Pathway analysis by adaptive combination of P-values

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.

Genetic polymorphisms in the 9p21 region associated with risk of multiple cancers

The chromosome 9p21 region has been implicated in the pathogenesis of multiple cancers. We analyzed 9p21 single nucleotide polymorphisms (SNPs) from eight genome-wide association studies (GWAS) with data deposited in dbGaP, including studies of esophageal squamous cell carcinoma (ESCC), gastric cancer (GC), pancreatic cancer, renal cell carcinoma (RCC), lung cancer (LC), breast cancer (BrC), bladder cancer (BC) and prostate cancer (PrC). The number of subjects ranged from 2252 (PrC) to 7619 (LC). SNP-level analyses for each cancer were conducted by logistic regression or random-effects meta-analysis. A subset-based statistical approach (ASSET) was performed to combine SNP-level P values across multiple cancers. We calculated gene-level P values using the adaptive rank truncated product method. We identified that rs1063192 and rs2157719 in the CDKN2A/2B region were significantly associated with ESCC and rs2764736 (3' of TUSC1) was associated with BC (P ≤ 2.59 × 10(-6)). ASSET analyses identified four SNPs significantly associated with multiple cancers: rs3731239 (CDKN2A intronic) with ESCC, GC and BC (P = 3.96 × 10(-) (4)); rs10811474 (3' of IFNW1) with RCC and BrC (P = 0.001); rs12683422 (LINGO2 intronic) with RCC and BC (P = 5.93 × 10(-) (4)) and rs10511729 (3' of ELAVL2) with LC and BrC (P = 8.63 × 10(-) (4)). At gene level, CDKN2B, CDKN2A and CDKN2B-AS1 were significantly associated with ESCC (P ≤ 4.70 × 10(-) (5)). Rs10511729 and rs10811474 were associated with cis-expression of 9p21 genes in corresponding cancer tissues in the expression quantitative trait loci analysis. In conclusion, we identified several genetic variants in the 9p21 region associated with the risk of multiple cancers, suggesting that this region may contribute to a shared susceptibility across different cancer types.

Diet and lifestyle factors interact with MAPK genes to influence survival: the Breast Cancer Health Disparities Study

INTRODUCTION

MAPK genes are activated by a variety of factors related to growth factors, hormones, and environmental stress.

METHODS

We evaluated associations between 13 MAPK genes and survival among 1,187 nonHispanic White and 1,155 Hispanic/Native American (NA) women diagnosed with breast cancer. We assessed the influence of diet, lifestyle, and genetic ancestry on these associations. Percent NA ancestry was determined from 104 Ancestry Informative Markers. Adaptive rank truncation product (ARTP) was used to determine gene and pathway significance.

RESULTS

Associations were predominantly observed among women with lower NA ancestry. Specifically, the mitogen-activated protein kinases (MAPK) pathway was associated with all-cause mortality (P ARTP = 0.02), but not with breast cancer-specific mortality (P ARTP = 0.10). However, MAP2K1 and MAP3K9 were associated with both breast cancer-specific and all-cause mortality. MAPK12 (P ARTP = 0.05) was only associated with breast cancer-specific mortality, and MAP3K1 (P ARTP = 0.02) and MAPK1 (P ARTP = 0.05) were only associated with all-cause mortality. Among women with higher NA ancestry, MAP3K2 was significantly associated with all-cause mortality (P ARTP = 0.04). Several diet and lifestyle factors, including alcohol consumption, caloric intake, dietary folate, and cigarette smoking, significantly modified the associations with MAPK genes and all-cause mortality.

CONCLUSIONS

Our study supports an association between MAPK genes and survival after diagnosis with breast cancer, especially among women with low NA ancestry. The interaction between genetic variation in the MAPK pathway with diet and lifestyle factors for all women supports the important role of these factors for breast cancer survivorship.

Functional and genomic context in pathway analysis of GWAS data

Gene set analysis (GSA) is a promising tool for uncovering the polygenic effects associated with complex diseases. However, the available techniques reflect a wide variety of hypotheses about how genetic effects interact to contribute to disease susceptibility. The lack of consensus about the best way to perform GSA has led to confusion in the field and has made it difficult to compare results across methods. A clear understanding of the various choices made during GSA - such as how gene sets are defined, how single-nucleotide polymorphisms (SNPs) are assigned to genes, and how individual SNP-level effects are aggregated to produce gene- or pathway-level effects - will improve the interpretability and comparability of results across methods and studies. In this review we provide an overview of the various data sources used to construct gene sets and the statistical methods used to test for gene set association, as well as provide guidelines for ensuring the comparability of results.

Global adaptive rank truncated product method for gene-set analysis in association studies

Gene set analysis (GSA) aims to assess the overall association of a set of genetic variants with a phenotype and has the potential to detect subtle effects of variants in a gene or a pathway that might be missed when assessed individually. We present a new implementation of the Adaptive Rank Truncated Product method (ARTP) for analyzing the association of a set of Single Nucleotide Polymorphisms (SNPs) in a gene or pathway. The new implementation, referred to as globalARTP, improves the original one by allowing the different SNPs in the set to have different modes of inheritance. We perform a simulation study for exploring the power of the proposed methodology in a set of scenarios with different numbers of causal SNPs with different effect sizes. Moreover, we show the advantage of using the gene set approach in the context of an Alzheimer's disease case-control study where we explore the endocytosis pathway. The new method is implemented in the R function globalARTP of the globalGSA package available at http://cran.r-project.org.

A fast and powerful tree-based association test for detecting complex joint effects in case-control studies

MOTIVATION

Multivariate tests derived from the logistic regression model are widely used to assess the joint effect of multiple predictors on a disease outcome in case-control studies. These tests become less optimal if the joint effect cannot be approximated adequately by the additive model. The tree-structure model is an attractive alternative, as it is more apt to capture non-additive effects. However, the tree model is used most commonly for prediction and seldom for hypothesis testing, mainly because of the computational burden associated with the resampling-based procedure required for estimating the significance level.

RESULTS

We designed a fast algorithm for building the tree-structure model and proposed a robust TREe-based Association Test (TREAT) that incorporates an adaptive model selection procedure to identify the optimal tree model representing the joint effect. We applied TREAT as a multilocus association test on >20 000 genes/regions in a study of esophageal squamous cell carcinoma (ESCC) and detected a highly significant novel association between the gene CDKN2B and ESCC ([Formula: see text]). We also demonstrated, through simulation studies, the power advantage of TREAT over other commonly used tests.

AVAILABILITY AND IMPLEMENTATION

 The package TREAT is freely available for download at http://www.hanzhang.name/softwares/treat, implemented in C++ and R and supported on 64-bit Linux and 64-bit MS Windows.

CONTACT

yuka@mail.nih.gov

SUPPLEMENTARY INFORMATION

 Supplementary data are available at Bioinformatics online.

Breast cancer risk, nightwork, and circadian clock gene polymorphisms

Night shift work has been associated with an increased risk of breast cancer pointing to a role of circadian disruption. We investigated the role of circadian clock gene polymorphisms and their interaction with nightwork in breast cancer risk in a population-based case-control study in France including 1126 breast cancer cases and 1174 controls. We estimated breast cancer risk associated with each of the 577 single nucleotide polymorphisms (SNPs) in 23 circadian clock genes. We also used a gene- and pathway-based approach to investigate the overall effect on breast cancer of circadian clock gene variants that might not be detected in analyses based on individual SNPs. Interactions with nightwork were tested at the SNP, gene, and pathway levels. We found that two SNPs in RORA (rs1482057 and rs12914272) were associated with breast cancer in the whole sample and among postmenopausal women. In this subpopulation, we also reported an association with rs11932595 in CLOCK, and with CLOCK, RORA, and NPAS2 in the analyses at the gene level. Breast cancer risk in postmenopausal women was also associated with overall genetic variation in the circadian gene pathway (P=0.04), but this association was not detected in premenopausal women. There was some evidence of an interaction between PER1 and nightwork in breast cancer in the whole sample (P=0.024), although the effect was not statistically significant after correcting for multiple testing (P=0.452). Our results support the hypothesis that circadian clock gene variants modulate breast cancer risk.

Pathway analysis with next-generation sequencing data

Although pathway analysis methods have been developed and successfully applied to association studies of common variants, the statistical methods for pathway-based association analysis of rare variants have not been well developed. Many investigators observed highly inflated false-positive rates and low power in pathway-based tests of association of rare variants. The inflated false-positive rates and low true-positive rates of the current methods are mainly due to their lack of ability to account for gametic phase disequilibrium. To overcome these serious limitations, we develop a novel statistic that is based on the smoothed functional principal component analysis (SFPCA) for pathway association tests with next-generation sequencing data. The developed statistic has the ability to capture position-level variant information and account for gametic phase disequilibrium. By intensive simulations, we demonstrate that the SFPCA-based statistic for testing pathway association with either rare or common or both rare and common variants has the correct type 1 error rates. Also the power of the SFPCA-based statistic and 22 additional existing statistics are evaluated. We found that the SFPCA-based statistic has a much higher power than other existing statistics in all the scenarios considered. To further evaluate its performance, the SFPCA-based statistic is applied to pathway analysis of exome sequencing data in the early-onset myocardial infarction (EOMI) project. We identify three pathways significantly associated with EOMI after the Bonferroni correction. In addition, our preliminary results show that the SFPCA-based statistic has much smaller P-values to identify pathway association than other existing methods.

Variants associated with susceptibility to pancreatic cancer and melanoma do not reciprocally affect risk

BACKGROUND

Melanoma cases may exist in pancreatic cancer kindreds, whereas there is increased risk of pancreatic cancer in familial melanoma. The two cancers may share genetic susceptibility variants in common.

METHODS

Three dbGaP (datasets in Genotypes and Phenotypes)-deposited GWAS (genome-wide association study) datasets (MD Anderson melanoma, PanScan 1, and PanScan 2 for pancreatic cancer) were used. Thirty-seven melanoma susceptibility variants in 22 genomic regions from published GWAS, plus melanoma-related genes and pathways were examined for pancreatic cancer risk in the PanScan datasets. Conversely, nine known pancreatic cancer susceptibility variants were examined for melanoma risk in the MD Anderson dataset.

RESULTS

In the PanScan data, initial associations were found with melanoma susceptibility variants in NCOA6 [rs4911442; OR, 1.32; 95% confidence interval (CI), 1.03-1.70; P = 0.03], YWHAZP5 (rs17119461; OR, 2.62; 95% CI, 1.08-6.35; P = 0.03), and YWHAZP5 (rs17119490; OR, 2.62; 95% CI, 1.08-6.34; P = 0.03), TYRP1 (P = 0.04), and IFNA13 (P = 0.04). In the melanoma dataset, two pancreatic cancer susceptibility variants were associated: NR5A2 (rs12029406; OR, 1.39; 95% CI, 1.01-1.92; P = 0.04) and CLPTM1L-TERT (rs401681; OR, 1.16; 95% CI, 1.01-1.34; P = 0.04). None of these associations remained significant after correcting for multiple comparisons.

CONCLUSION

Reported variants of melanoma genes and pathways do not play a role in pancreatic cancer predisposition. Reciprocally, pancreatic cancer susceptibility variants are not associated with melanoma risk.

IMPACT

Known melanoma-related genes and pathways, as well as GWAS-derived susceptibility variants of melanoma and pancreatic cancer, do not explain the shared genetic etiology of these two cancers. Cancer Epidemiol Biomarkers Prev; 23(6); 1121-4. ©2014 AACR.

Genome-wide association study of circulating vitamin D-binding protein

BACKGROUND

Vitamin D status may influence a spectrum of health outcomes, including osteoporosis, arthritis, cardiovascular disease, and cancer. Vitamin D-binding protein (DBP) is the primary carrier of vitamin D in the circulation and regulates the bioavailability of 25-hydroxyvitamin D. Epidemiologic studies have shown direct DBP-risk relations and modification by DBP of vitamin D-disease associations.

OBJECTIVE

We aimed to characterize common genetic variants that influence the DBP biochemical phenotype.

DESIGN

We conducted a genome-wide association study (GWAS) of 1380 men through linear regression of single-nucleotide polymorphisms (SNPs) in the Illumina HumanHap500/550/610 array on fasting serum DBP, assuming an additive genetic model, with adjustment for age at blood collection.

RESULTS

We identified 2 independent SNPs located in the gene encoding DBP, GC, that were highly associated with serum DBP: rs7041 (P = 1.42 × 10⁻²⁴⁶) and rs705117 (P = 4.7 × 10⁻⁹¹). For both SNPs, mean serum DBP decreased with increasing copies of the minor allele: mean DBP concentrations (nmol/L) were 7335, 5149, and 3152 for 0, 1, and 2 copies of rs7041 (T), respectively, and 6339, 4280, and 2341, respectively, for rs705117 (G). DBP was also associated with rs12144344 (P = 5.9 × 10⁻⁷) in ST6GALNAC3.

CONCLUSIONS

In this GWAS analysis, to our knowledge the first to examine this biochemical phenotype, 2 variants in GC--one exonic and one intronic--were associated with serum DBP concentrations at the genome-wide level of significance. Understanding the genetic contributions to circulating DBP may provide greater insights into the vitamin D binding, transport, and other functions of DBP and the effect of vitamin D status on health outcomes.

Kernel-machine testing coupled with a rank-truncation method for genetic pathway analysis

Traditional genome-wide association studies (GWASs) usually focus on single-marker analysis, which only accesses marginal effects. Pathway analysis, on the other hand, considers biological pathway gene marker hierarchical structure and therefore provides additional insights into the genetic architecture underlining complex diseases. Recently, a number of methods for pathway analysis have been proposed to assess the significance of a biological pathway from a collection of single-nucleotide polymorphisms. In this study, we propose a novel approach for pathway analysis that assesses the effects of genes using the sequence kernel association test and the effects of pathways using an extended adaptive rank truncated product statistic. It has been increasingly recognized that complex diseases are caused by both common and rare variants. We propose a new weighting scheme for genetic variants across the whole allelic frequency spectrum to be analyzed together without any form of frequency cutoff for defining rare variants. The proposed approach is flexible. It is applicable to both binary and continuous traits, and incorporating covariates is easy. Furthermore, it can be readily applied to GWAS data, exome-sequencing data, and deep resequencing data. We evaluate the new approach on data simulated under comprehensive scenarios and show that it has the highest power in most of the scenarios while maintaining the correct type I error rate. We also apply our proposed methodology to data from a study of the association between bipolar disorder and candidate pathways from Wellcome Trust Case Control Consortium (WTCCC) to show its utility.

Association testing of clustered rare causal variants in case-control studies

Biological evidence suggests that multiple causal variants in a gene may cluster physically. Variants within the same protein functional domain or gene regulatory element would locate in close proximity on the DNA sequence. However, spatial information of variants is usually not used in current rare variant association analyses. We here propose a clustering method (abbreviated as "CLUSTER"), which is extended from the adaptive combination of P-values. Our method combines the association signals of variants that are more likely to be causal. Furthermore, the statistic incorporates the spatial information of variants. With extensive simulations, we show that our method outperforms several commonly-used methods in many scenarios. To demonstrate its use in real data analyses, we also apply this CLUSTER test to the Dallas Heart Study data. CLUSTER is among the best methods when the effects of causal variants are all in the same direction. As variants located in close proximity are more likely to have similar impact on disease risk, CLUSTER is recommended for association testing of clustered rare causal variants in case-control studies.

Association study between gene polymorphisms in PPAR signaling pathway and porcine meat quality traits

There is increasing evidence suggesting that fatty acids biosynthesis and metabolism are regulated by peroxisome proliferator-activated receptors (PPARs), mostly through the PPAR signaling pathway at the transcriptomic level. We hypothesized that the genetic variants of the enzymes in the PPAR signaling pathway may be associated with the traits of porcine meat quality (PMQ). We mined 77 potentially functional single nucleotide polymorphisms in the PPAR signaling pathway of the pig. There were 13 TagSNPs in 13 different genes mapped within the reported pig quantitative trait loci (QTLs) regions related to PMQ based on the Pig QTL database. Based on the association study with ten measured PMQ traits in both the pathway level and the SNP level, we tested eight significantly associated traits with additive effect in the PPAR signaling pathway and explored only one significant TagSNP in gene RXRB, which is directly associated with the trait of skin weight. Moreover, several interactions of TagSNPs were also significantly related to some of PMQ traits. In this large and comprehensive candidate gene set study, we found a modest association of genes and SNPs in the PPAR signaling pathway with PMQ. Further investigation of these gene polymorphisms jointly with fatty acid measures and other genetic factors would help us better understand the regulation mechanisms of PMQ.

Genetic variants in interleukin genes are associated with breast cancer risk and survival in a genetically admixed population: the Breast Cancer Health Disparities Study

Interleukins (ILs) are key regulators of immune response. Genetic variation in IL genes may influence breast cancer risk and mortality given their role in cell growth, angiogenesis and regulation of inflammatory process. We examined 16 IL genes with breast cancer risk and mortality in an admixed population of Hispanic/Native American (NA) (2111 cases and 2597 controls) and non-Hispanic white (NHW) (1481 cases and 1585 controls) women. Adaptive Rank Truncated Product (ARTP) analysis was conducted to determine gene significance and lasso (least absolute shrinkage and selection operator) was used to identify potential gene by gene and gene by lifestyle interactions. The pathway was statistically significant for breast cancer risk overall (P ARTP = 0.0006), for women with low NA ancestry (P(ARTP) = 0.01), for premenopausal women (P(ARTP) = 0.02), for estrogen receptor (ER)+/progesterone receptor (PR)+ tumors (P(ARTP) = 0.03) and ER-/PR- tumors (P(ARTP) = 0.02). Eight of the 16 genes evaluated were associated with breast cancer risk (IL1A, IL1B, IL1RN, IL2, IL2RA, IL4, IL6 and IL10); four genes were associated with breast cancer risk among women with low NA ancestry (IL1B, IL6, IL6R and IL10), two were associated with breast cancer risk among women with high NA ancestry (IL2 and IL2RA) and four genes were associated with premenopausal breast cancer risk (IL1A, IL1B, IL2 and IL3). IL4, IL6R, IL8 and IL17A were associated with breast cancer-specific mortality. We confirmed associations with several functional polymorphisms previously associated with breast cancer risk and provide support that their combined effect influences the carcinogenic process.

Supervised categorical principal component analysis for genome-wide association analyses

In order to have a better understanding of unexplained heritability for complex diseases in conventional Genome-Wide Association Studies (GWAS), aggregated association analyses based on predefined functional regions, such as genes and pathways, become popular recently as they enable evaluating joint effect of multiple Single-Nucleotide Polymorphisms (SNPs), which helps increase the detection power, especially when investigating genetic variants with weak individual effects. In this paper, we focus on aggregated analysis methods based on the idea of Principal Component Analysis (PCA). The past approaches using PCA mostly make some inherent genotype data and/or risk effect model assumptions, which may hinder the accurate detection of potential disease SNPs that influence disease phenotypes. In this paper, we derive a general Supervised Categorical Principal Component Analysis (SCPCA), which explicitly models categorical SNP data without imposing any risk effect model assumption. We have evaluated the efficacy of SCPCA with the comparison to a traditional Supervised PCA (SPCA) and a previously developed Supervised Logistic Principal Component Analysis (SLPCA) based on both the simulated genotype data by HAPGEN2 and the genotype data of Crohn's Disease (CD) from Wellcome Trust Case Control Consortium (WTCCC). Our preliminary results have demonstrated the superiority of SCPCA over both SPCA and SLPCA due to its modeling explicitly designed for categorical SNP data as well as its flexibility on the risk effect model assumption.

Rare variant association testing by adaptive combination of P-values

With the development of next-generation sequencing technology, there is a great demand for powerful statistical methods to detect rare variants (minor allele frequencies (MAFs)<1%) associated with diseases. Testing for each variant site individually is known to be underpowered, and therefore many methods have been proposed to test for the association of a group of variants with phenotypes, by pooling signals of the variants in a chromosomal region. However, this pooling strategy inevitably leads to the inclusion of a large proportion of neutral variants, which may compromise the power of association tests. To address this issue, we extend the -MidP method (Cheung et al., 2012, Genet Epidemiol 36: 675–685) and propose an approach (named ‘adaptive combination of P-values for rare variant association testing’, abbreviated as ‘ADA’) that adaptively combines per-site P-values with the weights based on MAFs. Before combining P-values, we first imposed a truncation threshold upon the per-site P-values, to guard against the noise caused by the inclusion of neutral variants. This ADA method is shown to outperform popular burden tests and non-burden tests under many scenarios. ADA is recommended for next-generation sequencing data analysis where many neutral variants may be included in a functional region.

Comprehensive analyses of DNA repair pathways, smoking and bladder cancer risk in Los Angeles and Shanghai

Tobacco smoking is a bladder cancer risk factor and a source of carcinogens that induce DNA damage to urothelial cells. Using data and samples from 988 cases and 1,004 controls enrolled in the Los Angeles County Bladder Cancer Study and the Shanghai Bladder Cancer Study, we investigated associations between bladder cancer risk and 632 tagSNPs that comprehensively capture genetic variation in 28 DNA repair genes from four DNA repair pathways: base excision repair (BER), nucleotide excision repair (NER), non-homologous end-joining (NHEJ) and homologous recombination repair (HHR). Odds ratios (ORs) and 95% confidence intervals (CIs) for each tagSNP were corrected for multiple testing for all SNPs within each gene using pACT and for genes within each pathway and across pathways with Bonferroni. Gene and pathway summary estimates were obtained using ARTP. We observed an association between bladder cancer and POLB rs7832529 (BER) (pACT = 0.003; ppathway = 0.021) among all, and SNPs in XPC (NER) and OGG1 (BER) among Chinese men and women, respectively. The NER pathway showed an overall association with risk among Chinese males (ARTP NER p = 0.034). The XRCC6 SNP rs2284082 (NHEJ), also in LD with SREBF2, showed an interaction with smoking (smoking status interaction pgene = 0.001, ppathway = 0.008, poverall = 0.034). Our findings support a role in bladder carcinogenesis for regions that map close to or within BER (POLB, OGG1) and NER genes (XPC). A SNP that tags both the XRCC6 and SREBF2 genes strongly modifies the association between bladder cancer risk and smoking.

Incorporating prior knowledge to increase the power of genome-wide association studies

Typical methods of analyzing genome-wide single nucleotide variant (SNV) data in cases and controls involve testing each variant's genotypes separately for phenotype association, and then using a substantial multiple-testing penalty to minimize the rate of false positives. This approach, however, can result in low power for modestly associated SNVs. Furthermore, simply looking at the most associated SNVs may not directly yield biological insights about disease etiology. SNVset methods attempt to address both limitations of the traditional approach by testing biologically meaningful sets of SNVs (e.g., genes or pathways). The number of tests run in a SNVset analysis is typically much lower (hundreds or thousands instead of millions) than in a traditional analysis, so the false-positive rate is lower. Additionally, by testing SNVsets that are biologically meaningful finding a significant set may more quickly yield insights into disease etiology.In this chapter we summarize the short history of SNVset testing and provide an overview of the many recently proposed methods. Furthermore, we provide detailed step-by-step instructions on how to perform a SNVset analysis, including a substantial number of practical tips and questions that researchers should consider before undertaking a SNVset analysis. Lastly, we describe a companion R package (snvset) that implements recently proposed SNVset methods. While SNVset testing is a new approach, with many new methods still being developed and many open questions, the promise of the approach is worth serious consideration when considering analytic methods for GWAS.

The influence of genetic ancestry and ethnicity on breast cancer survival associated with genetic variation in the TGF-β-signaling pathway: The Breast Cancer Health Disparities Study

The TGF-β signaling pathway regulates cellular proliferation and differentiation. We evaluated genetic variation in this pathway, its association with breast cancer survival, and survival differences by genetic ancestry and self-reported ethnicity. The Breast Cancer Health Disparities Study includes participants from the 4-Corners Breast Cancer Study (n = 1,391 cases) and the San Francisco Bay Area Breast Cancer Study (n = 946 cases) who have been followed for survival. We evaluated 28 genes in the TGF-β signaling pathway using a tagSNP approach. Adaptive rank truncated product (ARTP) was used to test the gene and pathway significance by Native American (NA) ancestry and by self-reported ethnicity (non-Hispanic white (NHW) and Hispanic/NA). Genetic variation in the TGF-β signaling pathway was associated with overall breast cancer survival (P ARTP = 0.05), especially for women with low NA ancestry (P ARTP = 0.007) and NHW women (P ARTP = 0.006). BMP2, BMP4, RUNX1, and TGFBR3 were significantly associated with breast cancer survival overall (P ARTP = 0.04, 0.02, 0.002, and 0.04, respectively). Among women with low NA, ancestry associations were as follows: BMP4 (P ARTP = 0.007), BMP6 (P ARTP = 0.001), GDF10 (P ARTP = 0.05), RUNX1 (P ARTP = 0.002), SMAD1 (P ARTP = 0.05), and TGFBR2 (P ARTP = 0.02). A polygenic risk model showed that women with low NA ancestry and high numbers of at-risk alleles had twice the risk of dying from breast cancer as did women with high NA ancestry. Our data suggest that genetic variation in the TGF-β signaling pathway influences breast cancer survival. Associations were similar when the analyses were stratified by genetic ancestry or by self-reported ethnicity.

GEE-based SNP set association test for continuous and discrete traits in family-based association studies

Family-based genetic association studies of related individuals provide opportunities to detect genetic variants that complement studies of unrelated individuals. Most statistical methods for family association studies for common variants are single marker based, which test one SNP a time. In this paper, we consider testing the effect of an SNP set, e.g., SNPs in a gene, in family studies, for both continuous and discrete traits. Specifically, we propose a generalized estimating equations (GEEs) based kernel association test, a variance component based testing method, to test for the association between a phenotype and multiple variants in an SNP set jointly using family samples. The proposed approach allows for both continuous and discrete traits, where the correlation among family members is taken into account through the use of an empirical covariance estimator. We derive the theoretical distribution of the proposed statistic under the null and develop analytical methods to calculate the P-values. We also propose an efficient resampling method for correcting for small sample size bias in family studies. The proposed method allows for easily incorporating covariates and SNP-SNP interactions. Simulation studies show that the proposed method properly controls for type I error rates under both random and ascertained sampling schemes in family studies. We demonstrate through simulation studies that our approach has superior performance for association mapping compared to the single marker based minimum P-value GEE test for an SNP-set effect over a range of scenarios. We illustrate the application of the proposed method using data from the Cleveland Family GWAS Study.

SEPP1 influences breast cancer risk among women with greater native american ancestry: the breast cancer health disparities study

Selenoproteins are a class of proteins containing a selenocysteine residue, many of which have been shown to have redox functions, acting as antioxidants to decrease oxidative stress. Selenoproteins have previously been associated with risk of various cancers and redox-related diseases. In this study we evaluated possible associations between breast cancer risk and survival and single nucleotide polymorphisms (SNPs) in the selenoprotein genes GPX1, GPX2, GPX3, GPX4, SELS, SEP15, SEPN1, SEPP1, SEPW1, TXNRD1, and TXNRD2 among Hispanic/Native American (2111 cases, 2597 controls) and non-Hispanic white (NHW) (1481 cases, 1586 controls) women in the Breast Cancer Health Disparities Study. Adaptive Rank Truncated Product (ARTP) analysis was used to determine both gene and pathway significance with these genes. The overall selenoprotein pathway P_ARTP was not significantly associated with breast cancer risk (P_ARTP = 0.69), and only one gene, GPX3, was of borderline significance for the overall population (P_ARTP =0.09) and marginally significant among women with 0-28% Native American (NA) ancestry (P_ARTP=0.06). The SEPP1 gene was statistically significantly associated with breast cancer risk among women with higher NA ancestry (P_ARTP=0.002) and contributed to a significant pathway among those women (P_ARTP=0.04). GPX1, GPX3, and SELS were associated with Estrogen Receptor-/Progesterone Receptor+ status (P_ARTP = 0.002, 0.05, and 0.01, respectively). Four SNPs (GPX3 rs2070593, rsGPX4 rs2074451, SELS rs9874, and TXNRD1 rs17202060) significantly interacted with dietary oxidative balance score after adjustment for multiple comparisons to alter breast cancer risk. GPX4 was significantly associated with breast cancer survival among those with the highest NA ancestry (P_ARTP = 0.05) only. Our data suggest that SEPP1 alters breast cancer risk among women with higher levels of NA ancestry.

RTK/ERK pathway under natural selection associated with prostate cancer

Prostate cancer (PCa) is a global disease causing large numbers of deaths every year. Recent studies have indicated the RTK/ERK pathway might be a key pathway in the development of PCa. However, the exact association and evolution-based mechanism remain unclear. This study was conducted by combining genotypic and phenotypic data from the Chinese Consortium for Prostate Cancer Genetics (ChinaPCa) with related databases such as the HapMap Project and Genevar. In this analysis, expression of quantitative trait loci (eQTLs) analysis, natural selection and gene-based pathway analysis were involved. The pathway analysis confirmed the positive relationship between PCa risk and several key genes. In addition, combined with the natural selection, it seems that 4 genes (EGFR, ERBB2, PTK2, and RAF1) with five SNPs (rs11238349, rs17172438, rs984654, rs11773818, and rs17172432) especially rs17172432, might be pivotal factors in the development of PCa. The results indicate that the RTK/ERK pathway under natural selection is a key link in PCa risk. The joint effect of the genes and loci with positive selection might be one reason for the development of PCa. Dealing with all the factors simultaneously might give insight into prevention and aid in predicting the success of potential therapies for PCa.

Genes-environment interactions in obesity- and diabetes-associated pancreatic cancer: a GWAS data analysis

BACKGROUND

Obesity and diabetes are potentially alterable risk factors for pancreatic cancer. Genetic factors that modify the associations of obesity and diabetes with pancreatic cancer have previously not been examined at the genome-wide level.

METHODS

Using genome-wide association studies (GWAS) genotype and risk factor data from the Pancreatic Cancer Case Control Consortium, we conducted a discovery study of 2,028 cases and 2,109 controls to examine gene-obesity and gene-diabetes interactions in relation to pancreatic cancer risk by using the likelihood-ratio test nested in logistic regression models and Ingenuity Pathway Analysis (IPA).

RESULTS

After adjusting for multiple comparisons, a significant interaction of the chemokine signaling pathway with obesity (P = 3.29 × 10(-6)) and a near significant interaction of calcium signaling pathway with diabetes (P = 1.57 × 10(-4)) in modifying the risk of pancreatic cancer were observed. These findings were supported by results from IPA analysis of the top genes with nominal interactions. The major contributing genes to the two top pathways include GNGT2, RELA, TIAM1, and GNAS. None of the individual genes or single-nucleotide polymorphism (SNP) except one SNP remained significant after adjusting for multiple testing. Notably, SNP rs10818684 of the PTGS1 gene showed an interaction with diabetes (P = 7.91 × 10(-7)) at a false discovery rate of 6%.

CONCLUSIONS

Genetic variations in inflammatory response and insulin resistance may affect the risk of obesity- and diabetes-related pancreatic cancer. These observations should be replicated in additional large datasets.

IMPACT

A gene-environment interaction analysis may provide new insights into the genetic susceptibility and molecular mechanisms of obesity- and diabetes-related pancreatic cancer.

Two novel pathway analysis methods based on a hierarchical model

Motivation: Over the past few years several pathway analysis methods have been proposed for exploring and enhancing the analysis of genome-wide association data. Hierarchical models have been advocated as a way to integrate SNP and pathway effects in the same model, but their computational complexity has prevented them being applied on a genome-wide scale to date.

Methods: We present two novel methods for identifying associated pathways. In the proposed hierarchical model, the SNP effects are analytically integrated out of the analysis, allowing computationally tractable model fitting to genome-wide data. The first method uses Bayes factors for calculating the effect of the pathways, whereas the second method uses a machine learning algorithm and adaptive lasso for finding a sparse solution of associated pathways.

Results: The performance of the proposed methods was explored on both simulated and real data. The results of the simulation study showed that the methods outperformed some well-established association methods: the commonly used Fisher’s method for combining P-values and also the recently published BGSA. The methods were applied to two genome-wide association study datasets that aimed to find the genetic structure of platelet function and body mass index, respectively. The results of the analyses replicated the results of previously published pathway analysis of these phenotypes but also identified novel pathways that are potentially involved.

Availability: An R package is under preparation. In the meantime, the scripts of the methods are available on request from the authors.

Contact: marina.evangelou@cimr.cam.ac.uk

Supplementary Information: Supplementary data are available at Bioinformatics online.

Integrated enrichment analysis of variants and pathways in genome-wide association studies indicates central role for IL-2 signaling genes in type 1 diabetes, and cytokine signaling genes in Crohn's disease

Pathway analyses of genome-wide association studies aggregate information over sets of related genes, such as genes in common pathways, to identify gene sets that are enriched for variants associated with disease. We develop a model-based approach to pathway analysis, and apply this approach to data from the Wellcome Trust Case Control Consortium (WTCCC) studies. Our method offers several benefits over existing approaches. First, our method not only interrogates pathways for enrichment of disease associations, but also estimates the level of enrichment, which yields a coherent way to promote variants in enriched pathways, enhancing discovery of genes underlying disease. Second, our approach allows for multiple enriched pathways, a feature that leads to novel findings in two diseases where the major histocompatibility complex (MHC) is a major determinant of disease susceptibility. Third, by modeling disease as the combined effect of multiple markers, our method automatically accounts for linkage disequilibrium among variants. Interrogation of pathways from eight pathway databases yields strong support for enriched pathways, indicating links between Crohn's disease (CD) and cytokine-driven networks that modulate immune responses; between rheumatoid arthritis (RA) and "Measles" pathway genes involved in immune responses triggered by measles infection; and between type 1 diabetes (T1D) and IL2-mediated signaling genes. Prioritizing variants in these enriched pathways yields many additional putative disease associations compared to analyses without enrichment. For CD and RA, 7 of 8 additional non-MHC associations are corroborated by other studies, providing validation for our approach. For T1D, prioritization of IL-2 signaling genes yields strong evidence for 7 additional non-MHC candidate disease loci, as well as suggestive evidence for several more. Of the 7 strongest associations, 4 are validated by other studies, and 3 (near IL-2 signaling genes RAF1, MAPK14, and FYN) constitute novel putative T1D loci for further study.

A fast multilocus test with adaptive SNP selection for large-scale genetic-association studies

As increasing evidence suggests that multiple correlated genetic variants could jointly influence the outcome, a multilocus test that aggregates association evidence across multiple genetic markers in a considered gene or a genomic region may be more powerful than a single-marker test for detecting susceptibility loci. We propose a multilocus test, AdaJoint, which adopts a variable selection procedure to identify a subset of genetic markers that jointly show the strongest association signal, and defines the test statistic based on the selected genetic markers. The P-value from the AdaJoint test is evaluated by a computationally efficient algorithm that effectively adjusts for multiple-comparison, and is hundreds of times faster than the standard permutation method. Simulation studies demonstrate that AdaJoint has the most robust performance among several commonly used multilocus tests. We perform multilocus analysis of over 26,000 genes/regions on two genome-wide association studies of pancreatic cancer. Compared with its competitors, AdaJoint identifies a much stronger association between the gene CLPTM1L and pancreatic cancer risk (6.0 × 10(-8)), with the signal optimally captured by two correlated single-nucleotide polymorphisms (SNPs). Finally, we show AdaJoint as a powerful tool for mapping cis-regulating methylation quantitative trait loci on normal breast tissues, and find many CpG sites whose methylation levels are jointly regulated by multiple SNPs nearby.

Genetic variants in fas signaling pathway genes and risk of gastric cancer

Populations in north central China are at high risk for gastric cancers (GC), and altered FAS-mediated cell signaling and/or apoptosis may contribute to this risk. We examined the association of 554 single nucleotide polymorphisms (SNPs) in 53 Fas signaling-related genes using a pathway-based approach in 1758 GC cases (1126 gastric cardia adenocarcinomas (GCA) and 632 gastric noncardia adenocarcinomas (GNCA)), and 2111 controls from a genome-wide association study (GWAS) of GC in ethnic Chinese. SNP associations with risk of overall GC, GCA and GNCA were evaluated using unconditional logistic regressions controlling for age, sex and study. Gene- and pathway-based associations were tested using the adaptive rank-truncated product (ARTP) method. Statistical significance was evaluated empirically by permutation. Significant pathway-based associations were observed for Fas signaling with risk of overall GC (p = 5.5E-04) and GCA (p = 6.3E-03), but not GNCA (p= 8.1E-02). Among examined genes in the Fas signaling pathway, MAP2K4, FAF1, MAPK8, CASP10, CASP8, CFLAR, MAP2K1, CAP8AP2, PAK2 and IKBKB were associated with risk of GC (nominal p < 0.05), and FAF1 and MAPK8 were significantly associated with risk of both GCA and GNCA (nominal p< 0.05). Our examination of genetic variation in the Fas signaling pathway is consistent with an association of altered Fas signaling and/or apoptosis with risk of GC. As one of the first attempts to investigate a pathway-level association, our results suggest that these genes and the Fas signaling pathway warrant further evaluation in relation to GC risk in other populations.

A powerful method for combining P-values in genomic studies

After genetic regions have been identified in genomewide association studies (GWAS), investigators often follow up with more targeted investigations of specific regions. These investigations typically are based on single nucleotide polymorphisms (SNPs) with dense coverage of a region. Methods are thus needed to test the hypothesis of any association in given genetic regions. Several approaches for combining P-values obtained from testing individual SNP hypothesis tests are available. We recently proposed a sequential procedure for testing the global null hypothesis of no association in a region. When this global null hypothesis is rejected, this method provides a list of significant hypotheses and has weak control of the family-wise error rate. In this paper, we devise a permutation-based version of the test that accounts for correlations of tests based on SNPs in the same genetic region. Based on simulated data, the method has correct control of the type I error rate and higher or comparable power to other tests.

Angiogenesis genes, dietary oxidative balance and breast cancer risk and progression: the Breast Cancer Health Disparities Study

Angiogenesis is essential for tumor development and progression. Genetic variation in angiogenesis-related genes may influence breast carcinogenesis. We evaluated dietary factors associated with oxidative balance, DDIT4 (one SNP), FLT1 (35 SNPs), HIF1A (four SNPs), KDR (19 SNPs), MPO (one SNP), NOS2A (15 SNPs), TEK (40 SNPs) and VEGFA (eight SNPs) and breast cancer risk among Hispanic (2,111 cases and 2,597 controls) and non-Hispanic white (1,481 cases and 1,586 controls) women in the Breast Cancer Health Disparities Study. Adaptive rank truncated product (ARTP) analysis was used to determine gene and pathway significance with breast cancer. TEK was associated with breast cancer overall (pARTP = 0.03) and with breast cancer survival (pARTP = 0.01). KDR was of borderline significance overall (pARTP = 0.07), although significantly associated with breast cancer in both low and intermediate Native American (NA) ancestry groups (pARTP = 0.02) and estrogen receptor (ER)+/progesterone receptor (PR)- tumor phenotype (pARTP = 0.008). Both VEGFA and NOS2A were associated with ER-/PR- tumor phenotype (pARTP = 0.01 and pARTP = 0.04, respectively). FLT1 was associated with breast cancer survival among those with low NA ancestry (pARTP = 0.009). With respect to diet, having a higher dietary oxidative balance score (DOBS) was significantly associated with lower breast cancer risk [odds ratio (OR) 0.74, 95% confidence interval (CI) 0.64-0.84], with the strongest associations observed for women with the highest NA ancestry (OR 0.44, 95% CI 0.30-0.65). We observed few interactions between DOBS and angiogenesis-related genes. Our data suggest that dietary factors and genetic variation in angiogenesis-related genes contribute to breast cancer carcinogenesis.

Genetic variation in the TLR and NF-κB pathways and cervical and vulvar cancer risk: a population-based case-control study

Genital infection with the oncogenic human papillomavirus is the necessary cause of cervical cancer and of a large fraction of vulvar cancers. The toll-like receptor and the nuclear factor κB (NF-κB) signaling pathways have been implicated in inflammation, autoimmune disease and cancer, but whether common nucleotide variation in these pathways is associated with the risk of cervical and vulvar cancers has received little study. Using data from a population-based case-control study of cervical and vulvar cancers, we genotyped 205 single nucleotide polymorphisms (SNPs) in and around 32 candidate gene regions within these pathways. Gene-based analyses were used to estimate the associations between individual gene regions and the risk of cervical and vulvar cancers. Odds ratio (OR) and 95% confidence intervals (CI) were calculated to assess the risk of cervical and vulvar cancers for each SNP. p-Values were adjusted for multiple testing. A total of 876 cervical cancer cases, 517 vulvar cancer cases and 1,100 controls were included in the analysis. The TNF region was significantly associated with the risks of cervical cancer (gene-based p-value: 2.0 × 10(-4) ) and vulvar cancer (gene-based p-value: 1.0 × 10(-4) ). The rare allele (A) of SNP rs2239704 in the 5' UTR of the LTA gene was significantly associated with increased risks of cervical cancer (OR=1.31, 95% CI: 1.15-1.50; adjusted p-value: 0.013) and vulvar cancer (OR=1.51, 95% CI: 1.30-1.75; adjusted p-value: 1.9 × 10(-5) ). These findings add to the evidence of the importance of the immune system in the etiology of cervical and vulvar cancers.

Associations with growth factor genes (FGF1, FGF2, PDGFB, FGFR2, NRG2, EGF, ERBB2) with breast cancer risk and survival: the Breast Cancer Health Disparities Study

Growth factors (GF) stimulate cell proliferation through binding to cell membrane receptors and are thought to be involved in cancer risk and survival. We examined how genetic variation in epidermal growth factor (EGF), neuregulin 2 (NRG2), ERBB2 (HER2/neu), fibroblast growth factors 1 and 2 (FGF1 and FGF2) and its receptor 2 (FGFR2), and platelet-derived growth factor B (PDGFB) independently and collectively influence breast cancer risk and survival. We analyzed data from the Breast Cancer Health Disparities Study which includes Hispanic (2,111 cases, 2,597 controls) and non-Hispanic white (1,481 cases, 1,586 controls) women. Adaptive rank-truncated product (ARTP) analysis was conducted to determine gene significance. Odds ratios (OR) and 95 % confidence intervals were obtained from conditional logistic regression models to estimate breast cancer risk and Cox proportional hazard models were used to estimate hazard ratios (HR) of dying from breast cancer. We assessed Native American (NA) ancestry using 104 ancestry informative markers. We observed few significant associations with breast cancer risk overall or by menopausal status other than for FGFR2 rs2981582. This SNP was significantly associated with ER+/PR+ (OR 1.66, 95 % CI 1.37-2.00) and ER+/PR- (OR 1.54, 95 % CI 1.03-2.31) tumors. Multiple SNPs in FGF1, FGF2, and NRG2 significantly interacted with multiple SNPs in EGFR, ERBB2, FGFR2, and PDGFB, suggesting that breast cancer risk is dependent on the collective effects of genetic variants in other GFs. Both FGF1 and ERBB2 significantly influenced overall survival, especially among women with low levels of NA ancestry (P ARTP = 0.007 and 0.003, respectively). Our findings suggest that genetic variants in growth factors signaling appear to influence breast cancer risk through their combined effects. Genetic variation in ERBB2 and FGF1 appear to be associated with survival after diagnosis with breast cancer.

Genetic variants in epidermal growth factor receptor pathway genes and risk of esophageal squamous cell carcinoma and gastric cancer in a Chinese population

The epidermal growth factor receptor (EGFR) signaling pathway regulates cell proliferation, differentiation, and survival, and is frequently dysregulated in esophageal and gastric cancers. Few studies have comprehensively examined the association between germline genetic variants in the EGFR pathway and risk of esophageal and gastric cancers. Based on a genome-wide association study in a Han Chinese population, we examined 3443 SNPs in 127 genes in the EGFR pathway for 1942 esophageal squamous cell carcinomas (ESCCs), 1758 gastric cancers (GCs), and 2111 controls. SNP-level analyses were conducted using logistic regression models. We applied the resampling-based adaptive rank truncated product approach to determine the gene- and pathway-level associations. The EGFR pathway was significantly associated with GC risk (P = 2.16×10(-3)). Gene-level analyses found 10 genes to be associated with GC, including FYN, MAPK8, MAP2K4, GNAI3, MAP2K1, TLN1, PRLR, PLCG2, RPS6KB2, and PIK3R3 (P<0.05). For ESCC, we did not observe a significant pathway-level association (P = 0.72), but gene-level analyses suggested associations between GNAI3, CHRNE, PAK4, WASL, and ITCH, and ESCC (P<0.05). Our data suggest an association between specific genes in the EGFR signaling pathway and risk of GC and ESCC. Further studies are warranted to validate these associations and to investigate underlying mechanisms.

Facial pain with localized and widespread manifestations: separate pathways of vulnerability

Human association studies of common genetic polymorphisms have identified many loci that are associated with risk of complex diseases, although individual loci typically have small effects. However, by envisaging genetic associations in terms of cellular pathways, rather than any specific polymorphism, combined effects of many biologically relevant alleles can be detected. The effects are likely to be most apparent in investigations of phenotypically homogenous subtypes of complex diseases. We report findings from a case-control, genetic association study of relationships between 2925 single nucleotide polymorphisms (SNPs) and 2 subtypes of a commonly occurring chronic facial pain condition, temporomandibular disorder (TMD): 1) localized TMD and 2) TMD with widespread pain. When compared to healthy controls, cases with localized TMD differed in allelic frequency of SNPs that mapped to a serotonergic receptor pathway (P=0.0012), while cases of TMD with widespread pain differed in allelic frequency of SNPs that mapped to a T-cell receptor pathway (P=0.0014). A risk index representing combined effects of 6 SNPs from the serotonergic pathway was associated with greater odds of localized TMD (odds ratio 2.7, P=1.3 E-09), and the result was reproduced in a replication case-control cohort study of 639 people (odds ratio 1.6, P=0.014). A risk index representing combined effects of 8 SNPs from the T-cell receptor pathway was associated with greater odds of TMD with widespread pain (P=1.9 E-08), although the result was not significant in the replication cohort. These findings illustrate potential for clinical classification of chronic pain based on distinct molecular profiles and genetic background.

VEGFA, FLT1, KDR and colorectal cancer: assessment of disease risk, tumor molecular phenotype, and survival

Angiogenesis is essential for tumor progression. Vascular endothelial growth factor (VEGFA) and its receptors 1 (FLT1) and 2 (KDR), have been identified as major mediators of this process. We hypothesized that genetic variation in FLT1 (38 SNPs), KDR (22 SNPS), and VEGFA (11 SNPs) would be associated with colon and rectal cancer development and survival. Data from a case-control study of 1555 colon cancer cases and 1956 controls and 754 rectal cancer cases and 959 controls were used. An adaptive rank truncation product (ARTP), based on 10,000 permutations, was used to determine the statistical significance of the candidate genes and angiogenesis pathway. Based on ARTP results, FLT1 was significantly associated with risk of colon cancer (P(ARTP) = 0.045) and VEGFA was significantly associated with rectal cancer (P(ARTP) = 0.036). After stratifying by tumor molecular subtype, SNP associations observed for colon cancer were: VEGFA rs2010963 with CIMP+ colon tumors; FLT1 rs4771249 and rs7987649 with TP53; FLT1 rs3751397, rs7337610, rs7987649, and rs9513008 and KDR rs10020464, rs11941492, and rs12498529 with MSI+ and CIMP+/KRAS2-mutated tumors. FLT1 rs2296189 and rs600640 were associated with CIMP+ rectal tumors and FLT1 rs7983774 was associated with TP53-mutated rectal tumors. Four SNPs in FLT1 were associated with colon cancer survival while three SNPs in KDR were associated with survival after diagnosis with rectal cancer. Aspirin/NSAID use, smoking cigarettes, and BMI modified the associations. These findings suggest the importance of inflammation and angiogenesis in the etiology of colorectal cancer and that genetic and lifestyle factors may be targets for modulating disease risk.

Pathway-based analysis using genome-wide association data from a Korean non-small cell lung cancer study

Pathway-based analysis, used in conjunction with genome-wide association study (GWAS) techniques, is a powerful tool to detect subtle but systematic patterns in genome that can help elucidate complex diseases, like cancers. Here, we stepped back from genetic polymorphisms at a single locus and examined how multiple association signals can be orchestrated to find pathways related to lung cancer susceptibility. We used single-nucleotide polymorphism (SNP) array data from 869 non-small cell lung cancer (NSCLC) cases from a previous GWAS at the National Cancer Center and 1,533 controls from the Korean Association Resource project for the pathway-based analysis. After mapping single-nucleotide polymorphisms to genes, considering their coding region and regulatory elements (±20 kbp), multivariate logistic regression of additive and dominant genetic models were fitted against disease status, with adjustments for age, gender, and smoking status. Pathway statistics were evaluated using Gene Set Enrichment Analysis (GSEA) and Adaptive Rank Truncated Product (ARTP) methods. Among 880 pathways, 11 showed relatively significant statistics compared to our positive controls (P_GSEA≤0.025, false discovery rate≤0.25). Candidate pathways were validated using the ARTP method and similarities between pathways were computed against each other. The top-ranked pathways were ABC Transporters (P_GSEA<0.001, P_ARTP = 0.001), VEGF Signaling Pathway (P_GSEA<0.001, P_ARTP = 0.008), G1/S Check Point (P_GSEA = 0.004, P_ARTP = 0.013), and NRAGE Signals Death through JNK (P_GSEA = 0.006, P_ARTP = 0.001). Our results demonstrate that pathway analysis can shed light on post-GWAS research and help identify potential targets for cancer susceptibility.

Matrix metalloproteinase genes are associated with breast cancer risk and survival: the Breast Cancer Health Disparities Study

Matrix metalloproteinases (MMPs) contribute to cancer through their involvement in cancer invasion and metastasis. We evaluated genetic variation in MMP1 (9 SNPs), MMP2 (8 SNPs), MMP3 (4 SNPs), and MMP9 (3 SNPs) and breast cancer risk among Hispanic (2111 cases, 2597 controls) and non-Hispanic white (NHW) (1481 cases, 1586 controls) women in the Breast Cancer Health Disparities Study. Ancestral informative markers (n = 104) were assessed to determine Native American (NA) ancestry. MMP1 [4 single nucleotide polymorphisms (SNPs)] and MMP2 (2 SNPs) were associated with breast cancer overall. MMP1 rs996999 had strongest associations among women with the most NA ancestry (OR 1.61,95% CI 1.09,2.40) as did MMP3 rs650108 (OR 1.36, 95% CI 1.05,1.75) and MMP9 rs3787268 (OR 1.52, 95% CI 1.09,2.13). The adaptive rank truncated product (ARTP) showed a significant pathway p_artp value of 0.04, with a stronger association among women with the most NA ancestry (p_artp = 0.02). Significant pathway genes using the ARTP were MMP1 for all women (p_artp = 0.02) and MMP9 for women with the most NA ancestry (p_artp = 0.024); MMP2 was borderline significant overall (p_artp = 0.06) and MMP1 and MMP3 were borderline significant for women with the most NA ancestry (p_artp = 0.07 and 0.06 respectively). MMP1 and MMP2 were associated with ER+/PR+ and ER+/PR-tumors; MMP3 and MMP9 were associated with ER−/PR− tumors. The pathway was highly significant with survival (p_artp = 0.0041) with MMP2 having the strongest gene association (p_artp = 0.0007). Our findings suggest that genetic variation in MMP genes influence breast cancer development and survival in this genetically admixed population.

An adaptive truncated product method for combining dependent p-values

We propose an adaptive truncated product method that facilitates the selection of the truncation point among a set of candidates. To efficiently estimate the distribution of the proposed method when the p-values are correlated, we develop a single-layer bootstrap procedure.

Genetic variants in DNA repair pathway genes and risk of esophageal squamous cell carcinoma and gastric adenocarcinoma in a Chinese population

The DNA repair pathways help to maintain genomic integrity and therefore genetic variation in the pathways could affect the propensity to develop cancer. Selected germline single nucleotide polymorphisms (SNPs) in the pathways have been associated with esophageal cancer and gastric cancer (GC) but few studies have comprehensively examined the pathway genes. We aimed to investigate associations between DNA repair pathway genes and risk of esophageal squamous cell carcinoma (ESCC) and GC, using data from a genome-wide association study in a Han Chinese population where ESCC and GC are the predominant cancers. In sum, 1942 ESCC cases, 1758 GC cases and 2111 controls from the Shanxi Upper Gastrointestinal Cancer Genetics Project (discovery set) and the Linxian Nutrition Intervention Trials (replication set) were genotyped for 1675 SNPs in 170 DNA repair-related genes. Logistic regression models were applied to evaluate SNP-level associations. Gene- and pathway-level associations were determined using the resampling-based adaptive rank-truncated product approach. The DNA repair pathways overall were significantly associated with risk of ESCC (P = 6.37 × 10(-4)), but not with GC (P = 0.20). The most significant gene in ESCC was CHEK2 (P = 2.00 × 10(-6)) and in GC was CLK2 (P = 3.02 × 10(-4)). We observed several other genes significantly associated with either ESCC (SMUG1, TDG, TP53, GTF2H3, FEN1, POLQ, HEL308, RAD54B, MPG, FANCE and BRCA1) or GC risk (MRE11A, RAD54L and POLE) (P < 0.05). We provide evidence for an association between specific genes in the DNA repair pathways and the risk of ESCC and GC. Further studies are warranted to validate these associations and to investigate underlying mechanisms.

Common single nucleotide polymorphisms in genes related to immune function and risk of papillary thyroid cancer

Accumulating evidence suggests that alterations in immune function may be important in the etiology of papillary thyroid cancer (PTC). To identify genetic markers in immune-related pathways, we evaluated 3,985 tag single nucleotide polymorphisms (SNPs) in 230 candidate gene regions (adhesion-extravasation-migration, arachidonic acid metabolism/eicosanoid signaling, complement and coagulation cascade, cytokine signaling, innate pathogen detection and antimicrobials, leukocyte signaling, TNF/NF-kB pathway or other) in a case-control study of 344 PTC cases and 452 controls. We used logistic regression models to estimate odds ratios (OR) and calculate one degree of freedom P values of linear trend (P(SNP-trend) ) for the association between genotype (common homozygous, heterozygous, variant homozygous) and risk of PTC. To correct for multiple comparisons, we applied the false discovery rate method (FDR). Gene region- and pathway-level associations (P(Region) and P(Pathway)) were assessed by combining individual P(SNP-trend) values using the adaptive rank truncated product method. Two SNPs (rs6115, rs6112) in the SERPINA5 gene were significantly associated with risk of PTC (P(SNP-FDR)/P(SNP-trend)= 0.02/6×10(-6) and P(SNP-FDR)/P(SNP-trend)= 0.04/2×10(-5), respectively). These associations were independent of a history of autoimmune thyroiditis (OR = 6.4; 95% confidence interval: 3.0-13.4). At the gene region level, SERPINA5 was suggestively associated with risk of PTC (P(Region-FDR)/P(Region)= 0.07/0.0003). Overall, the complement and coagulation cascade pathway was the most significant pathway (P(Pathway)= 0.02) associated with PTC risk largely due to the strong effect of SERPINA5. Our results require replication but suggest that the SERPINA5 gene, which codes for the protein C inhibitor involved in many biological processes including inflammation, may be a new susceptibility locus for PTC.

Hormone metabolism genes and mammographic density in Singapore Chinese women

BACKGROUND

Female steroid hormone levels and exogenous hormone use influence breast cancer risk. We investigated the association between genetic variation in the hormone metabolism and signaling pathway and mammographic density, a strong predictor of breast cancer risk.

METHODS

We genotyped 161 SNPs in 15 hormone metabolism pathway gene regions and evaluated mammographic density in 2,038 Singapore Chinese women. Linear regression analysis was used to investigate single-nucleotide polymorphism (SNP) and mammographic density association. An overall pathway summary was obtained using the adaptive ranked truncated product test.

RESULTS

We did not find any of the individually tested SNPs to be associated with mammographic density after a multiple testing correction. There was no evidence of an overall effect on mammographic density of genetic variation in the hormone metabolism pathway.

CONCLUSIONS

In this cross-sectional study, genetic variation in hormone metabolism pathway was not associated with mammographic density in Singapore Chinese women.

IMPACT

Consistent with existing data from Caucasian populations, polymorphisms in hormone pathway genes are not likely to be strong predictors of mammographic density in Asian women.

Common genetic variants in the 9p21 region and their associations with multiple tumours

BACKGROUND

The chromosome 9p21.3 region has been implicated in the pathogenesis of multiple cancers.

METHODS

We systematically examined up to 203 tagging SNPs of 22 genes on 9p21.3 (19.9-32.8 Mb) in eight case-control studies: thyroid cancer, endometrial cancer (EC), renal cell carcinoma, colorectal cancer (CRC), colorectal adenoma (CA), oesophageal squamous cell carcinoma (ESCC), gastric cardia adenocarcinoma and osteosarcoma (OS). We used logistic regression to perform single SNP analyses for each study separately, adjusting for study-specific covariates. We combined SNP results across studies by fixed-effect meta-analyses and a newly developed subset-based statistical approach (ASSET). Gene-based P-values were obtained by the minP method using the Adaptive Rank Truncated Product program. We adjusted for multiple comparisons by Bonferroni correction.

RESULTS

Rs3731239 in cyclin-dependent kinase inhibitors 2A (CDKN2A) was significantly associated with ESCC (P=7 × 10(-6)). The CDKN2A-ESCC association was further supported by gene-based analyses (Pgene=0.0001). In the meta-analyses by ASSET, four SNPs (rs3731239 in CDKN2A, rs615552 and rs573687 in CDKN2B and rs564398 in CDKN2BAS) showed significant associations with ESCC and EC (P<2.46 × 10(-4)). One SNP in MTAP (methylthioadenosine phosphorylase) (rs7023329) that was previously associated with melanoma and nevi in multiple genome-wide association studies was associated with CRC, CA and OS by ASSET (P=0.007).

CONCLUSION

Our data indicate that genetic variants in CDKN2A, and possibly nearby genes, may be associated with ESCC and several other tumours, further highlighting the importance of 9p21.3 genetic variants in carcinogenesis.

Genetic variants in sex hormone metabolic pathway genes and risk of esophageal squamous cell carcinoma

In China, esophageal cancer is the fourth leading cause of cancer death where essentially all cases are histologically esophageal squamous cell carcinoma (ESCC), in contrast to esophageal adenocarcinoma in the West. Globally, ESCC is 2.4 times more common among men than women and recently it has been suggested that sex hormones may be associated with the risk of ESCC. We examined the association between genetic variants in sex hormone metabolic genes and ESCC risk in a population from north central China with high-incidence rates. A total of 1026 ESCC cases and 1452 controls were genotyped for 797 unique tag single-nucleotide polymorphisms (SNPs) in 51 sex hormone metabolic genes. SNP-, gene- and pathway-based associations with ESCC risk were evaluated using unconditional logistic regression adjusted for age, sex and geographical location and the adaptive rank truncated product (ARTP) method. Statistical significance was determined through use of permutation for pathway- and gene-based associations. No associations were observed for the overall sex hormone metabolic pathway (P = 0.14) or subpathways (androgen synthesis: P = 0.30, estrogen synthesis: P = 0.15 and estrogen removal: P = 0.19) with risk of ESCC. However, six individual genes (including SULT2B1, CYP1B1, CYP3A7, CYP3A5, SHBG and CYP11A1) were significantly associated with ESCC risk (P < 0.05). Our examination of genetic variation in the sex hormone metabolic pathway is consistent with a potential association with risk of ESCC. These positive findings warrant further evaluation in relation to ESCC risk and replication in other populations.

Polymorphisms in genes related to one-carbon metabolism are not related to pancreatic cancer in PanScan and PanC4

PURPOSE

The evidence of a relation between folate intake and one-carbon metabolism (OCM) with pancreatic cancer (PanCa) is inconsistent. In this study, the association between genes and single-nucleotide polymorphisms (SNPs) related to OCM and PanCa was assessed.

METHODS

Using biochemical knowledge of the OCM pathway, we identified thirty-seven genes and 834 SNPs to examine in association with PanCa. Our study included 1,408 cases and 1,463 controls nested within twelve cohorts (PanScan). The ten SNPs and five genes with lowest p values (<0.02) were followed up in 2,323 cases and 2,340 controls from eight case-control studies (PanC4) that participated in PanScan2. The correlation of SNPs with metabolite levels was assessed for 649 controls from the European Prospective Investigation into Cancer and Nutrition.

RESULTS

When both stages were combined, we observed suggestive associations with PanCa for rs10887710 (MAT1A) (OR 1.13, 95 %CI 1.04-1.23), rs1552462 (SYT9) (OR 1.27, 95 %CI 1.02-1.59), and rs7074891 (CUBN) (OR 1.91, 95 %CI 1.12-3.26). After correcting for multiple comparisons, no significant associations were observed in either the first or second stage. The three suggested SNPs showed no correlations with one-carbon biomarkers.

CONCLUSIONS

This is the largest genetic study to date to examine the relation between germline variations in OCM-related genes polymorphisms and the risk of PanCa. Suggestive evidence for an association between polymorphisms and PanCa was observed among the cohort-nested studies, but this did not replicate in the case-control studies. Our results do not strongly support the hypothesis that genes related to OCM play a role in pancreatic carcinogenesis.

Genome-wide pathway association studies of multiple correlated quantitative phenotypes using principle component analyses

Genome-wide pathway association studies provide novel insight into the biological mechanism underlying complex diseases. Current pathway association studies primarily focus on single important disease phenotype, which is sometimes insufficient to characterize the clinical manifestations of complex diseases. We present a multi-phenotypes pathway association study(MPPAS) approach using principle component analysis(PCA). In our approach, PCA is first applied to multiple correlated quantitative phenotypes for extracting a set of orthogonal phenotypic components. The extracted phenotypic components are then used for pathway association analysis instead of original quantitative phenotypes. Four statistics were proposed for PCA-based MPPAS in this study. Simulations using the real data from the HapMap project were conducted to evaluate the power and type I error rates of PCA-based MPPAS under various scenarios considering sample sizes, additive and interactive genetic effects. A real genome-wide association study data set of bone mineral density (BMD) at hip and spine were also analyzed by PCA-based MPPAS. Simulation studies illustrated the performance of PCA-based MPPAS for identifying the causal pathways underlying complex diseases. Genome-wide MPPAS of BMD detected associations between BMD and KENNY_CTNNB1_TARGETS_UP as well as LONGEVITYPATHWAY pathways in this study. We aim to provide a applicable MPPAS approach, which may help to gain deep understanding the potential biological mechanism of association results for complex diseases.

On the interplay of telomeres, nevi and the risk of melanoma

The relationship between telomeres, nevi and melanoma is complex. Shorter telomeres have been found to be associated with many cancers and with number of nevi, a known risk factor for melanoma. However, shorter telomeres have also been found to decrease melanoma risk. We performed a systematic analysis of telomere-related genes and tagSNPs within these genes, in relation to the risk of melanoma, dysplastic nevi, and nevus count combining data from four studies conducted in Italy. In addition, we examined whether telomere length measured in peripheral blood leukocytes is related to the risk of melanoma, dysplastic nevi, number of nevi, or telomere-related SNPs. A total of 796 cases and 770 controls were genotyped for 517 SNPs in 39 telomere-related genes genotyped with a custom-made array. Replication of the top SNPs was conducted in two American populations consisting of 488 subjects from 53 melanoma-prone families and 1,086 cases and 1,024 controls from a case-control study. We estimated odds ratios for associations with SNPs and combined SNP P-values to compute gene region-specific, functional group-specific, and overall P-value using an adaptive rank-truncated product algorithm. In the Mediterranean population, we found suggestive evidence that RECQL4, a gene involved in genome stability, RTEL1, a gene regulating telomere elongation, and TERF2, a gene implicated in the protection of telomeres, were associated with melanoma, the presence of dysplastic nevi and number of nevi, respectively. However, these associations were not found in the American samples, suggesting variable melanoma susceptibility for these genes across populations or chance findings in our discovery sample. Larger studies across different populations are necessary to clarify these associations.