IL8 polymorphisms and overall survival in pazopanib- or sunitinib-treated patients with renal cell carcinoma.

BACKGROUND: We evaluated germline single nucleotide polymorphisms (SNPs) for association with overall survival (OS) in pazopanib- or sunitinib-treated patients with advanced renal cell carcinoma (aRCC). METHODS: The discovery analysis tested 27 SNPs within 13 genes from a phase III pazopanib trial (N=241, study 1). Suggestive associations were then pursued in two independent datasets: a phase III trial (COMPARZ) comparing pazopanib vs sunitinib (N=729, study 2) and an observational study of sunitinib-treated patients (N=89, study 3). RESULTS: In study 1, four SNPs showed nominally significant association (P≤0.05) with OS; two of these SNPs (rs1126647, rs4073) in IL8 were associated (P≤0.05) with OS in study 2. Because rs1126647 and rs4073 were highly correlated, only rs1126647 was evaluated in study 3, which also showed association (P≤0.05). In the combined data, rs1126647 was associated with OS after conservative multiple-test adjustment (P=8.8 × 10(-5); variant vs reference allele hazard ratio 1.32, 95% confidence interval: 1.15-1.52), without evidence for heterogeneity of effects between studies or between pazopanib- and sunitinib-treated patients. CONCLUSIONS: Variant alleles of IL8 polymorphisms are associated with poorer survival outcomes in pazopanib- or sunitinib-treated patients with aRCC. These findings provide insight in aRCC prognosis and may advance our thinking in development of new therapies.

Genetic association analysis of vitamin D pathway with obesity traits.

OBJECTIVE: Observational studies have examined the link between vitamin D deficiency and obesity traits. Some studies have reported associations between vitamin D pathway genes such as VDR, GC and CYP27B1 with body mass index (BMI) and waist circumference (WC); however, the findings have been inconsistent. Therefore, we investigated the involvement of vitamin D metabolic pathway genes in obesity-related traits in a large population-based study. METHODS: We undertook a comprehensive analysis between 100 tagging single nucleotide polymorphisms (tagSNPs) in genes encoding for DHCR7, CYP2R1, VDBP, CYP27B1, CYP27A1, CYP24A1, VDR and RXRG, and obesity traits in 5224 participants (aged 45 years) in the 1958 British birth cohort (1958BC). We further extended our analyses to investigate the associations between SNPs and obesity traits using the summary statistics from the GIANT (Genetic Investigation of Anthropometric Traits) consortium (n=123 865). RESULTS: In the 1958BC (n=5224), after Bonferroni correction, none of the tagSNPs were associated with obesity traits except for one tagSNP from CYP24A1 that was associated with waist-hip ratio (WHR) (rs2296239, P=0.001). However, the CYP24A1 SNP was not associated with BMI-adjusted WHR (WHRadj) in the 1958BC (rs2296239, P=1.00) and GIANT results (n=123 865, P=0.18). There was also no evidence for an interaction between the tagSNPs and obesity on BMI, WC, WHR and WHRadj in the 1958BC. In the GIANT consortium, none of the tagSNPs were associated with obesity traits. CONCLUSIONS: Despite a very large study, our findings suggest that the vitamin D pathway genes are unlikely to have a major role in obesity-related traits in the general population.

How informative is a negative finding in a small pharmacogenetic study?

Many pharmacogenetic studies fail to yield any statistically significant associations. Such negative findings may be due to the absence of, or inadequate statistical power to test for, an effect at the genetic variants tested. In many instances, sample sizes are small, making it unclear how to interpret the absence of statistically significant findings. We demonstrate that the amount of information that can be drawn from a negative study is improved by incorporating statistical power and the added context of well-validated pharmacogenetic effects into the interpretation process. This approach permits clearer inferences to be made about the possible range of genetic effects that may be present in, or are likely absent from, small drug studies.

Sequencing of Lp-PLA2-encoding PLA2G7 gene in 2000 Europeans reveals several rare loss-of-function mutations.

Elevated plasma levels of lipoprotein-associated phospholipase A(2) (Lp-PLA2) activity have been shown to be associated with increased risk of coronary heart disease and an inhibitor of this enzyme is under development for the treatment of that condition. A Val279Phe null allele in this gene, that may influence patient eligibility for treatment, is relatively common in East Asians but has not been observed in Europeans. We investigated the existence and functional effects of low frequency alleles in a Western European population by re-sequencing the exons of PLA2G7 in 2000 samples. In all, 19 non-synonymous single-nucleotide polymorphisms (nsSNPs) were found, 14 in fewer than four subjects (minor allele frequency <0.1%). Lp-PLA2 activity was significantly lower in rare nsSNP carriers compared with non-carriers (167.8±63.2 vs 204.6±41.8, P=0.01) and seven variants had enzyme activities consistent with a null allele. The cumulative frequency of these null alleles was 0.25%, so <1 in 10,000 Europeans would be expected to be homozygous, and thus not potentially benefit from treatment with an Lp-PLA2 inhibitor.

A variance components factor model for genetic association studies: a Bayesian analysis.

Studies of gene-trait associations for complex diseases often involve multiple traits that may vary by genotype groups or patterns. Such traits are usually manifestations of lower-dimensional latent factors or disease syndromes. We illustrate the use of a variance components factor (VCF) model to model the association between multiple traits and genotype groups as well as any other existing patient-level covariates. This model characterizes the correlations between traits as underlying latent factors that can be used in clinical decision-making. We apply it within the Bayesian framework and provide a straightforward implementation using the WinBUGS software. The VCF model is illustrated with simulated data and an example that comprises changes in plasma lipid measurements of patients who were treated with statins to lower low-density lipoprotein cholesterol, and polymorphisms from the apolipoprotein-E gene. The simulation shows that this model clearly characterizes existing multiple trait manifestations across genotype groups where individuals' group assignments are fully observed or can be deduced from the observed data. It also allows one to investigate covariate by genotype group interactions that may explain the variability in the traits. The flexibility to characterize such multiple trait manifestations makes the VCF model more desirable than the univariate variance components model, which is applied to each trait separately. The Bayesian framework offers a flexible approach that allows one to incorporate prior information.

The effect of exclusion of cases with unrecorded best estimate of gestational age on the estimates of preterm birth rate.

OBJECTIVES: To investigate the effect of excluding cases with unrecorded best estimate of gestational age at birth on pregnancy outcome reporting and to determine the reasons for unrecorded gestational age data. DESIGN: Prospective study. SETTING: Fifteen maternity units in North West London. POPULATION: 497,105 women who booked for antenatal care from 1988 to 1998. METHOD: Multiple logistic regression analysis. MAIN OUTCOME MEASURES: Preterm birth rate of, and the factors associated with, cases with unrecorded best estimate of gestational age at birth. RESULTS: Of the 53,981 cases with an unrecorded best estimate of gestational age at birth, by using additional data, it was possible to compute a new best estimate of gestational age in 80%. In this latter group, the preterm birth rate was 42% (95% CI 41.5-42.6). The corrected, overall preterm birth rate in North West London (9.8%, 9.7-9.9) was higher than the original estimate (7.6%, 7.5-7.7), which included only cases with recorded data on gestational age at birth. The most significant factors associated with an unrecorded gestational age were no ultrasound scan (OR 49, P < 0.001), and preterm birth <31 weeks (OR 30, P < 0.001). CONCLUSIONS: The incidence of preterm birth are likely to be under-reported in studies where only cases with readily available gestational age data are included. In routinely collected maternity data, human omission is an important contributing factor for an unrecorded best estimate of gestational age at birth. This is associated with the urgent transfer of babies to the neonatal intensive care unit.

The heritability and genetics of complement C3 expression in UK SLE families.

As the central component of the complement system, C3 has sensory and effector functions bridging innate and adaptive immunity. It is plausible that common genetic variation at C3 determines either serum C3 level or susceptibility to systemic lupus erythematosus (SLE), but only a single, Japanese, study has currently showed genetic association. In a cohort of 1371 individuals from 393 UK white European SLE families, we quantified serum C3 and genotyped C3 tagSNPs. Using a Bayesian variance components model, we estimated 39.6% serum C3 heritability. Genotype/serum C3 association was determined by mixed linear models. Single nucleotide polymorphism (SNP) rs344555, located in a haplotype block incorporating the 3' end of C3, was associated with serum C3 (P=0.007), with weaker associations observed for other SNPs in this block. In an extended cohort of 585 SLE families the association between C3 variants and SLE was assessed by transmission disequilibrium test. SNP rs3745568 was associated with SLE (P=0.0046), but not with serum C3. Our disease associated SNP differs from that highlighted in the Japanese study; however, we replicate their finding that genetic variants at the 3' end of C3 are associated with serum C3. Larger studies and further fine mapping will be required to definitively identify functional variants.

Quantification of the genetic component of basal C-reactive protein expression in SLE nuclear families.

C-reactive protein (CRP) is a heritable acute-phase plasma protein also expressed at low, basal, levels in healthy individuals. Elevated basal CRP has been associated with increased cardiovascular risk, while CRP dysregulation may be a feature of systemic lupus erythematosus (SLE). In this cohort of 496 Caucasian SLE families we estimated basal CRP heritability, h(2)= 27.7%. We typed a dense map of CRP single nucleotide polymorphisms (SNPs) and found that seven were associated with basal CRP using both a regression approach and an orthogonal family-based test (P = 0.001-0.011), as were haplotypes carrying the minor allele of these SNPs. SNPs in the interleukin-1beta and interleukin-6 genes were associated with basal CRP. No association was seen between CRP genotype and SLE. Using a variance components approach we estimated that the CRP genotype accounted for only 15% of the total genetic component of basal CRP variation, perhaps explaining the limited evidence of association between CRP and disease. Most of the genetic determinants of basal CRP variation therefore remain unknown. Multiple genes may be involved and identifying them will provide an insight into pathways regulating CRP expression, highlight potential cardiovascular disease and SLE candidates and improve the ability of basal CRP to predict cardiovascular risk.

The use of meta-analysis risk estimates for candidate genes in combination to predict coronary heart disease risk.

Although the risk for coronary heart disease (CHD) associated with single SNPs is modest it has been suggested that, in combination, several common risk-associated alleles could lead to a substantially better heart disease risk prediction. We have modelled this using 10 SNPs in ten candidate genes (APOB, NOS3, APOE, ACE, SERPINE1, MTHFR, ITGA2B, PON 1, LPL, and CETP) and their predicted summary risk estimates from meta-analysis. Based on published allele frequencies, approximately 29% of the general population would be expected to carry less than three risk alleles, approximately 55% would carry 3 or 4 risk alleles, 4% would have 6 and 1% 7 or more risk alleles. Compared to the mean of those with 3 or 4 risk associated genotypes, those with 6 and 7-or-more alleles have a significantly higher risk odds ratio (OR) of CHD (mean OR (95% Confidence Intervals), 1.70 (1.14 to 2.55); and 4.51 (2.89 to 7.04) respectively), while compared to those in the lowest decile of risk, those in the highest decile have a CHD odds ratio in the range of 3.05 (2.24 to 4.14). Taking into account age and the risk alleles carried, the mean 10 year probability for developing CHD for a 55 year old man was calculated to be 15% (8.6% to 24.8%), with nearly 1 in 5 having more than 20% risk. Whether this particular group of 10 SNPs will improve the accuracy of CHD predictions over the combination of classical risk factors in clinical use requires further experimental evidence.

Elevated placental expression of the imprinted PHLDA2 gene is associated with low birth weight.

The identification of genes that regulate fetal growth will help establish the reasons for intrauterine growth restriction. Most autosomal genes are expressed biallelically, but some are imprinted, expressed only from one parental allele. Imprinted genes are associated with fetal growth and development. The growth of the fetus in utero relies on effective nutrient transfer from the mother to the fetus via the placenta. Some current research on the genetic control of fetal growth has focused on genes that display imprinted expression in utero. The expression levels of four imprinted genes, the paternally expressed insulin growth factor 2 (IGF2), the mesoderm-specific transcript isoform 1 (MEST); the maternally expressed pleckstrin homology-like domain, family A, member 2 (PHLDA2); and the polymorphically imprinted insulin-like growth factor 2 (IGF2R) gene are all known to have roles in fetal growth and were studied in the placentae of 200 white European, normal term babies. The quantitative expression analysis with real-time PCR showed the maternally expressing PHLDA2 but not the paternally expressing IGF2 and MEST, nor the polymorphic maternally expressing IGF2R placental levels to have a statistically significant effect on birth weight. PHLDA2 expression levels are negatively correlated with size at birth. These data implicate PHLDA2 as an imprinted gene important in fetal growth and also as a potential marker of fetal growth.

Evidence for unique association signals in SLE at the CD28-CTLA4-ICOS locus in a family-based study.

CD28, CTLA4 (cytotoxic T lymphocyte-associated protein 4) and ICOS (inducible T cell co-stimulator) are good candidate genes for systemic lupus erythematosus (SLE) because of their role in regulating T cell activation. CTLA4 inhibits CD28-mediated T cell activation. CTLA4 is expressed on CD4+ and CD8+ activated T cells, and also B cells, but CD28 and ICOS are largely restricted to T cells. An interval encompassing the CD28-CTLA4-ICOS locus on chromosome 2q33 was linked to lupus in two genome-wide linkage scans. This large family-based association study in 532 UK SLE families represents the first high-density genetic screen of 80 SNPs at this locus. There are seven haplotype blocks across the locus. In CTLA4, the strongest signal comes from two variants, located 2.1 kb downstream from the 3'-UTR. These polymorphisms, rs231726 (SNP 43) and rs231726 (SNP 44), are in complete linkage disequilibrium (LD) (r(2)=1) and are associated with SLE P=0.0008 (GH) and P=0.01 (family-based association test). There is also a signal in the distal 3' flanking region of CTLA4/ICOS promoter (P=0.003). There was no confirmation of published associations for SLE in the promoter or coding region of CTLA4. These SLE risk alleles are more distal than those identified in Graves' disease and are in LD with Graves' disease protective alleles identified in both of these regions of CTLA4 (Ueda et al. 2003). These factors suggest an SLE-specific pattern of association. The functional consequences of the associated polymorphisms are likely to influence CTLA4 expression, although it is possible that genetically modulated ICOS expression is involved in SLE susceptibility.

A likelihood ratio approach to family-based association studies with covariates.

We introduce a procedure for association based analysis of nuclear families that allows for dichotomous and more general measurements of phenotype and inclusion of covariate information. Standard generalized linear models are used to relate phenotype and its predictors. Our test procedure, based on the likelihood ratio, unifies the estimation of all parameters through the likelihood itself and yields maximum likelihood estimates of the genetic relative risk and interaction parameters. Our method has advantages in modelling the covariate and gene-covariate interaction terms over recently proposed conditional score tests that include covariate information via a two-stage modelling approach. We apply our method in a study of human systemic lupus erythematosus and the C-reactive protein that includes sex as a covariate.

Fine mapping of disease genes via haplotype clustering.

We propose an algorithm for analysing SNP-based population association studies, which is a development of that introduced by Molitor et al. [2003: Am J Hum Genet 73:1368-1384]. It uses clustering of haplotypes to overcome the major limitations of many current haplotype-based approaches. We define a between-haplotype score that is simple, yet appears to capture much of the information about evolutionary relatedness of the haplotypes in the vicinity of a (unobserved) putative causal locus. Haplotype clusters can then be defined via a putative ancestral haplotype and a cut-off distance. The number of an individual's two haplotypes that lie within the cluster predicts the individual's genotype at the causal locus. This predicted genotype can then be investigated for association with the phenotype of interest. We implement our approach within a Markov-chain Monte Carlo algorithm that, in effect, searches over locations and ancestral haplotypes to identify large, case-rich clusters. The algorithm successfully fine-maps a causal mutation in a test analysis using real data, and achieves almost 98% accuracy in predicting the genotype at the causal locus. A simulation study indicates that the new algorithm is substantially superior to alternative approaches, and it also allows us to identify situations in which multi-point approaches can substantially improve over single-SNP analyses. Our algorithm runs quickly and there is scope for extension to a wide range of disease models and genomic scales.

SNP selection for association studies: maximizing power across SNP choice and study size.

Selection of single nucleotide polymorphisms (SNPs) is a problem of primary importance in association studies and several approaches have been proposed. However, none provides a satisfying answer to the problem of how many SNPs should be selected, and how this should depend on the pattern of linkage disequilibrium (LD) in the region under consideration. Moreover, SNP selection is usually considered as independent from deciding the sample size of the study. However, when resources are limited there is a tradeoff between the study size and the number of SNPs to genotype. We show that tuning the SNP density to the LD pattern can be achieved by looking for the best solution to this tradeoff. Our approach consists of formulating SNP selection as an optimization problem: the objective is to maximize the power of the final association study, whilst keeping the total costs below a given budget. We also propose two alternative algorithms for the solution of this optimization problem: a genetic algorithm and a hill climbing search. These standard techniques efficiently find good solutions, even when the number of possible SNPs to choose from is large. We compare the performance of these two algorithms on different chromosomal regions and show that, as expected, the selected SNPs reflect the LD pattern: the optimal SNP density varies dramatically between chromosomal regions.

No association between E- and L-selectin genes and SLE: soluble L-selectin levels do correlate with genotype and a subset in SLE.

Altered function of selectin glycoprotein adhesion molecules may modulate severity and organ-specific manifestations of autoimmune and inflammatory disease via changes in leukocyte trafficking. Serum concentrations of selectin molecules have been suggested as useful biomarkers in systemic lupus erythematosus (SLE). We identified increased levels of soluble L-selectin (sL-selectin), but not soluble E-selectin (sE-selectin) in 278 European-Caucasian lupus patients compared to 230 healthy siblings (P=0.002). sL-selectin levels were markedly elevated in patients with IgG antiphospholipid autoantibodies (P=0.002), suggesting that perhaps sL-selectin defines a subgroup of lupus with vasculopathy. sL-selectin level was also influenced by two L-selectin polymorphisms: 665C>T, F206L in the epidermal growth factor-like domain (P=0.015) and rs12938 in the 3'-untranslated region (P=0.06). Having shown increased sL-selectin levels in lupus patients, we used genetics to investigate whether this was a secondary phenomena or the result of an underlying genetic mechanism. The inheritance of nine single-nucleotide polymorphisms (SNP) spanning the selectin locus was tested in 523 UK simplex SLE families. No association with SLE, or related phenotypes, was evident with any single SNP, or haplotype in family-based tests of association. Selectin polymorphisms are, therefore, unlikely to be independent factors in SLE susceptibility.

Little loss of information due to unknown phase for fine-scale linkage-disequilibrium mapping with single-nucleotide-polymorphism genotype data.

We present the results of a simulation study that indicate that true haplotypes at multiple, tightly linked loci often provide little extra information for linkage-disequilibrium fine mapping, compared with the information provided by corresponding genotypes, provided that an appropriate statistical analysis method is used. In contrast, a two-stage approach to analyzing genotype data, in which haplotypes are inferred and then analyzed as if they were true haplotypes, can lead to a substantial loss of information. The study uses our COLDMAP software for fine mapping, which implements a Markov chain-Monte Carlo algorithm that is based on the shattered coalescent model of genetic heterogeneity at a disease locus. We applied COLDMAP to 100 replicate data sets simulated under each of 18 disease models. Each data set consists of haplotype pairs (diplotypes) for 20 SNPs typed at equal 50-kb intervals in a 950-kb candidate region that includes a single disease locus located at random. The data sets were analyzed in three formats: (1). as true haplotypes; (2). as haplotypes inferred from genotypes using an expectation-maximization algorithm; and (3). as unphased genotypes. On average, true haplotypes gave a 6% gain in efficiency compared with the unphased genotypes, whereas inferring haplotypes from genotypes led to a 20% loss of efficiency, where efficiency is defined in terms of root mean integrated square error of the location of the disease locus. Furthermore, treating inferred haplotypes as if they were true haplotypes leads to considerable overconfidence in estimates, with nominal 50% credibility intervals achieving, on average, only 19% coverage. We conclude that (1). given appropriate statistical analyses, the costs of directly measuring haplotypes will rarely be justified by a gain in the efficiency of fine mapping and that (2). a two-stage approach of inferring haplotypes followed by a haplotype-based analysis can be very inefficient for fine mapping, compared with an analysis based directly on the genotypes.

SNP subset selection for genetic association studies.

Association studies for disease susceptibility genes rely on the high density of SNPs within candidate genes. However, the linkage disequilibrium between SNPs imply that not all SNPs identified in the candidate region need be genotyped. Here we develop several approaches to SNP subset selection, which can substantially reduce the number of SNPs to be genotyped in an association study. We apply clustering algorithms to pairwise linkage disequilibrium measures, with SNP subsets determined for different cut-off values of Delta using nearest and furthest neighbour clusters. Alternatively, SNP subsets may be determined by the proportion of haplotypes they identify. We also show how power calculations, based on the average power to identify a SNP as the disease susceptibility mutation using haplotype-based or logistic regression based statistical analyses, can be used to choose SNP subsets. All these methods provide a ranking method for subsets of a specific size, but do not provide criteria for overall choice of SNP subset size. We develop such criteria by incorporating power calculations into a decision analysis, where the choice of SNP subset size depends on the genotyping costs and the perceived benefits of identifying association. These methods are illustrated using eleven SNPs in the MMP2 gene.

Fine-scale mapping of disease loci via shattered coalescent modeling of genealogies.

We present a Bayesian, Markov-chain Monte Carlo method for fine-scale linkage-disequilibrium gene mapping using high-density marker maps. The method explicitly models the genealogy underlying a sample of case chromosomes in the vicinity of a putative disease locus, in contrast with the assumption of a star-shaped tree made by many existing multipoint methods. Within this modeling framework, we can allow for missing marker information and for uncertainty about the true underlying genealogy and the makeup of ancestral marker haplotypes. A crucial advantage of our method is the incorporation of the shattered coalescent model for genealogies, allowing for multiple founding mutations at the disease locus and for sporadic cases of disease. Output from the method includes approximate posterior distributions of the location of the disease locus and population-marker haplotype proportions. In addition, output from the algorithm is used to construct a cladogram to represent genetic heterogeneity at the disease locus, highlighting clusters of case chromosomes sharing the same mutation. We present detailed simulations to provide evidence of improvements over existing methodology. Furthermore, inferences about the location of the disease locus are shown to remain robust to modeling assumptions.

Mapping quantitative trait Loci using generalized estimating equations.

A number of statistical methods are now available to map quantitative trait loci (QTL) relative to markers. However, no existing methodology can simultaneously map QTL for multiple nonnormal traits. In this article we rectify this deficiency by developing a QTL-mapping approach based on generalized estimating equations (GEE). Simulation experiments are used to illustrate the application of the GEE-based approach.

On prediction of genetic values in marker-assisted selection.

We suggest a new approximation for the prediction of genetic values in marker-assisted selection. The new approximation is compared to the standard approach. It is shown that the new approach will often provide substantially better prediction of genetic values; furthermore the new approximation avoids some of the known statistical problems of the standard approach. The advantages of the new approach are illustrated by a simulation study in which the new approximation outperforms both the standard approach and phenotypic selection.