Epistasis, complexity, and multifactor dimensionality reduction

Common genetic variants associated with melanoma risk or naevus count in patients with wildtype MC1R melanoma

BACKGROUND

Hypomorphic MC1R variants are the most prevalent genetic determinants of melanoma risk in the white population. However, the genetic background of patients with wildtype (WT) MC1R melanoma is poorly studied.

OBJECTIVES

To analyse the role of candidate common genetic variants on the melanoma risk and naevus count in Spanish patients with WT MC1R melanoma.

METHODS

We examined 753 individuals with WT MC1R from Spain (497 patients and 256 controls). We used OpenArray reverse-transcriptase polymerase chain reaction to genotype a panel of 221 common genetic variants involved in melanoma, naevogenesis, hormonal pathways and proinflammatory pathways. Genetic models were tested using multivariate logistic regression models. Nonparametric multifactor dimensionality reduction (MDR) was used to detect gene-gene interactions within each biological subgroup of variants.

RESULTS

We found that variant rs12913832 in the HERC2 gene, which is associated with blue eye colour, increased melanoma risk in individuals with WT MC1R [odds ratio (OR) 1·97, 95% confidence interval (CI) 1·48-2·63; adjusted P < 0·001; corrected P < 0·001]. We also observed a trend between the rs3798577 variant in the oestrogen receptor alpha gene (ESR1) and a lower naevus count, which was restricted to female patients with WT MC1R (OR 0·51, 95% CI 0·33-0·79; adjusted P = 0·002; corrected P = 0·11). This sex-dependent association was statistically significant in a larger cohort of patients with melanoma regardless of their MC1R status (n = 1497; OR 0·71, 95% CI 0·57-0·88; adjusted P = 0·002), reinforcing the hypothesis of an association between hormonal pathways and susceptibility to melanocytic proliferation. Last, the MDR analysis revealed four genetic combinations associated with melanoma risk or naevus count in patients with WT MC1R.

CONCLUSIONS

Our data suggest that epistatic interaction among common variants related to melanocyte biology or proinflammatory pathways might influence melanocytic proliferation in individuals with WT MC1R. What is already known about this topic? Genetic variants in the MC1R gene are the most prevalent melanoma genetic risk factor in the white population. Still, 20-40% of cases of melanoma occur in individuals with wildtype MC1R. Multiple genetic variants have a pleiotropic effect in melanoma and naevogenesis. Additional variants in unexplored pathways might also have a role in melanocytic proliferation in these patients. Epidemiological evidence suggests an association of melanocytic proliferation with hormonal pathways and proinflammatory pathways. What does this study add? Variant rs12913832 in the HERC2 gene, which is associated with blue eye colour, increases the melanoma risk in individuals with wildtype MC1R. Variant rs3798577 in the oestrogen receptor gene is associated with naevus count regardless of the MC1R status in female patients with melanoma. We report epistatic interactions among common genetic variants with a role in modulating the risk of melanoma or the number of naevi in individuals with wildtype MC1R. What is the translational message? We report a potential role of hormonal signalling pathways in melanocytic proliferation, providing a basis for better understanding of sex-based differences observed at the epidemiological level. We show that gene-gene interactions among common genetic variants might be responsible for an increased risk for melanoma development in individuals with a low-risk phenotype, such as darkly pigmented hair and skin.

A pharmacogenetic interaction analysis of bevacizumab with paclitaxel in advanced breast cancer patients

To investigate pharmacogenetic interactions among VEGF-A, VEGFR-2, IL-8, HIF-1α, EPAS-1, and TSP-1 SNPs and their role on progression-free survival (PFS) in metastatic breast cancer (MBC) patients treated with bevacizumab plus first-line paclitaxel or with paclitaxel alone. Analyses were performed on germline DNA, and SNPs were investigated by real-time PCR technique. The multifactor dimensionality reduction (MDR) methodology was applied to investigate the interaction between SNPs. The present study was an explorative, ambidirectional cohort study: 307 patients from 11 Oncology Units were evaluated retrospectively from 2009 to 2016, then followed prospectively (NCT01935102). Two hundred and fifteen patients were treated with paclitaxel and bevacizumab, whereas 92 patients with paclitaxel alone. In the bevacizumab plus paclitaxel group, the MDR software provided two pharmacogenetic interaction profiles consisting of the combination between specific VEGF-A rs833061 and VEGFR-2 rs1870377 genotypes. Median PFS for favorable genetic profile was 16.8 vs. the 10.6 months of unfavorable genetic profile (p = 0.0011). Cox proportional hazards model showed an adjusted hazard ratio of 0.64 (95% CI, 0.5–0.9; p = 0.004). Median OS for the favorable genetic profile was 39.6 vs. 28 months of unfavorable genetic profile (p = 0.0103). Cox proportional hazards model revealed an adjusted hazard ratio of 0.71 (95% CI, 0.5–1.01; p = 0.058). In the 92 patients treated with paclitaxel alone, the results showed no effect of the favorable genetic profile, as compared to the unfavorable genetic profile, either on the PFS (p = 0.509) and on the OS (p = 0.732). The pharmacogenetic statistical interaction between VEGF-A rs833061 and VEGFR-2 rs1870377 genotypes may identify a population of bevacizumab-treated patients with a better PFS.

What Can Machine Learning Approaches in Genomics Tell Us about the Molecular Basis of Amyotrophic Lateral Sclerosis?

Amyotrophic Lateral Sclerosis (ALS) is the most common late-onset motor neuron disorder, but our current knowledge of the molecular mechanisms and pathways underlying this disease remain elusive. This review (1) systematically identifies machine learning studies aimed at the understanding of the genetic architecture of ALS, (2) outlines the main challenges faced and compares the different approaches that have been used to confront them, and (3) compares the experimental designs and results produced by those approaches and describes their reproducibility in terms of biological results and the performances of the machine learning models. The majority of the collected studies incorporated prior knowledge of ALS into their feature selection approaches, and trained their machine learning models using genomic data combined with other types of mined knowledge including functional associations, protein-protein interactions, disease/tissue-specific information, epigenetic data, and known ALS phenotype-genotype associations. The importance of incorporating gene-gene interactions and cis-regulatory elements into the experimental design of future ALS machine learning studies is highlighted. Lastly, it is suggested that future advances in the genomic and machine learning fields will bring about a better understanding of ALS genetic architecture, and enable improved personalized approaches to this and other devastating and complex diseases.

CHIMGEN: a Chinese imaging genetics cohort to enhance cross-ethnic and cross-geographic brain research

The Chinese Imaging Genetics (CHIMGEN) study establishes the largest Chinese neuroimaging genetics cohort and aims to identify genetic and environmental factors and their interactions that are associated with neuroimaging and behavioral phenotypes. This study prospectively collected genomic, neuroimaging, environmental, and behavioral data from more than 7000 healthy Chinese Han participants aged 18-30 years. As a pioneer of large-sample neuroimaging genetics cohorts of non-Caucasian populations, this cohort can provide new insights into ethnic differences in genetic-neuroimaging associations by being compared with Caucasian cohorts. In addition to micro-environmental measurements, this study also collects hundreds of quantitative macro-environmental measurements from remote sensing and national survey databases based on the locations of each participant from birth to present, which will facilitate discoveries of new environmental factors associated with neuroimaging phenotypes. With lifespan environmental measurements, this study can also provide insights on the macro-environmental exposures that affect the human brain as well as their timing and mechanisms of action.

Advances in the genome-wide association study of chronic hepatitis B susceptibility in Asian population

Chronic hepatitis B (CHB) is the most common chronic liver disease resulting from viral infection and has become a serious threat to human health. Each year, about 1.2 million people in the world die from diseases caused by chronic infection of hepatitis B virus. The genetic polymorphism is significantly associated with the susceptibility to chronic hepatitis B. Genome-wide association study was recently developed and has become an important tool to detect susceptibility genes of CHB. To date, a number of CHB-associated susceptibility loci and regions have been identified by scientists over the world. To clearly understand the role of susceptibility loci in the occurrence of CHB is important for the early diagnosis and prevention of CHB.

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.

Genotype distribution-based inference of collective effects in genome-wide association studies: insights to age-related macular degeneration disease mechanism

BACKGROUND

Genome-wide association studies provide important insights to the genetic component of disease risks. However, an existing challenge is how to incorporate collective effects of interactions beyond the level of independent single nucleotide polymorphism (SNP) tests. While methods considering each SNP pair separately have provided insights, a large portion of expected heritability may reside in higher-order interaction effects.

RESULTS

We describe an inference approach (discrete discriminant analysis; DDA) designed to probe collective interactions while treating both genotypes and phenotypes as random variables. The genotype distributions in case and control groups are modeled separately based on empirical allele frequency and covariance data, whose differences yield disease risk parameters. We compared pairwise tests and collective inference methods, the latter based both on DDA and logistic regression. Analyses using simulated data demonstrated that significantly higher sensitivity and specificity can be achieved with collective inference in comparison to pairwise tests, and with DDA in comparison to logistic regression. Using age-related macular degeneration (AMD) data, we demonstrated two possible applications of DDA. In the first application, a genome-wide SNP set is reduced into a small number (∼100) of variants via filtering and SNP pairs with significant interactions are identified. We found that interactions between SNPs with highest AMD association were epigenetically active in the liver, adipocytes, and mesenchymal stem cells. In the other application, multiple groups of SNPs were formed from the genome-wide data and their relative strengths of association were compared using cross-validation. This analysis allowed us to discover novel collections of loci for which interactions between SNPs play significant roles in their disease association. In particular, we considered pathway-based groups of SNPs containing up to ∼10, 000 variants in each group. In addition to pathways related to complement activation, our collective inference pointed to pathway groups involved in phospholipid synthesis, oxidative stress, and apoptosis, consistent with the AMD pathogenesis mechanism where the dysfunction of retinal pigment epithelium cells plays central roles.

CONCLUSIONS

The simultaneous inference of collective interaction effects within a set of SNPs has the potential to reveal novel aspects of disease association.

Incorporating epistasis interaction of genetic susceptibility single nucleotide polymorphisms in a lung cancer risk prediction model

Incorporation of genetic variants such as single nucleotide polymorphisms (SNPs) into risk prediction models may account for a substantial fraction of attributable disease risk. Genetic data, from 2385 subjects recruited into the Liverpool Lung Project (LLP) between 2000 and 2008, consisting of 20 SNPs independently validated in a candidate-gene discovery study was used. Multifactor dimensionality reduction (MDR) and random forest (RF) were used to explore evidence of epistasis among 20 replicated SNPs. Multivariable logistic regression was used to identify similar risk predictors for lung cancer in the LLP risk model for the epidemiological model and extended model with SNPs. Both models were internally validated using the bootstrap method and model performance was assessed using area under the curve (AUC) and net reclassification improvement (NRI). Using MDR and RF, the overall best classifier of lung cancer status were SNPs rs1799732 (DRD2), rs5744256 (IL-18), rs2306022 (ITGA11) with training accuracy of 0.6592 and a testing accuracy of 0.6572 and a cross-validation consistency of 10/10 with permutation testing P<0.0001. The apparent AUC of the epidemiological model was 0.75 (95% CI 0.73–0.77). When epistatic data were incorporated in the extended model, the AUC increased to 0.81 (95% CI 0.79–0.83) which corresponds to 8% increase in AUC (DeLong's test P=2.2e-¹⁶); 17.5% by NRI. After correction for optimism, the AUC was 0.73 for the epidemiological model and 0.79 for the extended model. Our results showed modest improvement in lung cancer risk prediction when the SNP epistasis factor was added.

Phenome-Wide Association Studies: Embracing Complexity for Discovery

The inherent complexity of biological systems can be leveraged for a greater understanding of the impact of genetic architecture on outcomes, traits, and pharmacological response. The genome-wide association study (GWAS) approach has well-developed methods and relatively straight-forward methodologies; however, the bigger picture of the impact of genetic architecture on phenotypic outcome still remains to be elucidated even with an ever-growing number of GWAS performed. Greater consideration of the complexity of biological processes, using more data from the phenome, exposome, and diverse -omic resources, including considering the interplay of pleiotropy and genetic interactions, may provide additional leverage for making the most of the incredible wealth of information available for study. Here, we describe how incorporating greater complexity into analyses through the use of additional phenotypic data and widespread deployment of phenome-wide association studies may provide new insights into genetic factors influencing diseases, traits, and pharmacological response.

Sparse models for correlative and integrative analysis of imaging and genetic data

The development of advanced medical imaging technologies and high-throughput genomic measurements has enhanced our ability to understand their interplay as well as their relationship with human behavior by integrating these two types of datasets. However, the high dimensionality and heterogeneity of these datasets presents a challenge to conventional statistical methods; there is a high demand for the development of both correlative and integrative analysis approaches. Here, we review our recent work on developing sparse representation based approaches to address this challenge. We show how sparse models are applied to the correlation and integration of imaging and genetic data for biomarker identification. We present examples on how these approaches are used for the detection of risk genes and classification of complex diseases such as schizophrenia. Finally, we discuss future directions on the integration of multiple imaging and genomic datasets including their interactions such as epistasis.

Association of 5 Well-Defined Polymorphisms in the Gene Encoding Transforming Growth Factor-β1 With Coronary Artery Disease Among Chinese Patients With Hypertension

We assessed the association between 5 well-defined polymorphisms of the transforming growth factor-β1 (TGFB1) gene and coronary artery disease (CAD) among patients with hypertension from northeast China. All study participants were classified into patients with CAD (n = 679) and controls (n = 686) according to angiographic results. Genotyping was carried out with the ligase detection reaction method. In single-locus analysis, only genotypes of rs1800469 differed significantly between patients with CAD and controls ( P = .001); patients carrying the mutant allele of rs1800469 exhibited a 73% increased risk of CAD ( P < .001). Haplotype analysis indicated that haplotype A-T-T-C-C (alleles in the order of rs1800468, rs1800469, rs1800470, rs1800471, and rs1800472) was associated with a 1.49-fold increased risk ( P = .003). Interaction analysis identified an overall best 3-locus model including rs1800469, rs1800468, and rs1800471 ( P = .003). Taken together, we identified a synergistic interaction between TGFB1 gene multiple polymorphisms that entailed greater risk of CAD in Chinese patients.

Genetic analysis of quantitative phenotypes in AD and MCI: imaging, cognition and biomarkers

The Genetics Core of the Alzheimer's Disease Neuroimaging Initiative (ADNI), formally established in 2009, aims to provide resources and facilitate research related to genetic predictors of multidimensional Alzheimer's disease (AD)-related phenotypes. Here, we provide a systematic review of genetic studies published between 2009 and 2012 where either ADNI APOE genotype or genome-wide association study (GWAS) data were used. We review and synthesize ADNI genetic associations with disease status or quantitative disease endophenotypes including structural and functional neuroimaging, fluid biomarker assays, and cognitive performance. We also discuss the diverse analytical strategies used in these studies, including univariate and multivariate analysis, meta-analysis, pathway analysis, and interaction and network analysis. Finally, we perform pathway and network enrichment analyses of these ADNI genetic associations to highlight key mechanisms that may drive disease onset and trajectory. Major ADNI findings included all the top 10 AD genes and several of these (e.g., APOE, BIN1, CLU, CR1, and PICALM) were corroborated by ADNI imaging, fluid and cognitive phenotypes. ADNI imaging genetics studies discovered novel findings (e.g., FRMD6) that were later replicated on different data sets. Several other genes (e.g., APOC1, FTO, GRIN2B, MAGI2, and TOMM40) were associated with multiple ADNI phenotypes, warranting further investigation on other data sets. The broad availability and wide scope of ADNI genetic and phenotypic data has advanced our understanding of the genetic basis of AD and has nominated novel targets for future studies employing next-generation sequencing and convergent multi-omics approaches, and for clinical drug and biomarker development.

A review of multivariate analyses in imaging genetics

Recent advances in neuroimaging technology and molecular genetics provide the unique opportunity to investigate genetic influence on the variation of brain attributes. Since the year 2000, when the initial publication on brain imaging and genetics was released, imaging genetics has been a rapidly growing research approach with increasing publications every year. Several reviews have been offered to the research community focusing on various study designs. In addition to study design, analytic tools and their proper implementation are also critical to the success of a study. In this review, we survey recent publications using data from neuroimaging and genetics, focusing on methods capturing multivariate effects accommodating the large number of variables from both imaging data and genetic data. We group the analyses of genetic or genomic data into either a priori driven or data driven approach, including gene-set enrichment analysis, multifactor dimensionality reduction, principal component analysis, independent component analysis (ICA), and clustering. For the analyses of imaging data, ICA and extensions of ICA are the most widely used multivariate methods. Given detailed reviews of multivariate analyses of imaging data available elsewhere, we provide a brief summary here that includes a recently proposed method known as independent vector analysis. Finally, we review methods focused on bridging the imaging and genetic data by establishing multivariate and multiple genotype-phenotype-associations, including sparse partial least squares, sparse canonical correlation analysis, sparse reduced rank regression and parallel ICA. These methods are designed to extract latent variables from both genetic and imaging data, which become new genotypes and phenotypes, and the links between the new genotype-phenotype pairs are maximized using different cost functions. The relationship between these methods along with their assumptions, advantages, and limitations are discussed.

Evidence for an epistatic effect between TP53 R72P and MDM2 T309G SNPs in HIV infection: a cross-sectional study in women from South Brazil

OBJECTIVE

To investigate the associations of TP53 R72P and MDM2 T309G SNPs with HPV infection status, HPV oncogenic risk and HIV infection status.

DESIGN

Cross-sectional study combining two groups (150 HIV-negative and 100 HIV-positive) of women.

METHODS

Data was collected using a closed questionnaire. DNA was extracted from cervical samples. HPV infection status was determined by nested-PCR, and HPV oncogenic risk group by Sanger sequencing. Both SNPS were genotyped by PCR-RFLP. Crude and adjusted associations involving each exposure (R72P and T309G SNPs, as well as 13 models of epistasis) and each outcome (HPV status, HPV oncogenic risk group and HIV infection) were assessed using logistic regression.

RESULTS

R72P SNP was protectively associated with HPV status (overdominant model), as well as T309G SNP with HPV oncogenic risk (strongest in the overdominant model). No epistatic model was associated with HPV status, but a dominant (R72P over T309G) protective epistatic effect was observed for HPV oncogenic risk. HIV status was strongly associated (risk factor) with different epistatic models, especially in models based on a visual inspection of the results. Moreover, HIV status was evidenced to be an effect mediator of the associations involving HPV oncogenic risk.

CONCLUSIONS

We found evidence for a role of R72P and T309G SNPs in HPV status and HPV oncogenic risk (respectively), and strong associations were found for an epistatic effect in HIV status. Prospective studies in larger samples are warranted to validate our findings, which point to a novel role of these SNPs in HIV infection.

Integrating gene expression and epidemiological data for the discovery of genetic interactions associated with cancer risk

Dozens of common genetic variants associated with cancer risk have been identified through genome-wide association studies (GWASs). However, these variants only explain a modest fraction of the heritability of disease. The missing heritability has been attributed to several factors, among them the existence of genetic interactions (G × G). Systematic screens for G × G in model organisms have revealed their fundamental influence in complex phenotypes. In this scenario, G × G overlap significantly with other types of gene and/or protein relationships. Here, by integrating predicted G × G from GWAS data and complex- and context-defined gene coexpression profiles, we provide evidence for G × G associated with cancer risk. G × G predicted from a breast cancer GWAS dataset identified significant overlaps [relative enrichments (REs) of 8-36%, empirical P values < 0.05 to 10(-4)] with complex (non-linear) gene coexpression in breast tumors. The use of gene or protein data not specific for breast cancer did not reveal overlaps. According to the predicted G × G, experimental assays demonstrated functional interplay between lipoma-preferred partner and transforming growth factor-β signaling in the MCF10A non-tumorigenic mammary epithelial cell model. Next, integration of pancreatic tumor gene expression profiles with pancreatic cancer G × G predicted from a GWAS corroborated the observations made for breast cancer risk (REs of 25-59%). The method presented here can potentially support the identification of genetic interactions associated with cancer risk, providing novel mechanistic hypotheses for carcinogenesis.