Joint Linkage and Association Analysis Using GENEHUNTER-MODSCORE with an Application to Familial Pancreatic Cancer.

INTRODUCTION: Joint linkage and association (JLA) analysis combines two disease gene mapping strategies: linkage information contained in families and association information contained in populations. Such a JLA analysis can increase mapping power, especially when the evidence for both linkage and association is low to moderate. Similarly, an association analysis based on haplotypes instead of single markers can increase mapping power when the association pattern is complex. METHODS: In this paper, we present an extension to the GENEHUNTER-MODSCORE software package that enables a JLA analysis based on haplotypes and uses information from arbitrary pedigree types and unrelated individuals. Our new JLA method is an extension of the MOD score approach for linkage analysis, which allows the estimation of trait-model and linkage disequilibrium (LD) parameters, i.e., penetrance, disease-allele frequency, and haplotype frequencies. LD is modeled between alleles at a single diallelic disease locus and up to three diallelic test markers. Linkage information is contributed by additional multi-allelic flanking markers. We investigated the statistical properties of our JLA implementation using extensive simulations, and we compared our approach to another commonly used single-marker JLA test. To demonstrate the applicability of our new method in practice, we analyzed pedigree data from the German National Case Collection for Familial Pancreatic Cancer (FaPaCa). RESULTS: Based on the simulated data, we demonstrated the validity of our JLA-MOD score analysis implementation and identified scenarios in which haplotype-based tests outperformed the single-marker test. The estimated trait-model and LD parameters were in good accordance with the simulated values. Our method outperformed another commonly used JLA single-marker test when the LD pattern was complex. The exploratory analysis of the FaPaCa families led to the identification of a promising genetic region on chromosome 22q13.33, which can serve as a starting point for future mutation analysis and molecular research in pancreatic cancer. CONCLUSION: Our newly proposed JLA-MOD score method proves to be a valuable gene mapping and characterization tool, especially when either linkage or association information alone provide insufficient power to identify the disease-causing genetic variants.

Proteomics biomarker discovery for individualized prevention of familial pancreatic cancer using statistical learning.

BACKGROUND: The low five-year survival rate of pancreatic ductal adenocarcinoma (PDAC) and the low diagnostic rate of early-stage PDAC via imaging highlight the need to discover novel biomarkers and improve the current screening procedures for early diagnosis. Familial pancreatic cancer (FPC) describes the cases of PDAC that are present in two or more individuals within a circle of first-degree relatives. Using innovative high-throughput proteomics, we were able to quantify the protein profiles of individuals at risk from FPC families in different potential pre-cancer stages. However, the high-dimensional proteomics data structure challenges the use of traditional statistical analysis tools. Hence, we applied advanced statistical learning methods to enhance the analysis and improve the results' interpretability. METHODS: We applied model-based gradient boosting and adaptive lasso to deal with the small, unbalanced study design via simultaneous variable selection and model fitting. In addition, we used stability selection to identify a stable subset of selected biomarkers and, as a result, obtain even more interpretable results. In each step, we compared the performance of the different analytical pipelines and validated our approaches via simulation scenarios. RESULTS: In the simulation study, model-based gradient boosting showed a more accurate prediction performance in the small, unbalanced, and high-dimensional datasets than adaptive lasso and could identify more relevant variables. Furthermore, using model-based gradient boosting, we discovered a subset of promising serum biomarkers that may potentially improve the current screening procedure of FPC. CONCLUSION: Advanced statistical learning methods helped us overcome the shortcomings of an unbalanced study design in a valuable clinical dataset. The discovered serum biomarkers provide us with a clear direction for further investigations and more precise clinical hypotheses regarding the development of FPC and optimal strategies for its early detection.

Genetic Variation and Cardiovascular Risk Factors: A Cohort Study on Migrants from the Former Soviet Union and a Native German Population.

Resettlers are a large migrant group of more than 2 million people in Germany who migrated mainly from the former Soviet Union to Germany after 1989. We sought to compare the distribution of the major risk factors for cardiovascular disease (CVD) and to investigate the overall genetic differences in a study population which consisted of resettlers and native (autochthone) Germans. This was a joint analysis of two cohort studies which were performed in the region of Augsburg, Bavaria, Germany, with 3363 native Germans and 363 resettlers. Data from questionnaires and physical examinations were used to compare the risk factors for cardiovascular diseases between the resettlers and native Germans. A population-based genome-wide association analysis was performed in order to identify the genetic differences between the two groups. The distribution of the major risk factors for CVD differed between the two groups. The resettlers lead a less active lifestyle. While female resettlers smoked less than their German counterparts, the men showed similar smoking behavior. SNPs from three genes (BTNL2, DGKB, TGFBR3) indicated a difference in the two populations. In other studies, these genes have been shown to be associated with CVD, rheumatoid arthritis and osteoporosis, respectively.

Integrative clinical transcriptome analysis reveals TMPRSS2-ERG dependency of prognostic biomarkers in prostate adenocarcinoma.

In prostate adenocarcinoma (PCa), distinction between indolent and aggressive disease is challenging. Around 50% of PCa are characterized by TMPRSS2-ERG (T2E)-fusion oncoproteins defining two molecular subtypes (T2E-positive/negative). However, current prognostic tests do not differ between both molecular subtypes, which might affect outcome prediction. To investigate gene-signatures associated with metastasis in T2E-positive and T2E-negative PCa independently, we integrated tumor transcriptomes and clinicopathological data of two cohorts (total n = 783), and analyzed metastasis-associated gene-signatures regarding the T2E-status. Here, we show that the prognostic value of biomarkers in PCa critically depends on the T2E-status. Using gene-set enrichment analyses, we uncovered that metastatic T2E-positive and T2E-negative PCa are characterized by distinct gene-signatures. In addition, by testing genes shared by several functional gene-signatures for their association with event-free survival in a validation cohort (n = 272), we identified five genes (ASPN, BGN, COL1A1, RRM2 and TYMS)-three of which are included in commercially available prognostic tests-whose high expression was significantly associated with worse outcome exclusively in T2E-negative PCa. Among these genes, RRM2 and TYMS were validated by immunohistochemistry in another validation cohort (n = 135), and several of them proved to add prognostic information to current clinicopathological predictors, such as Gleason score, exclusively for T2E-negative patients. No prognostic biomarkers were identified exclusively for T2E-positive tumors. Collectively, our study discovers that the T2E-status, which is per se not a strong prognostic biomarker, crucially determines the prognostic value of other biomarkers. Our data suggest that the molecular subtype needs to be considered when applying prognostic biomarkers for outcome prediction in PCa.

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).

No evidence for genome-wide interactions on plasma fibrinogen by smoking, alcohol consumption and body mass index: results from meta-analyses of 80,607 subjects.

Plasma fibrinogen is an acute phase protein playing an important role in the blood coagulation cascade having strong associations with smoking, alcohol consumption and body mass index (BMI). Genome-wide association studies (GWAS) have identified a variety of gene regions associated with elevated plasma fibrinogen concentrations. However, little is yet known about how associations between environmental factors and fibrinogen might be modified by genetic variation. Therefore, we conducted large-scale meta-analyses of genome-wide interaction studies to identify possible interactions of genetic variants and smoking status, alcohol consumption or BMI on fibrinogen concentration. The present study included 80,607 subjects of European ancestry from 22 studies. Genome-wide interaction analyses were performed separately in each study for about 2.6 million single nucleotide polymorphisms (SNPs) across the 22 autosomal chromosomes. For each SNP and risk factor, we performed a linear regression under an additive genetic model including an interaction term between SNP and risk factor. Interaction estimates were meta-analysed using a fixed-effects model. No genome-wide significant interaction with smoking status, alcohol consumption or BMI was observed in the meta-analyses. The most suggestive interaction was found for smoking and rs10519203, located in the LOC123688 region on chromosome 15, with a p value of 6.2 × 10(-8). This large genome-wide interaction study including 80,607 participants found no strong evidence of interaction between genetic variants and smoking status, alcohol consumption or BMI on fibrinogen concentrations. Further studies are needed to yield deeper insight in the interplay between environmental factors and gene variants on the regulation of fibrinogen concentrations.

Empirical hierarchical bayes approach to gene-environment interactions: development and application to genome-wide association studies of lung cancer in TRICL.

The analysis of gene-environment (G × E) interactions remains one of the greatest challenges in the postgenome-wide association studies (GWASs) era. Recent methods constitute a compromise between the robust but underpowered case-control and powerful case-only methods. Inferences of the latter are biased when the assumption of gene-environment (G-E) independence in controls fails. We propose a novel empirical hierarchical Bayes approach to G × E interaction (EHB-GE), which benefits from greater rank power while accounting for population-based G-E correlation. Building on Lewinger et al.'s ([2007] Genet Epidemiol 31:871-882) hierarchical Bayes prioritization approach, the method first obtains posterior G-E correlation estimates in controls for each marker, borrowing strength from G-E information across the genome. These posterior estimates are then subtracted from the corresponding case-only G × E estimates. We compared EHB-GE with rival methods using simulation. EHB-GE has similar or greater rank power to detect G × E interactions in the presence of large numbers of G-E correlations with weak to strong effects or only a low number of such correlations with large effect. When there are no or only a few weak G-E correlations, Murcray et al.'s method ([2009] Am J Epidemiol 169:219-226) identifies markers with low G × E interaction effects better. We applied EHB-GE and competing methods to four lung cancer case-control GWAS from the Interdisciplinary Research in Cancer of the Lung/International Lung Cancer Consortium with smoking as environmental factor. A number of genes worth investigating were identified by the EHB-GE approach.

Deciphering the 8q24.21 association for glioma.

We have previously identified tagSNPs at 8q24.21 influencing glioma risk. We have sought to fine-map the location of the functional basis of this association using data from four genome-wide association studies, comprising a total of 4147 glioma cases and 7435 controls. To improve marker density across the 700 kb region, we imputed genotypes using 1000 Genomes Project data and high-coverage sequencing data generated on 253 individuals. Analysis revealed an imputed low-frequency SNP rs55705857 (P = 2.24 × 10(-38)) which was sufficient to fully capture the 8q24.21 association. Analysis by glioma subtype showed the association with rs55705857 confined to non-glioblastoma multiforme (non-GBM) tumours (P = 1.07 × 10(-67)). Validation of the non-GBM association was shown in three additional datasets (625 non-GBM cases, 2412 controls; P = 1.41 × 10(-28)). In the pooled analysis, the odds ratio for low-grade glioma associated with rs55705857 was 4.3 (P = 2.31 × 10(-94)). rs55705857 maps to a highly evolutionarily conserved sequence within the long non-coding RNA CCDC26 raising the possibility of direct functionality. These data provide additional insights into the aetiological basis of glioma development.

Leveraging ethnic group incidence variation to investigate genetic susceptibility to glioma: a novel candidate SNP approach.

OBJECTIVES: Using a novel candidate SNP approach, we aimed to identify a possible genetic basis for the higher glioma incidence in Whites relative to East Asians and African-Americans. METHODS: We hypothesized that genetic regions containing SNPs with extreme differences in allele frequencies across ethnicities are most likely to harbor susceptibility variants. We used International HapMap Project data to identify 3,961 candidate SNPs with the largest allele frequency differences in Whites compared to East Asians and Africans and tested these SNPs for association with glioma risk in a set of White cases and controls. Top SNPs identified in the discovery dataset were tested for association with glioma in five independent replication datasets. RESULTS: No SNP achieved statistical significance in either the discovery or replication datasets after accounting for multiple testing or conducting meta-analysis. However, the most strongly associated SNP, rs879471, was found to be in linkage disequilibrium with a previously identified risk SNP, rs6010620, in RTEL1. We estimate rs6010620 to account for a glioma incidence rate ratio of 1.34 for Whites relative to East Asians. CONCLUSION: We explored genetic susceptibility to glioma using a novel candidate SNP method which may be applicable to other diseases with appropriate epidemiologic patterns.

RANTES/CCL5 and risk for coronary events: results from the MONICA/KORA Augsburg case-cohort, Athero-Express and CARDIoGRAM studies.

BACKGROUND: The chemokine RANTES (regulated on activation, normal T-cell expressed and secreted)/CCL5 is involved in the pathogenesis of cardiovascular disease in mice, whereas less is known in humans. We hypothesised that its relevance for atherosclerosis should be reflected by associations between CCL5 gene variants, RANTES serum concentrations and protein levels in atherosclerotic plaques and risk for coronary events. METHODS AND FINDINGS: We conducted a case-cohort study within the population-based MONICA/KORA Augsburg studies. Baseline RANTES serum levels were measured in 363 individuals with incident coronary events and 1,908 non-cases (mean follow-up: 10.2±4.8 years). Cox proportional hazard models adjusting for age, sex, body mass index, metabolic factors and lifestyle factors revealed no significant association between RANTES and incident coronary events (HR [95% CI] for increasing RANTES tertiles 1.0, 1.03 [0.75-1.42] and 1.11 [0.81-1.54]). None of six CCL5 single nucleotide polymorphisms and no common haplotype showed significant associations with coronary events. Also in the CARDIoGRAM study (>22,000 cases, >60,000 controls), none of these CCL5 SNPs was significantly associated with coronary artery disease. In the prospective Athero-Express biobank study, RANTES plaque levels were measured in 606 atherosclerotic lesions from patients who underwent carotid endarterectomy. RANTES content in atherosclerotic plaques was positively associated with macrophage infiltration and inversely associated with plaque calcification. However, there was no significant association between RANTES content in plaques and risk for coronary events (mean follow-up 2.8±0.8 years). CONCLUSIONS: High RANTES plaque levels were associated with an unstable plaque phenotype. However, the absence of associations between (i) RANTES serum levels, (ii) CCL5 genotypes and (iii) RANTES content in carotid plaques and either coronary artery disease or incident coronary events in our cohorts suggests that RANTES may not be a novel coronary risk biomarker. However, the potential relevance of RANTES levels in platelet-poor plasma needs to be investigated in further studies.