the nature of the interaction between mutations can be explained through biophysical laws, combined with the basic knowledge of the logic of how genes regulate each other’s activities. Furthermore, the models allow Lagator, Paixão et al. to predict interactions between mutations in several different environments, such as the presence of a new food source or a toxin, defined by the amounts of relevant DNA-binding proteins in cells.

By providing new ways of understanding how genes are regulated in bacteria, and how gene regulation is affected by mutations, these findings contribute to our understanding of how organisms evolve. In addition, this work may help us to build artificial networks of genes that interact with each other to produce a desired response, such as more efficient production of fuel from ethanol or the break down of hazardous chemicals.

DOI:http://dx.doi.org/10.7554/eLife.25192.002

The "Mendelian Gene" and the "Molecular Gene": Two Relevant Concepts of Genetic Units

We focus here on two prevalent meanings of the word gene in research articles. On one hand, the gene, named here "molecular gene," is a stretch of DNA that is transcribed and codes for an RNA or a polypeptide with a known or presumed function (as in "gene network"), whose exact spatial delimitation on the chromosome remains a matter of debate, especially in cases with alternative splicing, antisense transcripts, etc. On the other hand, the gene, called here "Mendelian gene," is a segregating genetic unit which is detected through phenotypic differences associated with different alleles at the same locus (as in "gene flow"). We show that the "Mendelian gene" concept is still extensively used today in biology research and is sometimes confused with the "molecular gene." We try here to clarify the distinction between both concepts. Efforts to delineate the beginning and the end of the DNA sequence corresponding to the "Mendelian gene" and the "molecular gene" reveal that both entities do not always match. We argue that both concepts are part of two relevant frameworks for explaining the biological world.

High-order epistasis shapes evolutionary trajectories

High-order epistasis—where the effect of a mutation is determined by interactions with two or more other mutations—makes small, but detectable, contributions to genotype-fitness maps. While epistasis between pairs of mutations is known to be an important determinant of evolutionary trajectories, the evolutionary consequences of high-order epistasis remain poorly understood. To determine the effect of high-order epistasis on evolutionary trajectories, we computationally removed high-order epistasis from experimental genotype-fitness maps containing all binary combinations of five mutations. We then compared trajectories through maps both with and without high-order epistasis. We found that high-order epistasis strongly shapes the accessibility and probability of evolutionary trajectories. A closer analysis revealed that the magnitude of epistasis, not its order, predicts is effects on evolutionary trajectories. We further find that high-order epistasis makes it impossible to predict evolutionary trajectories from the individual and paired effects of mutations. We therefore conclude that high-order epistasis profoundly shapes evolutionary trajectories through genotype-fitness maps.

A key goal for evolutionary biologists is understanding why one evolutionary trajectory is taken rather than others. This requires understanding how individual mutations, as well as interactions between them, determine the accessibility of evolutionary pathways. We used a robust statistical analysis to reveal interactions between up to five mutations in published datasets, meaning that the effect of a mutation can depend on the presence or absence of four other mutations. Simulations reveal that these interactions strongly shape evolutionary trajectories. These interactions lead to profound unpredictability in evolution, as one cannot use the effect of a mutation in the ancestor to predict its effect later in the trajectory.

Ferguson L. SNP-SNP interactions as risk factors for aggressive prostate cancer

Prostate cancer (PCa) is one of the most significant male health concerns worldwide. Single nucleotide polymorphisms (SNPs) are becoming increasingly strong candidate biomarkers for identifying susceptibility to PCa. We identified a number of SNPs reported in genome-wide association analyses (GWAS) as risk factors for aggressive PCa in various European populations, and then defined SNP-SNP interactions, using PLINK software, with nucleic acid samples from a New Zealand cohort. We used this approach to find a gene x environment marker for aggressive PCa, as although statistically gene x environment interactions can be adjusted for, it is highly impossible in practicality, and thus must be incorporated in the search for a reliable biomarker for PCa. We found two intronic SNPs statistically significantly interacting with each other as a risk for aggressive prostate cancer on being compared to healthy controls in a New Zealand population.

Searching Genome-Wide Multi-Locus Associations for Multiple Diseases Based on Bayesian Inference

Taking the advantage of high-throughput single nucleotide polymorphism (SNP) genotyping technology, large genome-wide association studies (GWASs) have been considered to hold promise for unraveling complex relationships between genotypes and phenotypes. Current multi-locus-based methods are insufficient to detect interactions with diverse genetic effects on multifarious diseases. Also, statistic tests for high-order epistasis ( ≥ 2 SNPs) raise huge computational and analytical challenges because the computation increases exponentially as the growth of the cardinality of SNPs combinations. In this paper, we provide a simple, fast and powerful method, named DAM, using Bayesian inference to detect genome-wide multi-locus epistatic interactions in multiple diseases. Experimental results on simulated data demonstrate that our method is powerful and efficient. We also apply DAM on two GWAS datasets from WTCCC, i.e., Rheumatoid Arthritis and Type 1 Diabetes, and identify some novel findings. Therefore, we believe that our method is suitable and efficient for the full-scale analysis of multi-disease-related interactions in GWASs.

Gene-gene interaction of single nucleotide polymorphisms in 16p13.3 may contribute to the risk of non-syndromic cleft lip with or without cleft palate in Chinese case-parent trios

Non-syndromic cleft lip with or without cleft palate (NSCL/P) is a common birth defect with a complex and heterogeneous etiology. A recent genome-wide association study (GWAS) among Chinese populations has identified a new region at 16p13.3 as being associated with NSCL/P, which requires further replication. Here, we attempted to replicate and further clarify the genetic association between this region and NSCL/P, as well as testing for potential gene-gene (G × G) and gene-environment (G × E) interactions. We conducted transmission disequilibrium tests on 69 single nucleotide polymorphisms (SNPs) mapping to 16p13.3 among 806 Chinese case-parent trios ascertained through an international consortium where a GWAS of oral clefts was conducted. G × G, as well as G × E interactions involving maternal environmental tobacco smoke (ETS) and multivitamin supplementation, were explored using conditional logistic regression model. We applied Cordell's method as implemented in the R package TRIO to test for possible interactions. While no SNPs showed evidence of linkage and association with NSCL/P after Bonferroni correction, we found signals of G × G interactions between SNPs in 16p13.3. Nine pairs of SNP-SNP interactions attained significance after Bonferroni correction, among which the most significant interaction was found between rs2072346 (ADCY9) and rs11646137 (intergenic region, P = 7.2 × 10^-5 ). Linkage disequilibrium (LD) analysis revealed only low level of LD between these SNPs. This study failed to confirm the significant association between SNPs within 16p13.3 and the risk of NSCL/P, but underlined the importance of taking into account potential G × G interactions for the genetic association analysis of NSCL/P.

Statistical interactions and Bayes estimation of log odds in case-control studies

This paper is concerned with the estimation of the logarithm of disease odds (log odds) when evaluating two risk factors, whether or not interactions are present. Statisticians define interaction as a departure from an additive model on a certain scale of measurement of the outcome. Certain interactions, known as removable interactions, may be eliminated by fitting an additive model under an invertible transformation of the outcome. This can potentially provide more precise estimates of log odds than fitting a model with interaction terms. In practice, we may also encounter nonremovable interactions. The model must then include interaction terms, regardless of the choice of the scale of the outcome. However, in practical settings, we do not know at the outset whether an interaction exists, and if so whether it is removable or nonremovable. Rather than trying to decide on significance levels to test for the existence of removable and nonremovable interactions, we develop a Bayes estimator based on a squared error loss function. We demonstrate the favorable bias-variance trade-offs of our approach using simulations, and provide empirical illustrations using data from three published endometrial cancer case-control studies. The methods are implemented in an R program, and available freely at http://www.mskcc.org/biostatistics/~satagopj .

An empirical fuzzy multifactor dimensionality reduction method for detecting gene-gene interactions

BACKGROUND

Detection of gene-gene interaction (GGI) is a key challenge towards solving the problem of missing heritability in genetics. The multifactor dimensionality reduction (MDR) method has been widely studied for detecting GGIs. MDR reduces the dimensionality of multi-factor by means of binary classification into high-risk (H) or low-risk (L) groups. Unfortunately, this simple binary classification does not reflect the uncertainty of H/L classification. Thus, we proposed Fuzzy MDR to overcome limitations of binary classification by introducing the degree of membership of two fuzzy sets H/L. While Fuzzy MDR demonstrated higher power than that of MDR, its performance is highly dependent on the several tuning parameters. In real applications, it is not easy to choose appropriate tuning parameter values.

RESULT

In this work, we propose an empirical fuzzy MDR (EF-MDR) which does not require specifying tuning parameters values. Here, we propose an empirical approach to estimating the membership degree that can be directly estimated from the data. In EF-MDR, the membership degree is estimated by the maximum likelihood estimator of the proportion of cases(controls) in each genotype combination. We also show that the balanced accuracy measure derived from this new membership function is a linear function of the standard chi-square statistics. This relationship allows us to perform the standard significance test using p-values in the MDR framework without permutation. Through two simulation studies, the power of the proposed EF-MDR is shown to be higher than those of MDR and Fuzzy MDR. We illustrate the proposed EF-MDR by analyzing Crohn's disease (CD) and bipolar disorder (BD) in the Wellcome Trust Case Control Consortium (WTCCC) dataset.

CONCLUSION

We propose an empirical Fuzzy MDR for detecting GGI using the maximum likelihood of the proportion of cases(controls) as the membership degree of the genotype combination. The program written in R for EF-MDR is available at http://statgen.snu.ac.kr/software/EF-MDR .

Genetic Influences of OPRM1, OPRD1 and COMT on Morphine Analgesia in a Multi-Modal, Multi-Tissue Human Experimental Pain Model

Human studies on experimentally induced pain are of value to elucidate the genetic influence on morphine analgesia under controlled conditions. The aim of this study was to investigate whether genetic variants of mu-, kappa- and delta-opioid receptor genes (OPRM1, OPRK1 and OPRD1) and catechol-O-methyltransferase gene (COMT) are associated with the morphine analgesia. The study was a randomized, double-blind, two-way, crossover, single-dose study conducted in 40 healthy participants, where morphine was compared with placebo. Pain was induced by contact heat, muscle pressure, bone pressure, rectal stimulations (mechanical, electrical and thermal) and cold pressor test (immersion of the hand into ice water). Sixteen genetic polymorphisms of four candidate genes were explored. Variability in morphine analgesia to contact heat stimulation was associated with COMT rs4680 (p = 0.04), and rectal thermal stimulation was associated with OPRM1 rs9479757 (p = 0.03). Moreover, in males, variability in morphine analgesia to rectal thermal stimulation was associated with OPRD1 polymorphisms: rs2234918 (p = 0.01) and rs533123 (p = 0.046). The study was explorative and hypothesis-generating due to the relatively small study size. However, results suggest that genetic variants in the COMT and OPRM1 irrespective of gender, and OPRD1 in males may contribute to the variability in morphine analgesia in experimental pain models.

Interactions within the MHC contribute to the genetic architecture of celiac disease

Interaction analysis of GWAS can detect signal that would be ignored by single variant analysis, yet few robust interactions in humans have been detected. Recent work has highlighted interactions in the MHC region between known HLA risk haplotypes for various autoimmune diseases. To better understand the genetic interactions underlying celiac disease (CD), we have conducted exhaustive genome-wide scans for pairwise interactions in five independent CD case-control studies, using a rapid model-free approach to examine over 500 billion SNP pairs in total. We found 14 independent interaction signals within the MHC region that achieved stringent replication criteria across multiple studies and were independent of known CD risk HLA haplotypes. The strongest independent CD interaction signal corresponded to genes in the HLA class III region, in particular PRRC2A and GPANK1/C6orf47, which are known to contain variants for non-Hodgkin's lymphoma and early menopause, co-morbidities of celiac disease. Replicable evidence for statistical interaction outside the MHC was not observed. Both within and between European populations, we observed striking consistency of two-locus models and model distribution. Within the UK population, models of CD based on both interactions and additive single-SNP effects increased explained CD variance by approximately 1% over those of single SNPs. The interactions signal detected across the five cohorts indicates the presence of novel associations in the MHC region that cannot be detected using additive models. Our findings have implications for the determination of genetic architecture and, by extension, the use of human genetics for validation of therapeutic targets.

Gain of power of the general regression model compared to Cochran-Armitage Trend tests: simulation study and application to bipolar disorder

Background

Most genome-wide association studies assumed an additive model of inheritance which may result in significant loss of power when there is a strong departure from additivity. The General Regression Model (GRM), which allows performing an assumption-free test for association by testing for both additive effect and deviation from additive effect, may be more appropriate for association tests. Additionally, GRM allows testing the underlying genetic model. We compared the power of GRM association test to additive and other Cochran-Armitage Trend (CAT) tests through simulations and by applying GRM to a large case/control sample, the bipolar Welcome Trust Case Control Cohort data. Simulations were performed on two sets of case/control samples (1000/1000 and 2000/2000), using a large panel of genetic models. Four association tests (GRM and additive, recessive and dominant CAT tests) were applied to all replicates.

Results

We showed that GRM power to detect association was similar or greater than the additive CAT test, in particular in case of recessive inheritance, with up to 67% gain in power. GRM analysis of genome-wide bipolar disorder Welcome Trust Consortium data (1998 cases/3004 controls) showed significant association in the 16p12 region (rs420259; P = 3.4E-7) which has not been identified using the additive CAT test. As expected, rs42025 fitted a non-additive (recessive) model.

Conclusions

GRM provides increased power compared to the additive CAT test for association studies and is easily applicable.

Electronic supplementary material

The online version of this article (doi:10.1186/s12863-017-0486-6) contains supplementary material, which is available to authorized users.

High-density genetic map and identification of QTLs for responses to temperature and salinity stresses in the model brown alga Ectocarpus

Deciphering the genetic architecture of adaptation of brown algae to environmental stresses such as temperature and salinity is of evolutionary as well as of practical interest. The filamentous brown alga Ectocarpus sp. is a model for the brown algae and its genome has been sequenced. As sessile organisms, brown algae need to be capable of resisting the various abiotic stressors that act in the intertidal zone (e.g. osmotic pressure, temperature, salinity, UV radiation) and previous studies have shown that an important proportion of the expressed genes is regulated in response to hyposaline, hypersaline or oxidative stress conditions. Using the double digest RAD sequencing method, we constructed a dense genetic map with 3,588 SNP markers and identified 39 QTLs for growth-related traits and their plasticity under different temperature and salinity conditions (tolerance to high temperature and low salinity). GO enrichment tests within QTL intervals highlighted membrane transport processes such as ion transporters. Our study represents a significant step towards deciphering the genetic basis of adaptation of Ectocarpus sp. to stress conditions and provides a substantial resource to the increasing list of tools generated for the species.

Epistasis and the Evolution of Antimicrobial Resistance

The fitness effects of a mutation can depend, sometimes dramatically, on genetic background; this phenomenon is often referred to as “epistasis.” Epistasis can have important practical consequences in the context of antimicrobial resistance (AMR). For example, genetic background plays an important role in determining the costs of resistance, and hence in whether resistance will persist in the absence of antibiotic pressure. Furthermore, interactions between resistance mutations can have important implications for the evolution of multi-drug resistance. I argue that there is a need to better characterize the extent and nature of epistasis for mutations and horizontally transferred elements conferring AMR, particularly in clinical contexts. Furthermore, I suggest that epistasis should be an important consideration in attempts to slow or limit the evolution of AMR.

Detecting High-Order Epistasis in Nonlinear Genotype-Phenotype Maps

High-order epistasis has been observed in many genotype-phenotype maps. These multi-way interactions between mutations may be useful for dissecting complex traits and could have profound implications for evolution. Alternatively, they could be a statistical artifact. High-order epistasis models assume the effects of mutations should add, when they could in fact multiply or combine in some other nonlinear way. A mismatch in the “scale” of the epistasis model and the scale of the underlying map would lead to spurious epistasis. In this article, we develop an approach to estimate the nonlinear scales of arbitrary genotype-phenotype maps. We can then linearize these maps and extract high-order epistasis. We investigated seven experimental genotype-phenotype maps for which high-order epistasis had been reported previously. We find that five of the seven maps exhibited nonlinear scales. Interestingly, even after accounting for nonlinearity, we found statistically significant high-order epistasis in all seven maps. The contributions of high-order epistasis to the total variation ranged from 2.2 to 31.0%, with an average across maps of 12.7%. Our results provide strong evidence for extensive high-order epistasis, even after nonlinear scale is taken into account. Further, we describe a simple method to estimate and account for nonlinearity in genotype-phenotype maps.

Consequences of Epistasis on Growth in an Erhualian × White Duroc Pig Cross

Epistasis describes an interaction between the effects of loci. We included epistasis in quantitative trait locus (QTL) mapping of growth at a series of ages in a cross of a Chinese pig breed, Erhualian, with a commercial line, White Duroc. Erhualian pigs have much lower growth rates than White Duroc. We improved a method for genomewide testing of epistasis and present a clear analysis workflow. We also suggest a new approach for interpreting epistasis results where significant additive and dominance effects of a locus in specific backgrounds are determined. In total, seventeen QTL were found and eleven showed epistasis. Loci on chromosomes 2, 3, 4 and 7 were highlighted as affecting growth at more than one age or forming an interaction network. Epistasis resulted in both the QTL on chromosomes 3 and 7 having effects in opposite directions. We believe it is the first time for the chromosome 7 locus that an allele from a Chinese breed has been found to decrease growth. The consequences of epistasis were diverse. Results were impacted by using growth rather than body weight as the phenotype and by correcting for an effect of mother. Epistasis made a considerable contribution to growth in this population and modelling epistasis was important for accurately determining QTL effects.

Evaluating Additive Interaction Using Survival Percentiles

Evaluation of statistical interaction in time-to-event analysis is usually limited to the study of multiplicative interaction, via inclusion of a product term in a Cox proportional-hazard model. Measures of additive interaction are available but seldom used. All measures of interaction in survival analysis, whether additive or multiplicative, are in the metric of hazard, usually assuming that the interaction between two predictors of interest is constant during the follow-up period. We introduce a measure to evaluate additive interaction in survival analysis in the metric of time. This measure can be calculated by evaluating survival percentiles, defined as the time points by which different subpopulations reach the same incidence proportion. Using this approach, the probability of the outcome is fixed and the time variable is estimated. We also show that by using a regression model for the evaluation of conditional survival percentiles, including a product term between the two exposures in the model, interaction is evaluated as a deviation from additivity of the effects. In the simple case of two binary exposures, the product term is interpreted as excess/decrease in survival time (i.e., years, months, days) due to the presence of both exposures. This measure of interaction is dependent on the fraction of events being considered, thus allowing evaluation of how interaction changes during the observed follow-up. Evaluation of interaction in the context of survival percentiles allows deriving a measure of additive interaction without assuming a constant effect over time, overcoming two main limitations of commonly used approaches.

Evolutionary interplay between structure, energy and epistasis in the coat protein of the ϕX174 phage family

Viral capsids are structurally constrained by interactions among the amino acids (AAs) of their constituent proteins. Therefore, epistasis is expected to evolve among physically interacting sites and to influence the rates of substitution. To study the evolution of epistasis, we focused on the major structural protein of the ϕX174 phage family by first reconstructing the ancestral protein sequences of 18 species using a Bayesian statistical framework. The inferred ancestral reconstruction differed at eight AAs, for a total of 256 possible ancestral haplotypes. For each ancestral haplotype and the extant species, we estimated, in silico, the distribution of free energies and epistasis of the capsid structure. We found that free energy has not significantly increased but epistasis has. We decomposed epistasis up to fifth order and found that higher-order epistasis sometimes compensates pairwise interactions making the free energy seem additive. The dN/dS ratio is low, suggesting strong purifying selection, and that structure is under stabilizing selection. We synthesized phages carrying ancestral haplotypes of the coat protein gene and measured their fitness experimentally. Our findings indicate that stabilizing mutations can have higher fitness, and that fitness optima do not necessarily coincide with energy minima.

Evidence for SNP-SNP interaction identified through targeted sequencing of cleft case-parent trios

Nonsyndromic cleft lip with or without cleft palate (NSCL/P) is the most common craniofacial birth defect in humans, affecting 1 in 700 live births. This malformation has a complex etiology where multiple genes and several environmental factors influence risk. At least a dozen different genes have been confirmed to be associated with risk of NSCL/P in previous studies. However, all the known genetic risk factors cannot fully explain the observed heritability of NSCL/P, and several authors have suggested gene-gene (G × G) interaction may be important in the etiology of this complex and heterogeneous malformation. We tested for G × G interactions using common single nucleotide polymorphic (SNP) markers from targeted sequencing in 13 regions identified by previous studies spanning 6.3 Mb of the genome in a study of 1,498 NSCL/P case-parent trios. We used the R-package trio to assess interactions between polymorphic markers in different genes, using a 1 degree of freedom (1df) test for screening, and a 4 degree of freedom (4df) test to assess statistical significance of epistatic interactions. To adjust for multiple comparisons, we performed permutation tests. The most significant interaction was observed between rs6029315 in MAFB and rs6681355 in IRF6 (4df P = 3.8 × 10^-8 ) in case-parent trios of European ancestry, which remained significant after correcting for multiple comparisons. However, no significant interaction was detected in trios of Asian ancestry.

Bayesian reversible-jump for epistasis analysis in genomic studies

Background

The large amount of data used in genomic analysis has allowed geneticists to achieve some understanding of the genetic architecture of complex traits. Although the information gathered by molecular markers has permitted gains in predictive accuracy and gene discovery, epistatic effects have been ignored based on exhaustive searches requesting estimates of its effects on the whole genome. In this work, we propose the reversible-jump technique to estimate epistasis in the genome without drastically altering the model dimension. To this end, we used a real maize dataset based on 256 F_2:3 progenies plus a simulation data set based on 300 F₂ individuals. In the simulation scenario, six QTL presenting main effects (additive and dominance) were combined with seven other epistatic effects totaling 13 QTL controlling the trait.

Results

Our model explored 18,624 candidate epistases, but even in this vast space, only one spurious interaction was found. The three epistases selected by our model, named here as 18x26, 56x68 and 59x93, were very close to simulated ones (19x25, 54x72, 59x91 and 59x94). In the real dataset, we estimate 33,024 epistatic effects, and several minor epistatic combinations were found to explain a significant proportion of the genetic variance. The broad participation of epistasis in the real dataset may indicate the presence of pervasive epistasis acting on maize grain yield.

Conclusions

The power of selecting true epistasis in thousands of possible combinations suggests the attractiveness of our model to handle genomic data

Electronic supplementary material

The online version of this article (doi:10.1186/s12864-016-3342-6) contains supplementary material, which is available to authorized users.

Influences of XDH genotype by gene-gene interactions with SUCLA2 for thiopurine-induced leukopenia in Korean patients with Crohn's disease

BACKGROUND

The impact of genetic variation in the thiopurine S-methyltransferase (TPMT) gene on thiopurine-induced leukopenia has been well demonstrated. Although xanthine dehydrogenase (XDH) is the second major contributor to azathioprine breakdown, polymorphisms in XDH have rarely been studied in IBD patients. We aim to access association between XDH variants and thiopurine-induced leukopenia by gene-gene interaction in a Crohn's disease (CD) population.

STUDY

A total of 964 CD patients treated with thiopurines were recruited from a tertiary referral center. The association between four XDH variants (p.Gly172Arg, p.Asn1109Thr, p.Arg149Cys, and p.Thr910Lys) and thiopurine-induced leukopenia was analyzed in cases with early leukopenia (n = 66), late leukopenia (n = 264), and in controls without leukopenia (n = 632). Three non-synonymous SNPs, which we previously reported association with thiopurine-induced leukopenia, NUDT15 (p.Arg139Cys), SUCLA2 (p.Ser199Thr), and TPMT *3C were selected for epistasis analysis with the XDH variants.

RESULTS

There was no significant association for two variants of XDH and thiopurine-induced leukopenia. In the epistasis analysis, only XDH (p.Asn1109Thr) * SUCLA2 (p.Ser199Thr) showed a statistically significant association with early leukopenia [odds ratio (OR) = 0.16; p = 0.03]. After genotype stratification, a positive association on the background of SUCLA2 wild-type (199Ser) between the XDH (p.Asn1109Thr) and early leukopenia (OR = 4.39; p = 0.01) was detected.

CONCLUSION

Genes associated with thiopurine-induced leukopenia can act in a complex interactive manner. Further studies are warranted to explore the mechanisms underlying the effects of the combination of XDH (p.Asn1109Thr) and SUCLA2 (199Ser) on thiopurine-induced leukopenia.

Are Interactions between cis-Regulatory Variants Evidence for Biological Epistasis or Statistical Artifacts?

The importance of epistasis-or statistical interactions between genetic variants-to the development of complex disease in humans has been controversial. Genome-wide association studies of statistical interactions influencing human traits have recently become computationally feasible and have identified many putative interactions. However, statistical models used to detect interactions can be confounded, which makes it difficult to be certain that observed statistical interactions are evidence for true molecular epistasis. In this study, we investigate whether there is evidence for epistatic interactions between genetic variants within the cis-regulatory region that influence gene expression after accounting for technical, statistical, and biological confounding factors. We identified 1,119 (FDR = 5%) interactions that appear to regulate gene expression in human lymphoblastoid cell lines, a tightly controlled, largely genetically determined phenotype. Many of these interactions replicated in an independent dataset (90 of 803 tested, Bonferroni threshold). We then performed an exhaustive analysis of both known and novel confounders, including ceiling/floor effects, missing genotype combinations, haplotype effects, single variants tagged through linkage disequilibrium, and population stratification. Every interaction could be explained by at least one of these confounders, and replication in independent datasets did not protect against some confounders. Assuming that the confounding factors provide a more parsimonious explanation for each interaction, we find it unlikely that cis-regulatory interactions contribute strongly to human gene expression, which calls into question the relevance of cis-regulatory interactions for other human phenotypes. We additionally propose several best practices for epistasis testing to protect future studies from confounding.

Evidence for epistatic interaction between VDR and SLC13A2 genes in the pathogenesis of hypocitraturia in recurrent calcium oxalate stone formers

BACKGROUND

Genetic factors play a key role in the pathogenesis of hypocitraturia, a common risk factor for nephrolithiasis. The Na⁺-dicarboxylate cotransporter NaDC1, encoded by the sodium-dicarboxylate cotransporter (SLC13A2) gene, is a major determinant of urinary citrate excretion and its biological functions are regulated also by the vitamin D/Vitamin D receptor (VDR) biological system. The aim of this case-control study was to evaluate the possible epistatic interaction between VDR ^rs731236and SLC13A2 ^rs11567842 allelic variants in the pathogenesis of hypocitraturia.

METHODS

Recurrent calcium-oxalate stone formers (SF) with or without hypocitraturia and healthy controls (C) were genotyped. Gene-gene interactions were estimated by the 1.0 software package of multifactor dimensionality reduction (MDR).

RESULTS

The prevalence of VDR ^TT and SLC13A2 ^GG genotypes was higher in hypocitraturic SF compared to C (odds ratio [OR] 3.24, 95 % confidence interval [CI] 1.38-7.60 for VDR ^TT vs. VDR ^tt and OR 4.06, 95 % CI 1.75-9.42 for SLC13A2 ^GG vs. SLC13A2 ^AA ). MDR analysis indicated a significant interaction between VDR ^TT and SLC13A2 ^GG in hypocitraturic SF compared to C [OR 3.81 (2.11-6.88)]. These data are compatible with an epistatic interaction between the VDR ^TT and SLC13A2 ^GG genotypes with a significant impact on the magnitude of the effect (suppressive effect).

CONCLUSIONS

These results point to an epistatic interaction between the VDR and the SLC13A2 alleles in the pathogenesis of idiopathic hypocitraturia in calcium-oxalate SF.

Genetic Influences on Cerebral Outcome After Cardiac Surgery

Just as the influence of genetic variation on patient outcomes is being discussed in many other areas of medicine, so too are its effects on cerebral outcome after cardiac surgery now being described. Whereas early studies focused on neurocognitive outcome, where the single nucleotide polymorphisms of APOE4 and PLA2 were the first investigated genetic targets, stroke is now being elaborated on with related single and multi-gene single nucleotide polymorphisms having been identified. Our work has established key links between post-cardiac surgery stroke and C-reactive protein (3’UTR 1846C/T) and interleukin-6 (-174 G/C) single nucleotide polymorphisms.

Exhaustive Genome-Wide Search for SNP-SNP Interactions Across 10 Human Diseases

The identification of statistical SNP-SNP interactions may help explain the genetic etiology of many human diseases, but exhaustive genome-wide searches for these interactions have been difficult, due to a lack of power in most datasets. We aimed to use data from the Resource for Genetic Epidemiology Research on Adult Health and Aging (GERA) study to search for SNP-SNP interactions associated with 10 common diseases. FastEpistasis and BOOST were used to evaluate all pairwise interactions among approximately N = 300,000 single nucleotide polymorphisms (SNPs) with minor allele frequency (MAF) ≥ 0.15, for the dichotomous outcomes of allergic rhinitis, asthma, cardiac disease, depression, dermatophytosis, type 2 diabetes, dyslipidemia, hemorrhoids, hypertensive disease, and osteoarthritis. A total of N = 45,171 subjects were included after quality control steps were applied. These data were divided into discovery and replication subsets; the discovery subset had > 80% power, under selected models, to detect genome-wide significant interactions (P < 10(-12)). Interactions were also evaluated for enrichment in particular SNP features, including functionality, prior disease relevancy, and marginal effects. No interaction in any disease was significant in both the discovery and replication subsets. Enrichment analysis suggested that, for some outcomes, interactions involving SNPs with marginal effects were more likely to be nominally replicated, compared to interactions without marginal effects. If SNP-SNP interactions play a role in the etiology of the studied conditions, they likely have weak effect sizes, involve lower-frequency variants, and/or involve complex models of interaction that are not captured well by the methods that were utilized.

Genome-wide search identifies a gene-gene interaction between 20p13 and 2q14 in asthma

Background

Many studies have attempted to identify gene-gene interactions affecting asthma susceptibility. However, these studies have typically used candidate gene approaches in limiting the genetic search space, and there have been few searches for gene-gene interactions on a genome-wide scale. We aimed to conduct a genome-wide gene-gene interaction study for asthma, using data from the GABRIEL Consortium.

Results

A two-stage study design was used, including a screening analysis (N = 1625 subjects) and a follow-up analysis (N = 5264 subjects). In the screening analysis, all pairwise interactions among 301,547 SNPs were evaluated, encompassing a total of 4.55 × 10¹⁰ interactions. Those with a screening interaction p-value < 10⁻⁵ were evaluated in the follow-up analysis. No interaction selected from the screening analysis met strict statistical significance in the follow-up (p-value < 1.45 × 10⁻⁷). However, the top-ranked interaction (rs910652 [20p13] × rs11684871 [2q14]) in the follow-up (p-value = 1.58 × 10⁻⁶) was significant in one component of a replication analysis. This interaction was notable in that rs910652 is located within 78 kilobases of ADAM33, which is one of the most well studied asthma susceptibility genes. In addition, rs11684871 is located in or near GLI2, which may have biologically relevant roles in asthma.

Conclusions

Using a genome-wide approach, we identified and found suggestive evidence of replication for a gene-gene interaction in asthma involving loci that are potentially highly relevant in asthma pathogenesis.

Electronic supplementary material

The online version of this article (doi:10.1186/s12863-016-0376-3) contains supplementary material, which is available to authorized users.

Multikernel linear mixed models for complex phenotype prediction

Linear mixed models (LMMs) and their extensions have recently become the method of choice in phenotype prediction for complex traits. However, LMM use to date has typically been limited by assuming simple genetic architectures. Here, we present multikernel linear mixed model (MKLMM), a predictive modeling framework that extends the standard LMM using multiple-kernel machine learning approaches. MKLMM can model genetic interactions and is particularly suitable for modeling complex local interactions between nearby variants. We additionally present MKLMM-Adapt, which automatically infers interaction types across multiple genomic regions. In an analysis of eight case-control data sets from the Wellcome Trust Case Control Consortium and more than a hundred mouse phenotypes, MKLMM-Adapt consistently outperforms competing methods in phenotype prediction. MKLMM is as computationally efficient as standard LMMs and does not require storage of genotypes, thus achieving state-of-the-art predictive power without compromising computational feasibility or genomic privacy.

Construction of synergy networks from gene expression data related to disease

A few methods have been developed to determine whether genes collaborate with each other in relation to a particular disease using an information theoretic measure of synergy. Here, we propose an alternative definition of synergy and justify that our definition improves upon the existing measures of synergy in the context of gene interactions. We use this definition on a prostate cancer data set consisting of gene expression levels in both cancerous and non-cancerous samples and identify pairs of genes which are unable to discriminate between cancerous and non-cancerous samples individually but can do so jointly when we take their synergistic property into account. We also propose a very simple yet effective technique for computation of conditional entropy at a very low cost. The worst case complexity of our method is O(n) while the best case complexity of a state-of-the-art method is O(n(2)). Furthermore, our method can also be extended to find synergistic relation among triplets or even among a larger number of genes. Finally, we validate our results by demonstrating that these findings cannot be due to pure chance and provide the relevance of the synergistic pairs in cancer biology.

CINOEDV: a co-information based method for detecting and visualizing n-order epistatic interactions

BACKGROUND

Detecting and visualizing nonlinear interaction effects of single nucleotide polymorphisms (SNPs) or epistatic interactions are important topics in bioinformatics since they play an important role in unraveling the mystery of "missing heritability". However, related studies are almost limited to pairwise epistatic interactions due to their methodological and computational challenges.

RESULTS

We develop CINOEDV (Co-Information based N-Order Epistasis Detector and Visualizer) for the detection and visualization of epistatic interactions of their orders from 1 to n (n ≥ 2). CINOEDV is composed of two stages, namely, detecting stage and visualizing stage. In detecting stage, co-information based measures are employed to quantify association effects of n-order SNP combinations to the phenotype, and two types of search strategies are introduced to identify n-order epistatic interactions: an exhaustive search and a particle swarm optimization based search. In visualizing stage, all detected n-order epistatic interactions are used to construct a hypergraph, where a real vertex represents the main effect of a SNP and a virtual vertex denotes the interaction effect of an n-order epistatic interaction. By deeply analyzing the constructed hypergraph, some hidden clues for better understanding the underlying genetic architecture of complex diseases could be revealed.

CONCLUSIONS

Experiments of CINOEDV and its comparison with existing state-of-the-art methods are performed on both simulation data sets and a real data set of age-related macular degeneration. Results demonstrate that CINOEDV is promising in detecting and visualizing n-order epistatic interactions. CINOEDV is implemented in R and is freely available from R CRAN: http://cran.r-project.org and https://sourceforge.net/projects/cinoedv/files/ .

Association between Gene Polymorphisms and Pain Sensitivity Assessed in a Multi-Modal Multi-Tissue Human Experimental Model - An Explorative Study

The genetic influence on sensitivity to noxious stimuli (pain sensitivity) remains controversial and needs further investigation. In the present study, the possible influence of polymorphisms in three opioid receptor (OPRM, OPRD and OPRK) genes and the catechol-O-methyltransferase (COMT) gene on pain sensitivity in healthy participants was investigated. Catechol-O-methyltransferase has an indirect effect on the mu opioid receptor by changing its activity through an altered endogenous ligand effect. Blood samples for genetic analysis were withdrawn in a multi-modal and multi-tissue experimental pain model in 40 healthy participants aged 20-65. Seventeen different single nucleotide polymorphisms in different genes (OPRM, OPRK, OPRD and COMT) were included in the analysis. Experimental pain tests included thermal skin stimulation, mechanical muscle and bone stimulation and mechanical, electrical and thermal visceral stimulations. A cold pressor test was also conducted. DNA was available from 38 of 40 participants. Compared to non-carriers of the COMT rs4680A allele, carriers reported higher bone pressure pain tolerance threshold (i.e. less pain) by up to 23.8% (p < 0.015). Additionally, carriers of the C allele (CC/CT) of OPRK rs6473799 reported a 30.4% higher mechanical visceral pain tolerance threshold than non-carriers (TT; p < 0.019). For the other polymorphisms and stimulations, no associations were found (all p > 0.05). In conclusion, COMT rs4680 and OPRK rs6473799 polymorphisms seem to be associated with pain sensitivity. Thus, the findings support a possible genetic influence on pain sensitivity.

Functional Regression Models for Epistasis Analysis of Multiple Quantitative Traits

To date, most genetic analyses of phenotypes have focused on analyzing single traits or analyzing each phenotype independently. However, joint epistasis analysis of multiple complementary traits will increase statistical power and improve our understanding of the complicated genetic structure of the complex diseases. Despite their importance in uncovering the genetic structure of complex traits, the statistical methods for identifying epistasis in multiple phenotypes remains fundamentally unexplored. To fill this gap, we formulate a test for interaction between two genes in multiple quantitative trait analysis as a multiple functional regression (MFRG) in which the genotype functions (genetic variant profiles) are defined as a function of the genomic position of the genetic variants. We use large-scale simulations to calculate Type I error rates for testing interaction between two genes with multiple phenotypes and to compare the power with multivariate pairwise interaction analysis and single trait interaction analysis by a single variate functional regression model. To further evaluate performance, the MFRG for epistasis analysis is applied to five phenotypes of exome sequence data from the NHLBI’s Exome Sequencing Project (ESP) to detect pleiotropic epistasis. A total of 267 pairs of genes that formed a genetic interaction network showed significant evidence of epistasis influencing five traits. The results demonstrate that the joint interaction analysis of multiple phenotypes has a much higher power to detect interaction than the interaction analysis of a single trait and may open a new direction to fully uncovering the genetic structure of multiple phenotypes.

The widely used statistical methods test interaction for single phenotype. However, we often observe pleotropic genetic interaction effects. The simultaneous gene-gene (GxG) interaction analysis of multiple complementary traits will increase statistical power to detect GxG interactions. Although GxG interactions play an important role in uncovering the genetic structure of complex traits, the statistical methods for detecting GxG interactions in multiple phenotypes remains less developed owing to its potential complexity. Therefore, we extend functional regression model from single variate to multivariate for simultaneous GxG interaction analysis of multiple correlated phenotypes. Large-scale simulations are conducted to evaluate Type I error rates for testing interaction between two genes with multiple phenotypes and to compare power with traditional multivariate pair-wise interaction analysis and single trait interaction analysis by a single variate functional regression model. To further evaluate performance, the MFRG for interaction analysis is applied to five phenotypes of exome sequence data from the NHLBI’s Exome Sequencing Project (ESP) to detect pleiotropic GxG interactions. 267 pairs of genes that formed a genetic interaction network showed significant evidence of interactions influencing five traits.

An interaction quantitative trait loci tool implicates epistatic functional variants in an apoptosis pathway in smallpox vaccine eQTL data

Expression quantitative trait loci (eQTL) studies have functionalized nucleic acid variants through the regulation of gene expression. Although most eQTL studies only examine the effects of single variants on transcription, a more complex process of variant-variant interaction (epistasis) may regulate transcription. Herein, we describe a tool called interaction QTL (iQTL) designed to efficiently detect epistatic interactions that regulate gene expression. To maximize biological relevance and minimize the computational and hypothesis testing burden, iQTL restricts interactions such that one variant is within a user-defined proximity of the transcript (cis-regulatory). We apply iQTL to a data set of 183 smallpox vaccine study participants with genome-wide association study and gene expression data from unstimulated samples and samples stimulated by inactivated vaccinia virus. While computing only 0.15% of possible interactions, we identify 11 probe sets whose expression is regulated through a variant-variant interaction. We highlight the functional epistatic interactions among apoptosis-related genes, DIABLO, TRAPPC4 and FADD, in the context of smallpox vaccination. We also use an integrative network approach to characterize these iQTL interactions in a posterior network of known prior functional interactions. iQTL is an efficient, open-source tool to analyze variant interactions in eQTL studies, providing better understanding of the function of epistasis in immune response and other complex phenotypes.

AGGrEGATOr: A Gene-based GEne-Gene interActTiOn test for case-control association studies

Among the large of number of statistical methods that have been proposed to identify gene-gene interactions in case-control genome-wide association studies (GWAS), gene-based methods have recently grown in popularity as they confer advantage in both statistical power and biological interpretation. All of the gene-based methods jointly model the distribution of single nucleotide polymorphisms (SNPs) sets prior to the statistical test, leading to a limited power to detect sums of SNP-SNP signals. In this paper, we instead propose a gene-based method that first performs SNP-SNP interaction tests before aggregating the obtained p-values into a test at the gene level. Our method called AGGrEGATOr is based on a minP procedure that tests the significance of the minimum of a set of p-values. We use simulations to assess the capacity of AGGrEGATOr to correctly control for type-I error. The benefits of our approach in terms of statistical power and robustness to SNPs set characteristics are evaluated in a wide range of disease models by comparing it to previous methods. We also apply our method to detect gene pairs associated to rheumatoid arthritis (RA) on the GSE39428 dataset. We identify 13 potential gene-gene interactions and replicate one gene pair in the Wellcome Trust Case Control Consortium dataset at the level of 5%. We further test 15 gene pairs, previously reported as being statistically associated with RA or Crohn's disease (CD) or coronary artery disease (CAD), for replication in the Wellcome Trust Case Control Consortium dataset. We show that AGGrEGATOr is the only method able to successfully replicate seven gene pairs.

Using propensity score adjustment method in genetic association studies

BACKGROUND

The statistical tests for single locus disease association are mostly under-powered. If a disease associated causal single nucleotide polymorphism (SNP) operates essentially through a complex mechanism that involves multiple SNPs or possible environmental factors, its effect might be missed if the causal SNP is studied in isolation without accounting for these unknown genetic influences. In this study, we attempt to address the issue of reduced power that is inherent in single point association studies by accounting for genetic influences that negatively impact the detection of causal variant in single point association analysis. In our method we use propensity score (PS) to adjust for the effect of SNPs that influence the marginal association of a candidate marker. These SNPs might be in linkage disequilibrium (LD) and/or epistatic with the target-SNP and have a joint interactive influence on the disease under study. We therefore propose a propensity score adjustment method (PSAM) as a tool for dimension reduction to improve the power for single locus studies through an estimated PS to adjust for influence from these SNPs while regressing disease status on the target-genetic locus. The degree of freedom of such a test is therefore always restricted to 1.

RESULTS

We assess PSAM under the null hypothesis of no disease association to affirm that it correctly controls for the type-I-error rate (<0.05). PSAM displays reasonable power (>70%) and shows an average of 15% improvement in power as compared with commonly-used logistic regression method and PLINK under most simulated scenarios. Using the open-access multifactor dimensionality reduction dataset, PSAM displays improved significance for all disease loci. Through a whole genome study, PSAM was able to identify 21 SNPs from the GAW16 NARAC dataset by reducing their original trend-test p-values from within 0.001 and 0.05 to p-values less than 0.0009, and among which 6 SNPs were further found to be associated with immunity and inflammation.

CONCLUSIONS

PSAM improves the significance of single-locus association of causal SNPs which have had marginal single point association by adjusting for influence from other SNPs in a dataset. This would explain part of the missing heritability without increasing the complexity of the model due to huge multiple testing scenarios. The newly reported SNPs from GAW16 data would provide evidences for further research to elucidate the etiology of rheumatoid arthritis. PSAM is proposed as an exploratory tool that would be complementary to other existing methods. A downloadable user friendly program, PSAM, written in SAS, is available for public use.

A Novel Test for Detecting SNP-SNP Interactions in Case-Only Trio Studies

Epistasis plays a significant role in the genetic architecture of many complex phenotypes in model organisms. To date, there have been very few interactions replicated in human studies due in part to the multiple-hypothesis burden implicit in genome-wide tests of epistasis. Therefore, it is of paramount importance to develop the most powerful tests possible for detecting interactions. In this work we develop a new SNP-SNP interaction test for use in case-only trio studies called the trio correlation (TC) test. The TC test computes the expected joint distribution of marker pairs in offspring conditional on parental genotypes. This distribution is then incorporated into a standard 1 d.f. correlation test of interaction. We show via extensive simulations under a variety of disease models that our test substantially outperforms existing tests of interaction in case-only trio studies. We also demonstrate a bias in a previous case-only trio interaction test and identify its origin. Finally, we show that a previously proposed permutation scheme in trio studies mitigates the known biases of case-only tests in the presence of population stratification. We conclude that the TC test shows improved power to identify interactions in existing, as well as emerging, trio association studies. The method is publicly available at www.github.com/BrunildaBalliu/TrioEpi.

From Single Variants to Protein Cascades: MULTISCALE MODELING OF SINGLE NUCLEOTIDE VARIANT SETS IN GENETIC DISORDERS

Understanding the role of genetics in disease has become a central part of medical research. Non-synonymous single nucleotide variants (nsSNVs) in coding regions of human genes frequently lead to pathological phenotypes. Beyond single variations, the individual combination of nsSNVs may add to pathogenic processes. We developed a multiscale pipeline to systematically analyze the existence of quantitative effects of multiple nsSNVs and gene combinations in single individuals on pathogenicity. Based on this pipeline, we detected in a data set of 842 nsSNVs discovered in 76 genes related to cardiomyopathies, associated nsSNV combinations in seven genes present in at least 70% of all 639 patient samples, but not in a control cohort of healthy humans. Structural analyses of these revealed primarily an influence on the protein stability. For amino acid substitutions located at the protein surface, we generally observed a proximity to putative binding pockets. To computationally analyze cumulative effects and their impact, pathogenicity methods are currently being developed. Our approach supports this process, as shown on the example of a cardiac phenotype but can be likewise applied to other diseases such as cancer.

A Mutation in DAOA Modifies the Age of Onset in PSEN1 E280A Alzheimer's Disease

We previously reported age of onset (AOO) modifier genes in the world's largest pedigree segregating early-onset Alzheimer's disease (AD), caused by the p.Glu280Ala (E280A) mutation in the PSEN1 gene. Here we report the results of a targeted analysis of functional exonic variants in those AOO modifier genes in sixty individuals with PSEN1 E280A AD who were whole-exome genotyped for ~250,000 variants. Standard quality control, filtering, and annotation for functional variants were applied, and common functional variants located in those previously reported as AOO modifier loci were selected. Multiloci linear mixed-effects models were used to test the association between these variants and AOO. An exonic missense mutation in the G72 (DAOA) gene (rs2391191, P = 1.94 × 10⁻⁴, P_FDR = 9.34 × 10⁻³) was found to modify AOO in PSEN1 E280A AD. Nominal associations of missense mutations in the CLUAP1 (rs9790, P = 7.63 × 10⁻³, P_FDR = 0.1832) and EXOC2 (rs17136239, P = 0.0325, P_FDR = 0.391) genes were also found. Previous studies have linked polymorphisms in the DAOA gene with the occurrence of neuropsychiatric symptoms such as depression, apathy, aggression, delusions, hallucinations, and psychosis in AD. Our findings strongly suggest that this new conspicuous functional AOO modifier within the G72 (DAOA) gene could be pivotal for understanding the genetic basis of AD.

Association genetics in Populus reveals the interactions between Pto-miR160a and its target Pto-ARF16

MicroRNAs (miRNAs) play important roles in the regulation of gene expression in various biological processes. However, the interactions between miRNAs and their targets are largely unknown in plants. As a powerful tool for identification of variation associated with traits, association genetics provides another strategy for exploration of interactions between miRNAs and their targets. Here, we conducted expression analysis and association mapping to evaluate the interaction between Pto-miR160a and its target Pto-ARF16 in Populus tomentosa. By examining the expression patterns of Pto-MIR160a and Pto-ARF16, we identified a significant, negative correlation between their expression levels, indicating that Pto-miR160a may affect the expression of Pto-ARF16. Among the single nucleotide polymorphisms (SNPs) identified in this study, one common SNP in the pre-miRNA region of Pto-miR160a altered its predicted secondary structure while another common SNP in the predicted miRNA target site changed the binding affinity of Pto-miR160a. Linkage disequilibrium (LD) analysis revealed low LD levels of Pto-MIR160a and Pto-ARF16, indicating that they are suitable for candidate gene-based association analysis. Single SNP-based association analysis identified 19 SNPs (false discovery rate Q < 0.05) in Pto-MIR160a and Pto-ARF16 associated with three phenotypic traits. Epistasis analysis further identified 36 SNP-SNP interactions between SNPs in Pto-MIR160a and SNPs in Pto-ARF16, reflecting the possible genetic interaction of Pto-miR160a and Pto-ARF16. Taking these results together, our study identified SNPs in Pto-MIR160a and Pto-ARF16 associated with tree growth and wood properties, providing SNPs with potential applications in marker-assisted breeding and evidence for the genetic interaction of Pto-miR160a and Pto-ARF16.

A Framework for Interpreting Type I Error Rates from a Product-Term Model of Interaction Applied to Quantitative Traits

Adequate control of type I error rates will be necessary in the increasing genome-wide search for interactive effects on complex traits. After observing unexpected variability in type I error rates from SNP-by-genome interaction scans, we sought to characterize this variability and test the ability of heteroskedasticity-consistent standard errors to correct it. We performed 81 SNP-by-genome interaction scans using a product-term model on quantitative traits in a sample of 1,053 unrelated European Americans from the NHLBI Family Heart Study, and additional scans on five simulated datasets. We found that the interaction-term genomic inflation factor (lambda) showed inflation and deflation that varied with sample size and allele frequency; that similar lambda variation occurred in the absence of population substructure; and that lambda was strongly related to heteroskedasticity but not to minor non-normality of phenotypes. Heteroskedasticity-consistent standard errors narrowed the range of lambda, with HC3 outperforming HC0, but in individual scans tended to create new P-value outliers related to sparse two-locus genotype classes. We explain the lambda variation as a result of non-independence of test statistics coupled with stochastic biases in test statistics due to a failure of the test to reach asymptotic properties. We propose that one way to interpret lambda is by comparison to an empirical distribution generated from data simulated under the null hypothesis and without population substructure. We further conclude that the interaction-term lambda should not be used to adjust test statistics and that heteroskedasticity-consistent standard errors come with limitations that may outweigh their benefits in this setting.