Pathway analysis by adaptive combination of P-values

Identification of additional loci associated with antibody response to Mycobacterium avium ssp. Paratuberculosis in cattle by GSEA-SNP analysis

Mycobacterium avium subsp. paratuberculosis: (MAP) causes a contagious chronic infection results in Johne's disease in a wide range of animal species, including cattle. Several genome-wide association studies (GWAS) have been carried out to identify loci putatively associated with MAP susceptibility by testing each marker separately and identifying SNPs that show a significant association with the phenotype, while SNP with modest effects are usually ignored. The objective of this study was to identify modest-effect genes associated with MAP susceptibility using a pathway-based approach. The Illumina BovineSNP50 BeadChip was used to genotype 966 Holstein cows, 483 positive and 483 negative for antibody response to MAP, data were then analyzed using novel SNP-based Gene Set Enrichment Analysis (GSEA-SNP) and validated with Adaptive Rank Truncated Product methodology. An allele-based test was carried out to estimate the statistical association for each marker with the phenotype, subsequently SNPs were mapped to the closest genes, considering for each gene the single variant with the highest value within a window of 50 kb, then pathway-statistics were tested using the GSEA-SNP method. The GO biological process "embryogenesis and morphogenesis" was most highly associated with antibody response to MAP. Within this pathway, five genes code for proteins which play a role in the immune defense relevant to response to bacterial infection. The immune response genes identified would not have been considered using a standard GWAS, thus demonstrating that the pathway approach can extend the interpretation of genome-wide association analyses and identify additional candidate genes for target traits.

Genome-wide association study of sepsis in extremely premature infants

OBJECTIVE

To identify genetic variants associated with sepsis (early-onset and late-onset) using a genome-wide association (GWA) analysis in a cohort of extremely premature infants.

STUDY DESIGN

Previously generated GWA data from the Neonatal Research Network's anonymised genomic database biorepository of extremely premature infants were used for this study. Sepsis was defined as culture-positive early-onset or late-onset sepsis or culture-proven meningitis. Genomic and whole-genome-amplified DNA was genotyped for 1.2 million single-nucleotide polymorphisms (SNPs); 91% of SNPs were successfully genotyped. We imputed 7.2 million additional SNPs. p Values and false discovery rates (FDRs) were calculated from multivariate logistic regression analysis adjusting for gender, gestational age and ancestry. Target statistical value was p<10-5. Secondary analyses assessed associations of SNPs with pathogen type. Pathway analyses were also run on primary and secondary end points.

RESULTS

Data from 757 extremely premature infants were included: 351 infants with sepsis and 406 infants without sepsis. No SNPs reached genome-wide significance levels (5×10-8); two SNPs in proximity to FOXC2 and FOXL1 genes achieved target levels of significance. In secondary analyses, SNPs for ELMO1, IRAK2 (Gram-positive sepsis), RALA, IMMP2L (Gram-negative sepsis) and PIEZO2 (fungal sepsis) met target significance levels. Pathways associated with sepsis and Gram-negative sepsis included gap junctions, fibroblast growth factor receptors, regulators of cell division and interleukin-1-associated receptor kinase 2 (p values<0.001 and FDR<20%).

CONCLUSIONS

No SNPs met genome-wide significance in this cohort of extremely low birthweight infants; however, areas of potential association and pathways meriting further study were identified.

OPATs: Omnibus P-value association tests

Combining statistical significances (P-values) from a set of single-locus association tests in genome-wide association studies is a proof-of-principle method for identifying disease-associated genomic segments, functional genes and biological pathways. We review P-value combinations for genome-wide association studies and introduce an integrated analysis tool, Omnibus P-value Association Tests (OPATs), which provides popular analysis methods of P-value combinations. The software OPATs programmed in R and R graphical user interface features a user-friendly interface. In addition to analysis modules for data quality control and single-locus association tests, OPATs provides three types of set-based association test: window-, gene- and biopathway-based association tests. P-value combinations with or without threshold and rank truncation are provided. The significance of a set-based association test is evaluated by using resampling procedures. Performance of the set-based association tests in OPATs has been evaluated by simulation studies and real data analyses. These set-based association tests help boost the statistical power, alleviate the multiple-testing problem, reduce the impact of genetic heterogeneity, increase the replication efficiency of association tests and facilitate the interpretation of association signals by streamlining the testing procedures and integrating the genetic effects of multiple variants in genomic regions of biological relevance. In summary, P-value combinations facilitate the identification of marker sets associated with disease susceptibility and uncover missing heritability in association studies, thereby establishing a foundation for the genetic dissection of complex diseases and traits. OPATs provides an easy-to-use and statistically powerful analysis tool for P-value combinations. OPATs, examples, and user guide can be downloaded from http://www.stat.sinica.edu.tw/hsinchou/genetics/association/OPATs.htm.

Genetic association analysis of the RTK/ERK pathway with aggressive prostate cancer highlights the potential role of CCND2 in disease progression

The RTK/ERK signaling pathway has been implicated in prostate cancer progression. However, the genetic relevance of this pathway to aggressive prostate cancer at the SNP level remains undefined. Here we performed a SNP and gene-based association analysis of the RTK/ERK pathway with aggressive prostate cancer in a cohort comprising 956 aggressive and 347 non-aggressive cases. We identified several loci including rs3217869/CCND2 within the pathway shown to be significantly associated with aggressive prostate cancer. Our functional analysis revealed a statistically significant relationship between rs3217869 risk genotype and decreased CCND2 expression levels in a collection of 119 prostate cancer patient samples. Reduced expression of CCND2 promoted cell proliferation and its overexpression inhibited cell growth of prostate cancer. Strikingly, CCND2 downregulation was consistently observed in the advanced prostate cancer in 18 available clinical data sets with a total amount of 1,095 prostate samples. Furthermore, the lower expression levels of CCND2 markedly correlated with prostate tumor progression to high Gleason score and elevated PSA levels, and served as an independent predictor of biochemical relapse and overall survival in a large cohort of prostate cancer patients. Together, we have identified an association of genetic variants and genes in the RTK/ERK pathway with prostate cancer aggressiveness, and highlighted the potential importance of CCND2 in prostate cancer susceptibility and tumor progression to metastasis.

An adaptive two-sample test for high-dimensional means

Several two-sample tests for high-dimensional data have been proposed recently, but they are powerful only against certain alternative hypotheses. In practice, since the true alternative hypothesis is unknown, it is unclear how to choose a powerful test. We propose an adaptive test that maintains high power across a wide range of situations and study its asymptotic properties. Its finite-sample performance is compared with that of existing tests. We apply it and other tests to detect possible associations between bipolar disease and a large number of single nucleotide polymorphisms on each chromosome based on data from a genome-wide association study. Numerical studies demonstrate the superior performance and high power of the proposed test across a wide spectrum of applications.

A novel TCF7L2 type 2 diabetes SNP identified from fine mapping in African American women

SNP rs7903146 in the Wnt pathway’s TCF7L2 gene is the variant most significantly associated with type 2 diabetes to date, with associations observed across diverse populations. We sought to determine whether variants in other Wnt pathway genes are also associated with this disease. We evaluated 69 genes involved in the Wnt pathway, including TCF7L2, for associations with type 2 diabetes in 2632 African American cases and 2596 controls from the Black Women’s Health Study. Tag SNPs for each gene region were genotyped on a custom Affymetrix Axiom Array, and imputation was performed to 1000 Genomes Phase 3 data. Gene-based analyses were conducted using the adaptive rank truncated product (ARTP) statistic. The PSMD2 gene was significantly associated with type 2 diabetes after correction for multiple testing (corrected p = 0.016), based on the nine most significant single variants in the +/- 20 kb region surrounding the gene, which includes nearby genes EIF4G1, ECE2, and EIF2B5. Association data on four of the nine variants were available from an independent sample of 8284 African American cases and 15,543 controls; associations were in the same direction, but weak and not statistically significant. TCF7L2 was the only other gene associated with type 2 diabetes at nominal p <0.01 in our data. One of the three variants in the best gene-based model for TCF7L2, rs114770437, was not correlated with the GWAS index SNP rs7903146 and may represent an independent association signal seen only in African ancestry populations. Data on this SNP were not available in the replication sample.

Single nucleotide variants in metastasis-related genes are associated with breast cancer risk, by lymph node involvement and estrogen receptor status, in women with European and African ancestry

Single nucleotide polymorphisms (SNPs) in pathways influencing lymph node (LN) metastasis and estrogen receptor (ER) status in breast cancer may partially explain inter-patient variability in prognosis. We examined 154 SNPs in 12 metastasis-related genes for associations with breast cancer risk, stratified by LN and ER status, in European-American (EA) and African-American (AA) women. Two-thousand six hundred and seventy-one women enrolled in the Women's Circle of Health Study were genotyped. Pathway analyses were conducted using the adaptive rank truncated product (ARTP) method, with pARTP  ≤ 0.10 as significant. Multi-allelic risk scores were created for the ARTP-significant gene(s). Single-SNP and risk score associations were modeled using logistic regression, with false discovery rate (FDR) P-value adjustment. Although single-SNP associations were not significant at pFDR  < 0.05, several genes were significant in the ARTP analyses. In AA women, significant ARTP gene-level associations included CDH1 with LN+ (pARTP  = 0.10; multi-allelic OR = 1.13, 95%CI 1.07-1.19, pFDR  = 0.0003) and SIPA1 with ER- breast cancer (pARTP  = 0.10; multi-allelic OR = 1.16, 95%CI 1.02-1.31, pFDR  = 0.03). In EA women, MTA2 was associated with overall breast cancer risk (pARTP  = 0.004), regardless of ER status, and with LN- disease (pARTP  = 0.01). Also significant were SATB1 in ER- (pARTP  = 0.03; multi-allelic OR = 1.12, 95%CI 1.05-1.20, pFDR  = 0.003) and KISS1 in LN- (pARTP  = 0.10; multi-allelic OR = 1.18, 95%CI 1.08-1.29, pFDR  = 0.002) analyses. Among LN+ cases, significant ARTP associations were observed for SNAI1, CD82, NME1, and CTNNB1 (multi-allelic OR = 1.09, 95%CI 1.04-1.14, pFDR  = 0.001). Our findings suggest that variants in several metastasis genes may affect breast cancer risk by LN or ER status, although verification in larger studies is required. © 2016 Wiley Periodicals, Inc.

Genetic Polymorphisms in Estrogen Metabolic Pathway Associated with Risks of Alzheimer's Disease: Evidence from a Southern Chinese Population

OBJECTIVES

To investigate whether genetic variations on the estrogen metabolic pathway would be associated with risk of Alzheimer's disease (AD).

DESIGN

Cross-sectional study.

SETTING

Individuals were recruited at the Memory Clinic, Queen Mary Hospital, Hong Kong.

PARTICIPANTS

Chinese individuals with (n = 426) and without (n = 350) AD.

MEASUREMENTS

All subjects underwent a standardized cognitive assessment and genotyping of four candidate genes on the estrogen metabolic pathway (estrogen receptor α gene (ESR1), estrogen receptor β gene (ESR2), cytochrome P450 19A1 gene (CYP19A1), cytochrome P450 11A1 gene (CYP11A1)).

RESULTS

Apart from consistent results showing an association between apolipoprotein (APO)E and AD, strong evidence of disease associations were found for polymorphisms in ESR2 and CYP11A1 based on the entire data set. For ESR2, significant protective effects were found for A alleles of rs4986938 (permuted P = .02) and rs867443 (permuted P = .02). For CYP11A1, significant risk effects were found for G alleles of rs11638442 (permuted P = .03) and rs11632698 (permuted P = .03). Stratifying subjects according to APOE ε4 status, their genetic effects continued to be significant in the APOE ε4-negative subgroup. Associations between CYP11A1 polymorphisms (rs2279357, rs2073475) and risk of AD were detected in women but not men. Further gene-level analysis confirmed the above association between ESR2 and CYP11A1, and pathway-level analysis highlighted the genetic effect of the estrogen metabolic pathway on disease susceptibility (permuted pathway-level P = .03).

CONCLUSION

Consistent with previous biological findings for sex steroid hormones in the central nervous system, genetic alterations on the estrogen metabolic pathway were revealed in the Chinese population. Confirmation of these present findings in an independent population is warranted to elucidate disease pathogenesis and to explore the potential of hormone therapy in the treatment of AD.

Exome-Wide Association Study Identifies Low-Frequency Coding Variants in 2p23.2 and 7p11.2 Associated with Survival of Non-Small Cell Lung Cancer Patients

INTRODUCTION

A growing body of evidence has suggested that low-frequency or rare coding variants might have strong effects on the development and prognosis of cancer. Here, we aim to assess the role of low-frequency and rare coding variants in the survival of NSCLC in Chinese populations.

METHODS

We performed an exome-wide scan of 247,870 variants in 1008 patients with NSCLC and replicated the promising variants by using imputed genotype data of The Cancer Genome Atlas (TCGA) with a Cox regression model. Gene-based and pathway-based analysis were also performed for nonsynonymous or splice site variants. Additionally, analysis of gene expression data in the TCGA was used to increase the reliability of candidate loci and genes.

RESULTS

A low-frequency missense variant in chaperonin containing TCP1 subunit 6A gene (CCT6A) (rs33922584: adjusted hazard ratio [HR_adjusted] = 1.75, p = 6.06 × 10^-4) was significantly related to the survival of patients with NSCLC, which was further replicated by the TCGA samples (HR_adjusted = 4.19, p = 0.015). Interestingly, the G allele of rs33922584 was significantly associated with high expression of CCT6A (p = 0.019) that might induce the worse survival in the TCGA samples (HR_adjusted = 1.15, p = 0.047). Besides, rs117512489 in gene phospholipase B1 gene (PLB1) (HR = 2.02, p = 7.28 × 10^-4) was also associated with survival of the patients with NSCLC in our samples, but it was supported only by gene expression analysis in the TCGA (HR_adjusted = 1.15, p = 0.023). Gene-based and pathway-based analysis revealed a total of 32 genes, including CCT6A and 34 potential pathways might account for the survival of NSCLC, respectively.

CONCLUSION

These results provided more evidence for the important role of low-frequency or rare variants in the survival of patients with NSCLC.

Identifying Host Genetic Variants Associated with Microbiome Composition by Testing Multiple Beta Diversity Matrices

OBJECTIVES

Host genetics have been recently reported to affect human microbiome composition. We previously developed a statistical framework, microbiomeGWAS, to identify host genetic variants associated with microbiome composition by testing a distance matrix. However, statistical power depends on the choice of a microbiome distance matrix. To achieve more robust statistical power, we aim to extend microbiomeGWAS to test the association with many distance matrices, which are defined based on multilevel taxa abundances and phylogenetic information.

METHODS

The main challenge is to accurately and rapidly evaluate the significance for millions of SNPs. We propose methods for approximating p values by correcting for the multiple testing introduced by testing many distance matrices and by correcting for the skewness and kurtosis of score statistics.

RESULTS

The accuracy of p value approximation was verified by simulations. We applied our method to a set of 147 lung cancer patients with 16S rRNA microbiome profiles from nonmalignant lung tissues. We show that correcting for skewness and kurtosis eliminated dramatic deviations in the quantile-quantile plot.

CONCLUSION

We developed computationally efficient methods for identifying host genetic variants associated with microbiome composition by testing many distance matrices. The algorithms are implemented in the package microbiomeGWAS (https://github.com/lsncibb/microbiomeGWAS).

THE SCREENING AND RANKING ALGORITHM FOR CHANGE-POINTS DETECTION IN MULTIPLE SAMPLES

The chromosome copy number variation (CNV) is the deviation of genomic regions from their normal copy number states, which may associate with many human diseases. Current genetic studies usually collect hundreds to thousands of samples to study the association between CNV and diseases. CNVs can be called by detecting the change-points in mean for sequences of array-based intensity measurements. Although multiple samples are of interest, the majority of the available CNV calling methods are single sample based. Only a few multiple sample methods have been proposed using scan statistics that are computationally intensive and designed toward either common or rare change-points detection. In this paper, we propose a novel multiple sample method by adaptively combining the scan statistic of the screening and ranking algorithm (SaRa), which is computationally efficient and is able to detect both common and rare change-points. We prove that asymptotically this method can find the true change-points with almost certainty and show in theory that multiple sample methods are superior to single sample methods when shared change-points are of interest. Additionally, we report extensive simulation studies to examine the performance of our proposed method. Finally, using our proposed method as well as two competing approaches, we attempt to detect CNVs in the data from the Primary Open-Angle Glaucoma Genes and Environment study, and conclude that our method is faster and requires less information while our ability to detect the CNVs is comparable or better.

Pathway analysis of complex diseases for GWAS, extending to consider rare variants, multi-omics and interactions

BACKGROUND

Genome-wide association studies (GWAS) is a major method for studying the genetics of complex diseases. Finding all sequence variants to explain fully the aetiology of a disease is difficult because of their small effect sizes. To better explain disease mechanisms, pathway analysis is used to consolidate the effects of multiple variants, and hence increase the power of the study. While pathway analysis has previously been performed within GWAS only, it can now be extended to examining rare variants, other "-omics" and interaction data.

SCOPE OF REVIEW

1. Factors to consider in the choice of software for GWAS pathway analysis. 2. Examples of how pathway analysis is used to analyse rare variants, other "-omics" and interaction data.

MAJOR CONCLUSIONS

To choose appropriate software tools, factors for consideration include covariate compatibility, null hypothesis, one- or two-step analysis required, curation method of gene sets, size of pathways, and size of flanking regions to define gene boundaries. For rare variants, analysis performance depends on consistency between assumed and actual effect distribution of variants. Integration of other "-omics" data and interaction can better explain gene functions.

GENERAL SIGNIFICANCE

Pathway analysis methods will be more readily used for integration of multiple sources of data, and enable more accurate prediction of phenotypes.

Efficient and Powerful Method for Combining P-Values in Genome-Wide Association Studies

The goal of Genome-wide Association Studies (GWAS) is the identification of genetic variants, usually single nucleotide polymorphisms (SNPs), that are associated with disease risk. However, SNPs detected so far with GWAS for most common diseases only explain a small proportion of their total heritability. Gene set analysis (GSA) has been proposed as an alternative to single-SNP analysis with the aim of improving the power of genetic association studies. Nevertheless, most GSA methods rely on expensive computational procedures that make unfeasible their implementation in GWAS. We propose a new GSA method, referred as globalEVT, which uses the extreme value theory to derive gene-level p-values. GlobalEVT reduces dramatically the computational requirements compared to other GSA approaches. In addition, this new approach improves the power by allowing different inheritance models for each genetic variant as illustrated in the simulation study performed and allows the existence of correlation between the SNPs. Real data analysis of an Attention-deficit/hyperactivity disorder (ADHD) study illustrates the importance of using GSA approaches for exploring new susceptibility genes. Specifically, the globalEVT method is able to detect genes related to Cyclophilin A like domain proteins which is known to play an important role in the mechanisms of ADHD development.

Energy homeostasis genes and survival after breast cancer diagnosis: the Breast Cancer Health Disparities Study

PURPOSE

The leptin-signaling pathway and other genes involved in energy homeostasis (EH) have been examined in relation to breast cancer risk as well as to obesity. We test the hypothesis that genetic variation in EH genes influences survival after diagnosis with breast cancer and that body mass index (BMI) will modify that risk.

METHODS

We evaluated associations between 10 EH genes and survival among 1,186 non-Hispanic white and 1,155 Hispanic/Native American women diagnosed with breast cancer. Percent Native American (NA) ancestry was determined from 104 ancestry-informative markers. Adaptive rank truncation product (ARTP) was used to determine gene and pathway significance.

RESULTS

The overall EH pathway was marginally significant for all-cause mortality among women with low NA ancestry (P(ARTP) = 0.057). Within the pathway, ghrelin(GHRL) and leptin receptor (LEPR) were significantly associated with all-cause mortality (P(ARTP) = 0.035 and 0.007, respectively). The EH pathway was significantly associated with breast cancer-specific mortality among women with low NA ancestry (P(ARTP) = 0.038). Three genes cholecystokinin (CCK), GHRL, and LEPR were significantly associated with breast cancer-specific mortality among women with low NA ancestry (P(ARTP) = 0.046,0.015, and 0.046, respectively), while neuropeptide Y (NPY) was significantly associated with breast cancer-specific mortality among women with higher NA ancestry(P(ARTP) = 0.038). BMI did not modify these associations.

CONCLUSIONS

Our data support our hypothesis that certain EH genes influence survival after diagnosis with breast cancer; associations appear to be most important among women with low NA ancestry.

Genetic variations in the Hippo signaling pathway and breast cancer risk in African American women in the AMBER Consortium

The Hippo signaling pathway regulates cellular proliferation and survival, thus exerting profound effects on normal cell fate and tumorigenesis. Dysfunction of the Hippo pathway components has been linked with breast cancer stem cell regulation, as well as breast tumor progression and metastasis. TAZ, a key component of the Hippo pathway, is highly expressed in triple negative breast cancer; however, the associations of genetic variations in this important pathway with breast cancer risk remain largely unexplored. Here, we analyzed 8309 germline variants in 15 genes from the Hippo pathway with a total of 3663 cases and 4687 controls from the African American Breast Cancer Epidemiology and Risk Consortium. Odds ratios (ORs) were estimated using logistic regression for overall breast cancer, by estrogen receptor (ER) status (1983 ER positive and 1098 ER negative), and for case-only analyses by ER status. The Hippo signaling pathway was significantly associated with ER-negative breast cancer (pathway level P = 0.02). Gene-based analyses revealed that CDH1 was responsible for the pathway association (P < 0.01), with rs4783673 in CDH1 statistically significant after gene-level adjustment for multiple comparisons (P = 9.2×10(-5), corrected P = 0.02). rs142697907 in PTPN14 was associated with ER-positive breast cancer and rs2456773 in CDK1 with ER-negativity in case-only analysis after gene-level correction for multiple comparisons (corrected P < 0.05). In conclusion, common genetic variations in the Hippo signaling pathway may contribute to both ER-negative and ER+ breast cancer risk in AA women.

A Powerful Procedure for Pathway-Based Meta-analysis Using Summary Statistics Identifies 43 Pathways Associated with Type II Diabetes in European Populations

Meta-analysis of multiple genome-wide association studies (GWAS) has become an effective approach for detecting single nucleotide polymorphism (SNP) associations with complex traits. However, it is difficult to integrate the readily accessible SNP-level summary statistics from a meta-analysis into more powerful multi-marker testing procedures, which generally require individual-level genetic data. We developed a general procedure called Summary based Adaptive Rank Truncated Product (sARTP) for conducting gene and pathway meta-analysis that uses only SNP-level summary statistics in combination with genotype correlation estimated from a panel of individual-level genetic data. We demonstrated the validity and power advantage of sARTP through empirical and simulated data. We conducted a comprehensive pathway-based meta-analysis with sARTP on type 2 diabetes (T2D) by integrating SNP-level summary statistics from two large studies consisting of 19,809 T2D cases and 111,181 controls with European ancestry. Among 4,713 candidate pathways from which genes in neighborhoods of 170 GWAS established T2D loci were excluded, we detected 43 T2D globally significant pathways (with Bonferroni corrected p-values < 0.05), which included the insulin signaling pathway and T2D pathway defined by KEGG, as well as the pathways defined according to specific gene expression patterns on pancreatic adenocarcinoma, hepatocellular carcinoma, and bladder carcinoma. Using summary data from 8 eastern Asian T2D GWAS with 6,952 cases and 11,865 controls, we showed 7 out of the 43 pathways identified in European populations remained to be significant in eastern Asians at the false discovery rate of 0.1. We created an R package and a web-based tool for sARTP with the capability to analyze pathways with thousands of genes and tens of thousands of SNPs.

Conditioning adaptive combination of P-values method to analyze case-parent trios with or without population controls

Detection of rare causal variants can help uncover the etiology of complex diseases. Recruiting case-parent trios is a popular study design in family-based studies. If researchers can obtain data from population controls, utilizing them in trio analyses can improve the power of methods. The transmission disequilibrium test (TDT) is a well-known method to analyze case-parent trio data. It has been extended to rare-variant association testing (abbreviated as "rvTDT"), with the flexibility to incorporate population controls. The rvTDT method is robust to population stratification. However, power loss may occur in the conditioning process. Here we propose a "conditioning adaptive combination of P-values method" (abbreviated as "conADA"), to analyze trios with/without unrelated controls. By first truncating the variants with larger P-values, we decrease the vulnerability of conADA to the inclusion of neutral variants. Moreover, because the test statistic is developed by conditioning on parental genotypes, conADA generates valid statistical inference in the presence of population stratification. With regard to statistical methods for next-generation sequencing data analyses, validity may be hampered by population stratification, whereas power may be affected by the inclusion of neutral variants. We recommend conADA for its robustness to these two factors (population stratification and the inclusion of neutral variants).

Group-combined P-values with applications to genetic association studies

MOTIVATION

In large-scale genetic association studies with tens of hundreds of single nucleotide polymorphisms (SNPs) genotyped, the traditional statistical framework of logistic regression using maximum likelihood estimator (MLE) to infer the odds ratios of SNPs may not work appropriately. This is because a large number of odds ratios need to be estimated, and the MLEs may be not stable when some of the SNPs are in high linkage disequilibrium. Under this situation, the P-value combination procedures seem to provide good alternatives as they are constructed on the basis of single-marker analysis.

RESULTS

The commonly used P-value combination methods (such as the Fisher's combined test, the truncated product method, the truncated tail strength and the adaptive rank truncated product) may lose power when the significance level varies across SNPs. To tackle this problem, a group combined P-value method (GCP) is proposed, where the P-values are divided into multiple groups and then are combined at the group level. With this strategy, the significance values are integrated at different levels, and the power is improved. Simulation shows that the GCP can effectively control the type I error rates and have additional power over the existing methods-the power increase can be as high as over 50% under some situations. The proposed GCP method is applied to data from the Genetic Analysis Workshop 16. Among all the methods, only the GCP and ARTP can give the significance to identify a genomic region covering gene DSC3 being associated with rheumatoid arthritis, but the GCP provides smaller P-value.

AVAILABILITY AND IMPLEMENTATION

http://www.statsci.amss.ac.cn/yjscy/yjy/lqz/201510/t20151027_313273.html

CONTACT

liqz@amss.ac.cn

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Efficient Software for Multi-marker, Region-Based Analysis of GWAS Data

Genome-wide association studies (GWAS) have associated many single variants with complex disease, yet the better part of heritable complex disease risk remains unexplained. Analytical tools designed to work under specific population genetic models are needed. Rare variants are increasingly shown to be important in human complex disease, but most existing GWAS data do not cover rare variants. Explicit population genetic models predict that genes contributing to complex traits and experiencing recurrent, unconditionally deleterious, mutation will harbor multiple rare, causative mutations of subtle effect. It is difficult to identify genes harboring rare variants of large effect that contribute to complex disease risk via the single marker association tests typically used in GWAS. Gene/region-based association tests may have the power detect associations by combining information from multiple markers, but have yielded limited success in practice. This is partially because many methods have not been widely applied. Here, we empirically demonstrate the utility of a procedure based on the rank truncated product (RTP) method, filtered to reduce the effects of linkage disequilibrium. We apply the procedure to the Wellcome Trust Case Control Consortium (WTCCC) data set, and uncover previously unidentified associations, some of which have been replicated in much larger studies. We show that, in the absence of significant rare variant coverage, RTP based methods still have the power to detect associated genes. We recommend that RTP-based methods be applied to all existing GWAS data to maximize the usefulness of those data. For this, we provide efficient software implementing our procedure.

Beyond Rare-Variant Association Testing: Pinpointing Rare Causal Variants in Case-Control Sequencing Study

Rare-variant association testing usually requires some method of aggregation. The next important step is to pinpoint individual rare causal variants among a large number of variants within a genetic region. Recently Ionita-Laza et al. propose a backward elimination (BE) procedure that can identify individual causal variants among the many variants in a gene. The BE procedure removes a variant if excluding this variant can lead to a smaller P-value for the BURDEN test (referred to as "BE-BURDEN") or the SKAT test (referred to as "BE-SKAT"). We here use the adaptive combination of P-values (ADA) method to pinpoint causal variants. Unlike most gene-based association tests, the ADA statistic is built upon per-site P-values of individual variants. It is straightforward to select important variants given the optimal P-value truncation threshold found by ADA. We performed comprehensive simulations to compare ADA with BE-SKAT and BE-BURDEN. Ranking these three approaches according to positive predictive values (PPVs), the percentage of truly causal variants among the total selected variants, we found ADA > BE-SKAT > BE-BURDEN across all simulation scenarios. We therefore recommend using ADA to pinpoint plausible rare causal variants in a gene.

Pathway, in silico and tissue-specific expression quantitative analyses of oesophageal squamous cell carcinoma genome-wide association studies data

BACKGROUND

Oesophageal cancer is the fourth leading cause of cancer death in China where essentially all cases are histologically oesophageal squamous cell carcinoma (ESCC). Agnostic pathway-based analyses of genome-wide association study (GWAS) data combined with tissue-specific expression quantitative trait loci (eQTL) analysis and publicly available functional data can identify biological pathways and/or genes enriched with functionally-relevant disease-associated variants.

METHOD

We used the adaptive multilocus joint test to analyse 1827 pathways containing 6060 genes using GWAS data from 1942 ESCC cases and 2111 controls with Chinese ancestry. We examined the function of risk alleles using in silico and eQTL analyses in oesophageal tissues.

RESULTS

Associations with ESCC risk were observed for 36 pathways predominantly involved in apoptosis, cell cycle regulation and DNA repair and containing known GWAS-associated genes. After excluding genes with previous GWAS signals, candidate pathways (and genes) for ESCC risk included taste transduction (KEGG_hsa04742; TAS2R13, TAS2R42, TAS2R14, TAS2R46,TAS2R50), long-patch base excision repair (Reactome_pid; POLD2) and the metabolics pathway (KEGG_hsa01100; MTAP, GAPDH, DCTD, POLD2, AMDHD1). We identified and validated CASP8 rs13016963 and IDH2 rs11630814 as eQTLs, and CASP8 rs3769823 and IDH2 rs4561444 as the potential functional variants in high-linkage disequilibrium with these single nucleotide polymorphisms (SNPs), respectively. Further, IDH2 mRNA levels were down-regulated in ESCC (tumour:normal-fold change = 0.69, P =  .75E-14).

CONCLUSION

Agnostic pathway-based analyses and integration of multiple types of functional data provide new evidence for the contribution of genes in taste transduction and metabolism to ESCC susceptibility, and for the functionality of both established and new ESCC risk-related SNPs.

Gene-based analysis of the fibroblast growth factor receptor signaling pathway in relation to breast cancer in African American women: the AMBER consortium

We conducted gene-based analysis in 26 genes in the FGFR signaling pathway to identify genes carrying genetic variation affecting risk of breast cancer and the specific estrogen receptor (ER) subtypes. Tagging single-nucleotide polymorphisms (SNPs) for each gene were selected and genotyped on a customized Illumina Exome Array. Imputation was carried out using 1000 Genomes haplotypes. The analysis included 3237 SNPs in 3663 breast cancer cases (including 1983 ER-positive, and 1098 ER-negative) and 4687 controls from the African American Breast Cancer Epidemiology and Risk consortium, a collaborative project of four large studies of breast cancer in African American women (Carolina Breast Cancer Study, Black Women's Health Study, Women's Circle of Health Study, and Multiethnic Cohort). We used a multi-locus adaptive joint (AdaJoint) test to determine the association of each gene in the FGFR signaling pathway with overall breast cancer and ER subtypes. The FGF1 gene was significantly associated with risk of ER-negative breast cancer (P = 0.001). The FGFR2 gene was associated with risk of overall breast cancer (P = 0.002) and ER-positive breast cancer (P = 0.002). The FGF1 gene affects risk of ER-negative breast cancer in African American women. We confirmed the association of the FGFR2 gene with risk of overall and ER-positive breast cancer. These results highlight the importance of the FGFR signaling pathway in the pathogenesis of breast cancer, and suggest that different genes in the same pathway may be associated with different ER breast cancer subtypes.

Genetic variations in vitamin D-related pathways and breast cancer risk in African American women in the AMBER consortium

Studies of genetic variations in vitamin D-related pathways and breast cancer risk have been conducted mostly in populations of European ancestry, and only sparsely in African Americans (AA), who are known for a high prevalence of vitamin D deficiency. We analyzed 24,445 germline variants in 63 genes from vitamin D-related pathways in the African American Breast Cancer Epidemiology and Risk (AMBER) consortium, including 3,663 breast cancer cases and 4,687 controls. Odds ratios (OR) were derived from logistic regression models for overall breast cancer, by estrogen receptor (ER) status (1,983 ER positive and 1,098 ER negative), and for case-only analyses of ER status. None of the three vitamin D-related pathways were associated with breast cancer risk overall or by ER status. Gene-level analyses identified associations with risk for several genes at a nominal p ≤ 0.05, particularly for ER- breast cancer, including rs4647707 in DDB2. In case-only analyses, vitamin D metabolism and signaling pathways were associated with ER- cancer (pathway-level p = 0.02), driven by a single gene CASR (gene-level p = 0.001). The top SNP in CASR was rs112594756 (p = 7 × 10(-5), gene-wide corrected p = 0.01), followed by a second signal from a nearby SNP rs6799828 (p = 1 × 10(-4), corrected p = 0.03). In summary, several variants in vitamin D pathways were associated with breast cancer risk in AA women. In addition, CASR may be related to tumor ER status, supporting a role of vitamin D or calcium in modifying breast cancer phenotypes.

Adaptive Set-Based Methods for Association Testing

With a typical sample size of a few thousand subjects, a single genome-wide association study (GWAS) using traditional one single nucleotide polymorphism (SNP)-at-a-time methods can only detect genetic variants conferring a sizable effect on disease risk. Set-based methods, which analyze sets of SNPs jointly, can detect variants with smaller effects acting within a gene, a pathway, or other biologically relevant sets. Although self-contained set-based methods (those that test sets of variants without regard to variants not in the set) are generally more powerful than competitive set-based approaches (those that rely on comparison of variants in the set of interest with variants not in the set), there is no consensus as to which self-contained methods are best. In particular, several self-contained set tests have been proposed to directly or indirectly "adapt" to the a priori unknown proportion and distribution of effects of the truly associated SNPs in the set, which is a major determinant of their power. A popular adaptive set-based test is the adaptive rank truncated product (ARTP), which seeks the set of SNPs that yields the best-combined evidence of association. We compared the standard ARTP, several ARTP variations we introduced, and other adaptive methods in a comprehensive simulation study to evaluate their performance. We used permutations to assess significance for all the methods and thus provide a level playing field for comparison. We found the standard ARTP test to have the highest power across our simulations followed closely by the global model of random effects (GMRE) and a least absolute shrinkage and selection operator (LASSO)-based test.

Regionally Smoothed Meta-Analysis Methods for GWAS Datasets

Genome-wide association studies are proven tools for finding disease genes, but it is often necessary to combine many cohorts into a meta-analysis to detect statistically significant genetic effects. Often the component studies are performed by different investigators on different populations, using different chips with minimal SNPs overlap. In some cases, raw data are not available for imputation so that only the genotyped single nucleotide polymorphisms (SNPs) results can be used in meta-analysis. Even when SNP sets are comparable, different cohorts may have peak association signals at different SNPs within the same gene due to population differences in linkage disequilibrium or environmental interactions. We hypothesize that the power to detect statistical signals in these situations will improve by using a method that simultaneously meta-analyzes and smooths the signal over nearby markers. In this study, we propose regionally smoothed meta-analysis methods and compare their performance on real and simulated data.

Genetic variants in the mTOR pathway and breast cancer risk in African American women

The phosphatidylinositol 3-kinase-AKT-mammalian target of rapamycin (mTOR) pathway has been implicated in breast carcinogenesis. However, there has been no large-scale investigation of genetic variants in the mTOR pathway and breast cancer risk. We examined 28847 single-nucleotide polymorphisms (SNPs) in 61 mTOR pathway genes in the African American Breast Cancer Epidemiology and Risk consortium of 3663 cases [1983 estrogen receptor-positive (ER+) and 1098 ER-negative (ER-)] and 4687 controls. Gene-level analyses were conducted using the adaptive rank truncated product (ARTP) test for 10773 SNPs that were not highly correlated (r (2) < 0.8), and SNP-level analyses were conducted with logistic regression. Among genes that were prioritized (nominal P < 0.05, ARTP tests), associations were observed for intronic SNPs TSC2 rs181088346 [odds ratio (OR) of each copy of variant allele = 0.77, 95% confidence interval (CI) = 0.65-0.88 for all breast cancer] and BRAF rs114729114 (OR = 1.53, 95% CI = 1.24-1.91 for all breast cancer and OR = 2.03, 95% CI = 1.50-2.76 for ER- tumors). For ER- tumors, intronic SNPs PGF rs11542848 (OR = 1.38, 95% CI = 1.15-1.66) and rs61759375 (OR = 1.34, 95% CI = 1.14-1.57) and MAPK3 rs78564187 (OR = 1.26, 95% CI = 1.11-1.43) were associated with increased risk. These SNPs were significant at a gene-wide level (Bonferroni-corrected P < 0.05). The variant allele of RPS6KB2 rs35363135, a synonymous coding SNP, was more likely to be observed in ER- than ER+ tumors (OR = 1.18, 95% CI = 1.05-1.31, gene-wide Bonferroni-corrected P = 0.06). In conclusion, specific mTOR pathway genes are potentially important to breast cancer risk and to the ER negativity in African American women.

META-GSA: Combining Findings from Gene-Set Analyses across Several Genome-Wide Association Studies

INTRODUCTION

Gene-set analysis (GSA) methods are used as complementary approaches to genome-wide association studies (GWASs). The single marker association estimates of a predefined set of genes are either contrasted with those of all remaining genes or with a null non-associated background. To pool the p-values from several GSAs, it is important to take into account the concordance of the observed patterns resulting from single marker association point estimates across any given gene set. Here we propose an enhanced version of Fisher's inverse χ2-method META-GSA, however weighting each study to account for imperfect correlation between association patterns.

SIMULATION AND POWER

We investigated the performance of META-GSA by simulating GWASs with 500 cases and 500 controls at 100 diallelic markers in 20 different scenarios, simulating different relative risks between 1 and 1.5 in gene sets of 10 genes. Wilcoxon's rank sum test was applied as GSA for each study. We found that META-GSA has greater power to discover truly associated gene sets than simple pooling of the p-values, by e.g. 59% versus 37%, when the true relative risk for 5 of 10 genes was assume to be 1.5. Under the null hypothesis of no difference in the true association pattern between the gene set of interest and the set of remaining genes, the results of both approaches are almost uncorrelated. We recommend not relying on p-values alone when combining the results of independent GSAs.

APPLICATION

We applied META-GSA to pool the results of four case-control GWASs of lung cancer risk (Central European Study and Toronto/Lunenfeld-Tanenbaum Research Institute Study; German Lung Cancer Study and MD Anderson Cancer Center Study), which had already been analyzed separately with four different GSA methods (EASE; SLAT, mSUMSTAT and GenGen). This application revealed the pathway GO0015291 "transmembrane transporter activity" as significantly enriched with associated genes (GSA-method: EASE, p = 0.0315 corrected for multiple testing). Similar results were found for GO0015464 "acetylcholine receptor activity" but only when not corrected for multiple testing (all GSA-methods applied; p ≈ 0.02).

In Search for the Genetic Basis of Quality of Life in Healthy Swedish Women--A GWAS Study Using the iCOGS Custom Genotyping Array

BACKGROUND

Quality of life (QoL) is increasingly measured in both research and clinical practice. QoL-assessments are built on a long, empirically-based, and stringent approach. There is ample evidence that QoL is, in part, heritable. We therefore performed a GWAS relating genetic variation to QoL in healthy females.

METHODS

In 5,142 healthy females, background characteristics (e.g. demographic, clinical, lifestyle and psychological factors) and QoL by means of the EORTC QLQ-C30 were measured. Moreover, women were genotyped using a custom array including ~210,000 single nucleotide polymorphisms (SNPs). Initially, SNPs were related to each QoL-domain, by means of partially adjusted (controlling for age and population stratification) and fully adjusted (controlling for age, population stratification, and background characteristics) regression analyses. Additionally, gene-based analyses were performed relating the combined effect of SNPs within each gene to QoL using the statistical software package VEGAS.

RESULTS

None of the associations between QoL and genetic variation (i.e. individual SNPs and genes) reached the bonferroni corrected significance level.

CONCLUSION

Reasons for a lack of association between genetic markers and QoL could be low variation in QoL-scores; selecting genetic markers not tagging QoL; or that the genetic effect that impacts one's QoL is mediated through biological pathways rather than the effect of single SNPs or genes. Therefore, we opt for a pathway-based or system biology approach as a complementary and powerful approach to analyze the combined effect of genes and their biological implications in future studies focusing on QoL-issues.

Hormone-related pathways and risk of breast cancer subtypes in African American women

We sought to investigate genetic variation in hormone pathways in relation to risk of overall and subtype-specific breast cancer in women of African ancestry (AA). Genotyping and imputation yielded data on 143,934 SNPs in 308 hormone-related genes for 3663 breast cancer cases (1098 ER-, 1983 ER+, 582 ER unknown) and 4687 controls from the African American Breast Cancer Epidemiology and Risk (AMBER) Consortium. AMBER includes data from four large studies of AA women: the Carolina Breast Cancer Study, the Women's Circle of Health Study, the Black Women's Health Study, and the Multiethnic Cohort Study. Pathway- and gene-based analyses were conducted, and single-SNP tests were run for the top genes. There were no strong associations at the pathway level. The most significantly associated genes were GHRH, CALM2, CETP, and AKR1C1 for overall breast cancer (gene-based nominal p ≤ 0.01); NR0B1, IGF2R, CALM2, CYP1B1, and GRB2 for ER+ breast cancer (p ≤ 0.02); and PGR, MAPK3, MAP3K1, and LHCGR for ER- disease (p ≤ 0.02). Single-SNP tests for SNPs with pairwise linkage disequilibrium r (2) < 0.8 in the top genes identified 12 common SNPs (in CALM2, CETP, NR0B1, IGF2R, CYP1B1, PGR, MAPK3, and MAP3K1) associated with overall or subtype-specific breast cancer after gene-level correction for multiple testing. Rs11571215 in PGR (progesterone receptor) was the SNP most strongly associated with ER- disease. We identified eight genes in hormone pathways that contain common variants associated with breast cancer in AA women after gene-level correction for multiple testing.

Gene set analysis: A step-by-step guide

To maximize the potential of genome-wide association studies, many researchers are performing secondary analyses to identify sets of genes jointly associated with the trait of interest. Although methods for gene-set analyses (GSA), also called pathway analyses, have been around for more than a decade, the field is still evolving. There are numerous algorithms available for testing the cumulative effect of multiple SNPs, yet no real consensus in the field about the best way to perform a GSA. This paper provides an overview of the factors that can affect the results of a GSA, the lessons learned from past studies, and suggestions for how to make analysis choices that are most appropriate for different types of data. © 2015 Wiley Periodicals, Inc.

Energy homeostasis genes and breast cancer risk: The influence of ancestry, body size, and menopausal status, the breast cancer health disparities study

BACKGROUND

Obesity and breast cancer risk is multifaceted and genes associated with energy homeostasis may modify this relationship.

METHODS

We evaluated 10 genes that have been associated with obesity and energy homeostasis to determine their association with breast cancer risk in Hispanic/Native American (2111 cases, 2597 controls) and non-Hispanic white (1481 cases, 1585 controls) women.

RESULTS

Cholecystokinin (CCK) rs747455 and proopiomelanocortin (POMC) rs6713532 and rs7565877 (for low Indigenous American (IA) ancestry); CCK rs8192472 and neuropeptide Y (NYP) rs16141 and rs14129 (intermediate IA ancestry); and leptin receptor (LEPR) rs11585329 (high IA ancestry) were strongly associated with multiple indicators of body size. There were no significant associations with breast cancer risk between genes and SNPs overall. However, LEPR was significantly associated with breast cancer risk among women with low IA ancestry (PARTP=0.024); POMC was significantly associated with breast cancer risk among women with intermediate (PARTP=0.015) and high (PARTP=0.012) IA ancestry. The overall pathway was statistically significant for pre-menopausal women with low IA ancestry (PARTP=0.05), as was cocaine and amphetamine regulated transcript protein (CARTPT) (PARTP=0.014) and ghrelin (GHRL) (PARTP=0.007). POMC was significantly associated with breast cancer risk among post-menopausal women with higher IA ancestry (PARTP=0.005). Three SNPs in LEPR (rs6704167, rs17412175, and rs7626141), and adiponectin (ADIPOQ); rs822391) showed significant 4-way interactions (GxExMenopausexAncestry) for multiple indicators of body size among pre-menopausal women.

CONCLUSIONS

Energy homeostasis genes were associated with breast cancer risk; menopausal status, body size, and genetic ancestry influenced this relationship.

A Novel Statistic for Global Association Testing Based on Penalized Regression

Natural genetic structures like genes may contain multiple variants that work as a group to determine a biologic outcome. The effect of rare variants, mutations occurring in less than 5% of samples, is hypothesized to be explained best as groups collectively associated with a biologic function. Therefore, it is important to develop powerful association tests to identify a true association between an outcome of interest and a group of variants, in particular a group with many rare variants. In this article we first delineate a novel penalized regression-based global test for the association between sets of variants and a disease phenotype. Next, we use Genetic Analysis Workshop 18 (GAW18) data to assess the power of the new global association test to capture a relationship between an aggregated group of variants and a simulated hypertension status. Rare variant only, common variant only, and combined variant groups are studied. The power values are compared to those obtained from eight well-regarded global tests (Score, Sum, SSU, SSUw, UminP, aSPU, aSPUw, and sequence kernel association test (SKAT)) that do not use penalized regression and a set of tests using either the SSU or score statistics and least absolute shrinkage and selection operator penalty (LASSO) logistic regression. Association testing of rare variants with our method was the top performer when there was low linkage disequilibrium (LD) between and within causal variants. This was similarly true when simultaneously testing rare and common variants in low LD scenarios. Finally, our method was able to provide meaningful variant-specific association information.

Genome-Wide Association Study of Behavioral Disinhibition in a Selected Adolescent Sample

Behavioral disinhibition (BD) is a quantitative measure designed to capture the heritable variation encompassing risky and impulsive behaviors. As a result, BD represents an ideal target for discovering genetic loci that predispose individuals to a wide range of antisocial behaviors and substance misuse that together represent a large cost to society as a whole. Published genome-wide association studies (GWAS) have examined specific phenotypes that fall under the umbrella of BD (e.g. alcohol dependence, conduct disorder); however no GWAS has specifically examined the overall BD construct. We conducted a GWAS of BD using a sample of 1,901 adolescents over-selected for characteristics that define high BD, such as substance and antisocial behavior problems, finding no individual locus that surpassed genome-wide significance. Although no single SNP was significantly associated with BD, restricted maximum likelihood analysis estimated that 49.3 % of the variance in BD within the Caucasian sub-sample was accounted for by the genotyped SNPs (p = 0.06). Gene-based tests identified seven genes associated with BD (p ≤ 2.0 × 10(-6)). Although the current study was unable to identify specific SNPs or pathways with replicable effects on BD, the substantial sample variance that could be explained by all genotyped SNPs suggests that larger studies could successfully identify common variants associated with BD.

Genetic contributions to urgency urinary incontinence in women

PURPOSE

We identify genetic variants associated with urgency urinary incontinence in postmenopausal women.

MATERIALS AND METHODS

A 2-stage genome-wide association analysis was conducted to identify variants associated with urgency urinary incontinence. The WHI GARNET substudy with 4,894 genotyped post-reproductive white women was randomly split into independent discovery and replication cohorts. Genome-wide imputation was performed using IMPUTE2 with the 1000 Genomes ALL Phase I integrated variant set as a reference. Controls reported no urgency urinary incontinence at enrollment or followup. Cases reported monthly or greater urgency urinary incontinence and leaked sufficiently to wet/soak underpants/clothes. Logistic regression models were used to predict urgency urinary incontinence case vs control status based on genotype, assuming additive inheritance. Age, obesity, diabetes and depression were included in the models as covariates.

RESULTS

Following quality control, 975,508 single nucleotide polymorphisms in 2,241 cases (discovery 1,102; replication 1,133) and 776 controls (discovery 405, replication 371) remained. Genotype imputation resulted in 9,077,347 single nucleotide polymorphisms and insertions/deletions with minor allele frequency greater than 0.01 available for analysis. Meta-analysis of the discovery and replication samples identified 6 loci on chromosomes 5, 10, 11, 12 and 18 associated with urgency urinary incontinence at p <10(-6). Of the loci 3 were within genes, the zinc finger protein 521 (ZFP521) gene on chromosome 18q11, the ADAMTS16 gene on chromosome 5p15 and the CIT gene on chromosome 12q24. The other 3 loci were intergenic.

CONCLUSIONS

Although environmental factors also likely contribute, this first exploratory genome-wide association study for urgency urinary incontinence suggests that genetic variants in the ZFP521, CIT and ADAMTS16 genes might account for some of the observed heritability of the condition.

A comparison of aggregate p-value methods and multivariate statistics for self-contained tests of metabolic pathway analysis

For pathway analysis of genomic data, the most common methods involve combining p-values from individual statistical tests. However, there are several multivariate statistical methods that can be used to test whether a pathway has changed. Because of the large number of variables and pathway sizes in genomics data, some of these statistics cannot be computed. However, in metabolomics data, the number of variables and pathway sizes are typically much smaller, making such computations feasible. Of particular interest is being able to detect changes in pathways that may not be detected for the individual variables. We compare the performance of both the p-value methods and multivariate statistics for self-contained tests with an extensive simulation study and a human metabolomics study. Permutation tests, rather than asymptotic results are used to assess the statistical significance of the pathways. Furthermore, both one and two-sided alternatives hypotheses are examined. From the human metabolomic study, many pathways were statistically significant, although the majority of the individual variables in the pathway were not. Overall, the p-value methods perform at least as well as the multivariate statistics for these scenarios.

Common genetic variants in epigenetic machinery genes and risk of upper gastrointestinal cancers

BACKGROUND

Populations in north central China are at high risk for oesophageal squamous cell carcinoma (ESCC) and gastric cancer (GC), and genetic variation in epigenetic machinery genes and pathways may contribute to this risk.

METHODS

We used the adaptive multilocus joint test to analyse 192 epigenetic genes involved in chromatin remodelling, DNA methylation and microRNA biosynthesis in 1942 ESCC and 1758 GC cases [1126 cardia (GCA) and 632 non-cardia adenocarcinoma (GNCA)] and 2111 controls with Chinese ancestry. We examined potential function of risk alleles using in silico and expression quantitative trait loci (eQTLs) analyses.

RESULTS

Suggestive pathway-based associations were observed for the overall epigenetic (P-value(PATH) = 0.034) and chromatin remodelling (P-value(PATH) = 0.039) pathways with risk of GCA, but not GC, GNCA or ESCC. Overall, 37 different epigenetic machinery genes were associated with risk of one or more upper gastrointestinal (UGI) cancer sites (P-value(GENE )< 0.05), including 14 chromatin remodelling genes whose products are involved in the regulation of HOX genes. We identified a gastric eQTL (rs12724079; rho = 0.37; P = 0.0006) which regulates mRNA expression of ASH1L. Several suggestive eQTLs were also found in oesophageal (rs10898459 in EED), gastric cardia (rs7157322 in DICER1; rs8179271 in ASH1L), and gastric non-cardia (rs1790733 in PPP1CA) tissues.

CONCLUSIONS

Results of our analyses provide limited but suggestive evidence for a role of epigenetic gene variation in the aetiology of UGI cancer.

Vitamin D metabolic pathway genes and pancreatic cancer risk

Evidence on the association between vitamin D status and pancreatic cancer risk is inconsistent. This inconsistency may be partially attributable to variation in vitamin D regulating genes. We selected 11 vitamin D-related genes (GC, DHCR7, CYP2R1, VDR, CYP27B1, CYP24A1, CYP27A1, RXRA, CRP2, CASR and CUBN) totaling 213 single nucleotide polymorphisms (SNPs), and examined associations with pancreatic adenocarcinoma. Our study included 3,583 pancreatic cancer cases and 7,053 controls from the genome-wide association studies of pancreatic cancer PanScans-I-III. We used the Adaptive Joint Test and the Adaptive Rank Truncated Product statistic for pathway and gene analyses, and unconditional logistic regression for SNP analyses, adjusting for age, sex, study and population stratification. We examined effect modification by circulating vitamin D concentration (≤50, >50 nmol/L) for the most significant SNPs using a subset of cohort cases (n = 713) and controls (n = 878). The vitamin D metabolic pathway was not associated with pancreatic cancer risk (p = 0.830). Of the individual genes, none were associated with pancreatic cancer risk at a significance level of p<0.05. SNPs near the VDR (rs2239186), LRP2 (rs4668123), CYP24A1 (rs2762932), GC (rs2282679), and CUBN (rs1810205) genes were the top SNPs associated with pancreatic cancer (p-values 0.008-0.037), but none were statistically significant after adjusting for multiple comparisons. Associations between these SNPs and pancreatic cancer were not modified by circulating concentrations of vitamin D. These findings do not support an association between vitamin D-related genes and pancreatic cancer risk. Future research should explore other pathways through which vitamin D status might be associated with pancreatic cancer risk.

Resampling-based multiple comparison procedure with application to point-wise testing with functional data

In this paper we describe a coherent multiple testing procedure for correlated test statistics such as are encountered in functional linear models. The procedure makes use of two different p-value combination methods: the Fisher combination method and the Šidák correction-based method. P-values for Fisher's and Šidák's test statistics are estimated through resampling to cope with the correlated tests. Building upon these two existing combination methods, we propose the smallest p-value as a new test statistic for each hypothesis. The closure principle is incorporated along with the new test statistic to obtain the overall p-value and appropriately adjust the individual p-values. Furthermore, a shortcut version for the proposed procedure is detailed, so that individual adjustments can be obtained even for a large number of tests. The motivation for developing the procedure comes from a problem of point-wise inference with smooth functional data where tests at neighboring points are related. A simulation study verifies that the methodology performs well in this setting. We illustrate the proposed method with data from a study on the aerial detection of the spectral effect of below ground carbon dioxide leakage on vegetation stress via spectral responses.

Genetic variants in the inositol phosphate metabolism pathway and risk of different types of cancer

Members of the inositol phosphate metabolism pathway regulate cell proliferation, migration and phosphatidylinositol-3-kinase (PI3K)/Akt signaling, and are frequently dysregulated in cancer. Whether germline genetic variants in inositol phosphate metabolism pathway are associated with cancer risk remains to be clarified. We examined the association between inositol phosphate metabolism pathway genes and risk of eight types of cancer using data from genome-wide association studies. Logistic regression models were applied to evaluate SNP-level associations. Gene- and pathway-based associations were tested using the permutation-based adaptive rank-truncated product method. The overall inositol phosphate metabolism pathway was significantly associated with risk of lung cancer (P = 2.00 × 10⁻⁴), esophageal squamous cell carcinoma (P = 5.70 × 10⁻³), gastric cancer (P = 3.03 × 10⁻²) and renal cell carcinoma (P = 1.26 × 10⁻²), but not with pancreatic cancer (P = 1.40 × 10⁻¹), breast cancer (P = 3.03 × 10⁻¹), prostate cancer (P = 4.51 × 10⁻¹), and bladder cancer (P = 6.30 × 10⁻¹). Our results provide a link between inherited variation in the overall inositol phosphate metabolism pathway and several individual genes and cancer. Further studies will be needed to validate these positive findings, and to explore its mechanisms.

MAPK genes interact with diet and lifestyle factors to alter risk of breast cancer: the Breast Cancer Health Disparities Study

Mitogen-activated protein kinases (MAPK) are integration points for multiple biochemical signals. We evaluated 13 MAPK genes with breast cancer risk and determined if diet and lifestyle factors mediated risk. Data from 3 population-based case-control studies conducted in Southwestern United States, California, and Mexico included 4183 controls and 3592 cases. Percent Indigenous American (IA) ancestry was determined from 104 ancestry informative markers. The adaptive rank truncated product (ARTP) was used to determine the significance of each gene and the pathway with breast cancer risk, by menopausal status, genetic ancestry level, and estrogen receptor (ER)/progesterone receptor (PR) strata. MAP3K9 was associated with breast cancer overall (P(ARTP) = 0.02) with strongest association among women with the highest IA ancestry (P(ARTP) = 0.04). Several SNPs in MAP3K9 were associated with ER+/PR+ tumors and interacted with dietary oxidative balance score (DOBS), dietary folate, body mass index (BMI), alcohol consumption, cigarette smoking, and a history of diabetes. DUSP4 and MAPK8 interacted with calories to alter breast cancer risk; MAPK1 interacted with DOBS, dietary fiber, folate, and BMI; MAP3K2 interacted with dietary fat; and MAPK14 interacted with dietary folate and BMI. The patterns of association across diet and lifestyle factors with similar biological properties for the same SNPs within genes provide support for associations.

Detection for pathway effect contributing to disease in systems epidemiology with a case-control design

OBJECTIVES

Identification of pathway effects responsible for specific diseases has been one of the essential tasks in systems epidemiology. Despite some advance in procedures for distinguishing specific pathway (or network) topology between different disease status, statistical inference at a population level remains unsolved and further development is still needed. To identify the specific pathways contributing to diseases, we attempt to develop powerful statistics which can capture the complex relationship among risk factors.

SETTING AND PARTICIPANTS

Acute myeloid leukaemia (AML) data obtained from 133 adults (98 patients and 35 controls; 47% female).

RESULTS

Simulation studies indicated that the proposed Pathway Effect Measures (PEM) were stable; bootstrap-based methods outperformed the others, with bias-corrected bootstrap CI method having the highest power. Application to real data of AML successfully identified the specific pathway (Treg→TGFβ→Th17) effect contributing to AML with p values less than 0.05 under various methods and the bias-corrected bootstrap CI (-0.214 to -0.020). It demonstrated that Th17-Treg correlation balance was impaired in patients with AML, suggesting that Th17-Treg imbalance potentially plays a role in the pathogenesis of AML.

CONCLUSIONS

The proposed bootstrap-based PEM are valid and powerful for detecting the specific pathway effect contributing to disease, thus potentially providing new insight into the underlying mechanisms and ways to study the disease effects of specific pathways more comprehensively.

The influence of the CHIEF pathway on colorectal cancer-specific mortality

Many components of the CHIEF (Convergence of Hormones, Inflammation, and Energy Related Factors) pathway could influence survival given their involvement in cell growth, apoptosis, angiogenesis, and tumor invasion stimulation. We used ARTP (Adaptive Rank Truncation Product) to test if genes in the pathway were associated with colorectal cancer-specific mortality. Colon cancer (n = 1555) and rectal cancer (n = 754) cases were followed over five years. Age, center, stage at diagnosis, and tumor molecular phenotype were considered when calculating ARTP p values. A polygenic risk score was used to summarize the magnitude of risk associated with this pathway. The JAK/STAT/SOC was significant for colon cancer survival (PARTP = 0.035). Fifteen genes (DUSP2, INFGR1, IL6, IRF2, JAK2, MAP3K10, MMP1, NFkB1A, NOS2A, PIK3CA, SEPX1, SMAD3, TLR2, TYK2, and VDR) were associated with colon cancer mortality (PARTP < 0.05); JAK2 (PARTP  = 0.0086), PIK3CA (PARTP = 0.0098), and SMAD3 (PARTP = 0.0059) had the strongest associations. Over 40 SNPs were significantly associated with survival within the 15 significant genes (PARTP < 0.05). SMAD3 had the strongest association with survival (HRGG 2.46 95% CI 1.44,4.21 PTtrnd = 0.0002). Seven genes (IL2RA, IL8RA, IL8RB, IRF2, RAF1, RUNX3, and SEPX1) were significantly associated with rectal cancer (PARTP < 0.05). The HR for colorectal cancer-specific mortality among colon cancer cases in the upper at-risk alleles group was 11.81 (95% CI 7.07, 19. 74) and was 10.99 (95% CI 5.30, 22.78) for rectal cancer. These results suggest that several genes in the CHIEF pathway are important for colorectal cancer survival; the risk associated with the pathway merits validation in other studies.

Hormone metabolism pathway genes and mammographic density change after quitting estrogen and progestin combined hormone therapy in the California Teachers Study

INTRODUCTION

Mammographic density (MD) is a strong biomarker of breast cancer risk. MD increases after women start estrogen plus progestin therapy (EPT) and decreases after women quit EPT. A large interindividual variation in EPT-associated MD change has been observed, but few studies have investigated genetic predictors of the EPT-associated MD change. Here, we evaluate the association between polymorphisms in hormone metabolism pathway genes and MD changes when women quit EPT.

METHODS

We collected mammograms before and after women quit EPT and genotyped 405 tagging single nucleotide polymorphisms (SNPs) in 30 hormone metabolism pathway genes in 284 non-Hispanic white participants of the California Teachers Study (CTS). Participants were ages 49 to 71 years at time of mammography taken after quitting EPT. We assessed percent MD using a computer-assisted method. MD change was calculated by subtracting MD of an 'off-EPT' mammogram from MD of an 'on-EPT' (that is baseline) mammogram. Linear regression analysis was used to investigate the SNP-MD change association, adjusting for the baseline 'on-EPT' MD, age and BMI at time of baseline mammogram, and time interval and BMI change between the two mammograms. An overall pathway and gene-level summary was obtained using the adaptive rank truncated product (ARTP) test. We calculated 'P values adjusted for correlated tests (P(ACT))' to account for multiple testing within a gene.

RESULTS

The strongest associations were observed for rs7489119 in SLCO1B1, and rs5933863 in ARSC. SLCO1B1 and ARSC are involved in excretion and activation of estrogen metabolites of EPT, respectively. MD change after quitting was 4.2% smaller per minor allele of rs7489119 (P = 0.0008; P(ACT) = 0.018) and 1.9% larger per minor allele of rs5933863 (P = 0.013; P(ACT) = 0.025). These individual SNP associations did not reach statistical significance when we further used Bonferroni correction to consider the number of tested genes. The pathway level summary ARTP P value was not statistically significant.

CONCLUSIONS

Data from this longitudinal study of EPT quitters suggest that genetic variation in two hormone metabolism pathway genes, SLCO1B1 and ARSC, may be associated with change in MD after women stop using EPT. Larger longitudinal studies are needed to confirm our findings.

Diet and lifestyle factors modify immune/inflammation response genes to alter breast cancer risk and prognosis: the Breast Cancer Health Disparities Study

Tumor necrosis factor-α (TNF) and toll-like receptors (TLR) are important mediators of inflammation. We examined 10 of these genes with respect to breast cancer risk and mortality in a genetically admixed population of Hispanic/Native American (NA) (2111 cases, 2597 controls) and non-Hispanic white (NHW) (1481 cases, 1585 controls) women. Additionally, we explored if diet and lifestyle factors modified associations with these genes. Overall, these genes (collectively) were associated with breast cancer risk among women with >70% NA ancestry (P(ARTP) = 0.0008), with TLR1 rs7696175 being the primary risk contributor (OR 1.77, 95% CI 1.25, 2.51). Overall, TLR1 rs7696175 (HR 1.40, 95% CI 1.03, 1.91; P(adj) = 0.032), TLR4 rs5030728 (HR 1.96, 95% CI 1.30, 2.95; P(adj) = 0.014), and TNFRSF1A rs4149578 (HR 2.71, 95% CI 1.28, 5.76; P(adj) = 0.029) were associated with increased breast cancer mortality. We observed several statistically significant interactions after adjustment for multiple comparisons, including interactions between our dietary oxidative balance score and CD40LG and TNFSF1A; between cigarette smoking and TLR1, TLR4, and TNF; between body mass index (BMI) among pre-menopausal women and TRAF2; and between regular use of aspirin/non-steroidal anti-inflammatory drugs and TLR3 and TRA2. In conclusion, our findings support a contributing role of certain TNF-α and TLR genes in both breast cancer risk and survival, particularly among women with higher NA ancestry. Diet and lifestyle factors appear to be important mediators of the breast cancer risk associated with these genes.

A set-based association test identifies sex-specific gene sets associated with type 2 diabetes

Single variant analysis in genome-wide association studies (GWAS) has been proven to be successful in identifying thousands of genetic variants associated with hundreds of complex diseases. However, these identified variants only explain a small fraction of inheritable variability in many diseases, suggesting that other resources, such as multilevel genetic variations, may contribute to disease susceptibility. In this work, we proposed to combine genetic variants that belong to a gene set, such as at gene- and pathway-level to form an integrated signal aimed to identify major players that function in a coordinated manner conferring disease risk. The integrated analysis provides novel insight into disease etiology while individual signals could be easily missed by single variant analysis. We applied our approach to a genome-wide association study of type 2 diabetes (T2D) with male and female data analyzed separately. Novel sex-specific genes and pathways were identified to increase the risk of T2D. We also demonstrated the performance of signal integration through simulation studies.

A method for gene-based pathway analysis using genomewide association study summary statistics reveals nine new type 1 diabetes associations

Pathway analysis can complement point-wise single nucleotide polymorphism (SNP) analysis in exploring genomewide association study (GWAS) data to identify specific disease-associated genes that can be candidate causal genes. We propose a straightforward methodology that can be used for conducting a gene-based pathway analysis using summary GWAS statistics in combination with widely available reference genotype data. We used this method to perform a gene-based pathway analysis of a type 1 diabetes (T1D) meta-analysis GWAS (of 7,514 cases and 9,045 controls). An important feature of the conducted analysis is the removal of the major histocompatibility complex gene region, the major genetic risk factor for T1D. Thirty-one of the 1,583 (2%) tested pathways were identified to be enriched for association with T1D at a 5% false discovery rate. We analyzed these 31 pathways and their genes to identify SNPs in or near these pathway genes that showed potentially novel association with T1D and attempted to replicate the association of 22 SNPs in additional samples. Replication P-values were skewed () with 12 of the 22 SNPs showing . Support, including replication evidence, was obtained for nine T1D associated variants in genes ITGB7 (rs11170466, ), NRP1 (rs722988, ), BAD (rs694739, ), CTSB (rs1296023, ), FYN (rs11964650, ), UBE2G1 (rs9906760, ), MAP3K14 (rs17759555, ), ITGB1 (rs1557150, ), and IL7R (rs1445898, ). The proposed methodology can be applied to other GWAS datasets for which only summary level data are available.

Pathway-based analysis tools for complex diseases: a review

Genetic studies are traditionally based on single-gene analysis. The use of these analyses can pose tremendous challenges for elucidating complicated genetic interplays involved in complex human diseases. Modern pathway-based analysis provides a technique, which allows a comprehensive understanding of the molecular mechanisms underlying complex diseases. Extensive studies utilizing the methods and applications for pathway-based analysis have significantly advanced our capacity to explore large-scale omics data, which has rapidly accumulated in biomedical fields. This article is a comprehensive review of the pathway-based analysis methods—the powerful methods with the potential to uncover the biological depths of the complex diseases. The general concepts and procedures for the pathway-based analysis methods are introduced and then, a comprehensive review of the major approaches for this analysis is presented. In addition, a list of available pathway-based analysis software and databases is provided. Finally, future directions and challenges for the methodological development and applications of pathway-based analysis techniques are discussed. This review will provide a useful guide to dissect complex diseases.

Associations between ALOX, COX, and CRP polymorphisms and breast cancer among Hispanic and non-Hispanic white women: The breast cancer health disparities study

Chronic inflammation is suggested to be associated with specific cancer sites, including breast cancer. Recent research has focused on the roles of genes involved in the leukotriene/lipoxygenase and prostaglandin/cyclooxygenase pathways in breast cancer etiology. We hypothesized that genes in ALOX/COX pathways and CRP polymorphisms would be associated with breast cancer risk and mortality in our sample of Hispanic/Native American (NA) (1430 cases, 1599 controls) and non-Hispanic white (NHW) (2093 cases, 2610 controls) women. A total of 104 Ancestral Informative Markers was used to distinguish European and NA ancestry. The adaptive rank truncated product (ARTP) method was used to determine the significance of associations for each gene and the inflammation pathway with breast cancer risk and by NA ancestry. Overall, the pathway was associated with breast cancer risk (PARTP = 0.01). Two-way interactions with NA ancestry (P(adj)  < 0.05) were observed for ALOX12 (rs2292350, rs2271316) and PTGS1 (rs10306194). We observed increases in breast cancer risk in stratified analyses by tertiles of polyunsaturated fat intake for ALOX12 polymorphisms; the largest increase in risk was among women in the highest tertile with ALOX12 rs9904779CC (Odds Ratio (OR), 1.49; 95% Confidence Interval (CI) 1.14-1.94, P(adj) = 0.01). In a sub-analysis stratified by NSAIDs use, two-way interactions with NSAIDs use were found for ALOX12 rs9904779 (P(adj)  = 0.02), rs434473 (P(adj ) = 0.02), and rs1126667 (P(adj)  =  0.01); ORs for ALOX12 polymorphisms ranged from 1.55 to 1.64 among regular users. Associations were not observed with breast cancer mortality. These findings could support advances in the discovery of new pathways related to inflammation for breast cancer treatment.

A pathway approach to evaluating the association between the CHIEF pathway and risk of colorectal cancer

Inflammation, hormones and energy-related factors have been associated with colorectal cancer (CRC) and it has been proposed that convergence and interactions of these factors importantly influence CRC risk. We have previously hypothesized that genetic variation in the CHIEF (convergence of hormones, inflammation and energy-related factors) pathway would influence risk of CRC. In this paper, we utilize an Adaptive Rank Truncation Product (ARTP) statistical method to determine the overall pathway significance and then use that method to identify the key elements within the pathway associated with disease risk. Data from two population-based case-control studies of colon (n = 1555 cases and 1956 controls) and rectal (n = 754 cases and 959 controls) cancer were used. We use ARTP to estimate pathway and gene significance and polygenic scores based on ARTP findings to further estimate the risk associated with the pathway. Associations were further assessed based on tumor molecular phenotype. The CHIEF pathway was statistically significant for colon cancer (P(ARTP)= 0.03) with the most significant interferons (P(ARTP) = 0.0253), JAK/STAT/SOCS (P(ARTP) = 0.0111), telomere (P(ARTP) = 0.0399) and transforming growth factor β (P(ARTP) = 0.0043) being the most significant subpathways for colon cancer. For rectal cancer, interleukins (P(ARTP) = 0.0235) and selenoproteins (P ARTP = 0.0047) were statistically significant although the pathway overall was of borderline significance (P(ARTP) = 0.06). Interleukins (P(ARTP) = 0.0456) and mitogen-activated protein kinase (P(ARTP) = 0.0392) subpathways were uniquely significant for CpG island methylator phenotype-positive colon tumors. Increasing number of at-risk alleles was significantly associated with both colon [odds ratio (OR) = 6.21, 95% confidence interval (CI): 4.72, 8.16] and rectal (OR = 7.82, 95% CI: 5.26, 11.62) cancer. We conclude that elements of the CHIEF pathway are important for CRC risk.

Family-based association test using both common and rare variants and accounting for directions of effects for sequencing data

Current family-based association tests for sequencing data were mainly developed for identifying rare variants associated with a complex disease. As the disease can be influenced by the joint effects of common and rare variants, common variants with modest effects may not be identified by the methods focusing on rare variants. Moreover, variants can have risk, neutral, or protective effects. Association tests that can effectively select groups of common and rare variants that are likely to be causal and consider the directions of effects have become important. We developed the Ordered Subset - Variable Threshold - Pedigree Disequilibrium Test (OVPDT), a combination of three algorithms, for association analysis in family sequencing data. The ordered subset algorithm is used to select a subset of common variants based on their relative risks, calculated using only parental mating types. The variable threshold algorithm is used to search for an optimal allele frequency threshold such that rare variants below the threshold are more likely to be causal. The PDT statistics from both rare and common variants selected by the two algorithms are combined as the OVPDT statistic. A permutation procedure is used in OVPDT to calculate the p-value. We used simulations to demonstrate that OVPDT has the correct type I error rates under different scenarios and compared the power of OVPDT with two other family-based association tests. The results suggested that OVPDT can have more power than the other tests if both common and rare variants have effects on the disease in a region.