Improved use of SNP information to detect the role of genes

Determination of the real effect of genes identified in GWAS: the example of IL2RA in multiple sclerosis

Genome-wide association studies (GWAS), although efficient to detect genes involved in complex diseases, are not designed to measure the real effect of the genes. This is illustrated here by the example of IL2RA in multiple sclerosis (MS). Association between IL2RA and MS is clearly established, although the functional variation is still unknown: the effect of IL2RA might be better described by several SNPs than by a single one. This study investigates whether a pair of SNPs better explains the observed linkage and association data than a single SNP. In total, 522 trio families and 244 affected sib-pairs were typed for 26 IL2RA SNPs. For each SNP and pairs of SNPs, the phased genotypes of patients and controls were compared to determine the SNP set offering the best risk discrimination. Consistency between the genotype risks provided by the retained set and the identical by descent allele sharing in affected sib-pairs was assessed. After controlling for multiple testing, the set of SNPs rs2256774 and rs3118470, provides the best discrimination between the case and control genotype distributions (P-corrected=0.009). The relative risk between the least and most at-risk genotypes is 3.54 with a 95% confidence interval of [2.14-5.94]. Furthermore, the linkage information provided by the allele sharing between affected sibs is consistent with the retained set (P=0.80) but rejects the SNP reported in the literature (P=0.006). Establishing a valid modeling of a disease gene is essential to test its potential interaction with other genes and to reconstruct the pathophysiological pathways.

Robust linear regression methods in association studies

MOTIVATION

It is well known that data deficiencies, such as coding/rounding errors, outliers or missing values, may lead to misleading results for many statistical methods. Robust statistical methods are designed to accommodate certain types of those deficiencies, allowing for reliable results under various conditions. We analyze the case of statistical tests to detect associations between genomic individual variations (SNP) and quantitative traits when deviations from the normality assumption are observed. We consider the classical analysis of variance tests for the parameters of the appropriate linear model and a robust version of those tests based on M-regression. We then compare their empirical power and level using simulated data with several degrees of contamination.

RESULTS

Data normality is nothing but a mathematical convenience. In practice, experiments usually yield data with non-conforming observations. In the presence of this type of data, classical least squares statistical methods perform poorly, giving biased estimates, raising the number of spurious associations and often failing to detect true ones. We show through a simulation study and a real data example, that the robust methodology can be more powerful and thus more adequate for association studies than the classical approach.

AVAILABILITY

The code of the robustified version of function lmekin() from the R package kinship is provided as Supplementary Material.

Efficient utilization of rare variants for detection of disease-related genomic regions

When testing association between rare variants and diseases, an efficient analytical approach involves considering a set of variants in a genomic region as the unit of analysis. One factor complicating this approach is that the vast majority of rare variants in practical applications are believed to represent background neutral variation. As a result, analyzing a single set with all variants may not represent a powerful approach. Here, we propose two alternative strategies. In the first, we analyze the subsets of rare variants exhaustively. In the second, we categorize variants selectively into two subsets: one in which variants are overrepresented in cases, and the other in which variants are overrepresented in controls. When the proportion of neutral variants is moderate to large we show, by simulations, that the both proposed strategies improve the statistical power over methods analyzing a single set with total variants. When applied to a real sequencing association study, the proposed methods consistently produce smaller p-values than their competitors. When applied to another real sequencing dataset to study the difference of rare allele distributions between ethnic populations, the proposed methods detect the overrepresentation of variants between the CHB (Chinese Han in Beijing) and YRI (Yoruba people of Ibadan) populations with small p-values. Additional analyses suggest that there is no difference between the CHB and CHD (Chinese Han in Denver) datasets, as expected. Finally, when applied to the CHB and JPT (Japanese people in Tokyo) populations, existing methods fail to detect any difference, while it is detected by the proposed methods in several regions.

Analysis of candidate genes in occurrence and growth of colorectal adenomas

Predisposition to sporadic colorectal tumours is influenced by genes with minor phenotypic effects. A case-control study was set up on 295 patients treated for a large adenoma matched with polyp-free individuals on gender, age, and geographic origin in a 1 : 2 proportion. A second group of 302 patients treated for a small adenoma was also characterized to distinguish effects on adenoma occurrence and growth. We focussed the study on 38 single nucleotide polymorphisms (SNPs) encompassing 14 genes involved in colorectal carcinogenesis. Effect of SNPs was tested using unconditional logistic regression. Comparisons were made for haplotypes within a given gene and for biologically relevant genes combinations using the combination test. The APC p.Glu1317Gly variant appeared to influence the adenoma growth (P = .04, exact test) but not its occurrence. This result needs to be replicated and genome-wide association studies may be necessary to fully identify low-penetrance alleles involved in early stages of colorectal tumorigenesis.

Immunogenetics of drug-induced skin blistering disorders. Part I: Perspective

The overall immunopathogenesis relevant to a large series of disorders caused by a drug or its associated hyperimmune condition is discussed based upon the examination of the genetics of severe drug-induced bullous skin problems (sporadic idiosyncratic adverse events, including Stevens-Johnson syndrome and toxic epidermal necrolysis). An overarching pharmacogenetic schema is proposed. Immune cognition and early-effector processes are focused upon and a challenging synthesis around systems evolution is explained by a variety of projective analogies. Etiology, human leukocyte antigen-B, immune stability, dysregulation, pharmacomimicry, viruses and an aggressive ethnically differentiated 'karmic' response are discussed.

Large recursive partitioning analysis of complex disease pharmacogenetic studies. II. Statistical considerations

Identifying genetic variations predictive of important phenotypes, such as disease susceptibility, drug efficacy, and adverse events, remains a challenging task. There are individual polymorphisms that can be tested one at a time, but there is the more difficult problem of the identification of combinations of polymorphisms or even more complex interactions of genes with environmental factors. Diseases, drug responses or side effects can result from different mechanisms. Identification of subgroups of people where there is a common mechanism is a problem for diagnosis and prescribing of treatment. Recursive partitioning (RP) is a simple statistical tool for segmenting a population into non-overlapping groups where the response of interest, disease susceptibility, drug efficacy and adverse events are more homogeneous within the segments. We suggest that the use of RP is not only more technically feasible than other search methods but it is less susceptible to multiple-testing problems. The numbers of combinations of gene-gene and gene-environment interactions is potentially astronomical and RP greatly reduces the effective search and inference space. Moreover, the certain reliance of RP on the presence of marginal effects is justifiable as was found by using analytical and numerical arguments. In the context of haplotype analysis, results suggest that the analysis of individual SNPs is likely to be successful even when susceptibilities are determined by haplotypes. Retrospective clinical studies where cases and controls are collected will be a common design. This report provides methods that can be used to adjust the RP analysis to reflect the population incidence of the response of interest. Confidence limits on the incidence of the response in the segmented subgroups are also discussed. RP is a straightforward way to create realistic subgroups, and prediction intervals for the within-subgroup disease incidence are easily obtained.

Genetic polymorphisms of matrix metalloproteinase 12 and 13 genes are implicated in endometriosis progression

BACKGROUND

Matrix metalloproteinases (MMPs) may contribute to endometriosis. We tested whether eight functional polymorphisms of these genes could modify the risk of endometriosis.

METHODS

In this case-control study, 227 endometriosis and 241 controls were genotyped for MMP1 -1607 1G/2G, MMP2 -1575 G/A (MMP2.1), -1306 C/T (MMP2.2), MMP3 -1612 5A/6A, MMP7 -153 C/T (MMP7.1), -181 A/G (MMP7.2), MMP12 -82 A/G and MMP13-77 A/G. Association between MMP genotypes and superficial (SUP), deep infiltrating (DIE) and endometriomas (OMA) was tested for each polymorphism separately, using unconditional logistic regression and then for combined genotypes, using the combination test.

RESULTS

When considering all cases, MMP2 polymorphisms were found to be significant, mainly due to DIE (P = 0.023). A small difference between SUP and controls was found for MMP7.2 (P = 0.032) and MMP12 (P = 0.035), in the absence of correction for multiple testing. Using the combination test, the best association when comparing SUP with controls was obtained for MMP12-MMP13 (P = 0.004) for the combined genotype A/G-A/A (odds ratio = 27.60, 95% confidence interval: 2.80-272.40).

CONCLUSIONS

These data show a potential role for MMP12 -82 A/G and MMP13 -77 A/G combined polymorphisms in superficial endometriosis. As no association was found with deep infiltrating endometriosis, this combination of polymorphisms might protect from a more in-depth penetration of tissues.

Modeling the effect of PTPN22 in rheumatoid arthritis

In order to model the effect of PTPN22 on rheumatoid arthritis (RA), we determined the combination of single-nucleotide-polymorphisms (SNPs) showing the strongest association with RA. Three SNPs (rs2476601-rs12730735-rs11102685) were selected for which we estimated the genotypic relative risks (GRRs) of the corresponding genotypes. On the basis of these GRRs we defined four at-risk genotypic classes. Relative to the class of reference risk, individuals had a risk approximately multiplied by two, three, or four. This classification was confirmed by the excess of identity-by-descent (IBD) sharing (IBD = 2) for the sibs of an index in the high-risk class and by excess of non-IBD sharing (IBD = 0) when the index belonged to the low-risk class. The observed data could not be explained by the role of a single variant but were compatible either with a joint effect of the three typed SNPs of PTPN22 on RA or with the role of two untyped variants.

ADAM33, a new candidate for psoriasis susceptibility

Psoriasis is a chronic skin disorder with multifactorial etiology. In a recent study, we reported results of a genome-wide scan on 46 French extended families presenting with plaque psoriasis. In addition to unambiguous linkage to the major susceptibility locus PSORS1 on Chromosome 6p21, we provided evidence for a susceptibility locus on Chromosome 20p13. To follow up this novel psoriasis susceptibility locus we used a family-based association test (FBAT) for an association scan over the 17 Mb candidate region. A total of 85 uncorrelated SNP markers located in 65 genes of the region were initially investigated in the same set of large families used for the genome wide search, which consisted of 295 nuclear families. When positive association was obtained for a SNP, candidate genes nearby were explored more in detail using a denser set of SNPs. Thus, the gene ADAM33 was found to be significantly associated with psoriasis in this family set (The best association was on a 3-SNP haplotype P = 0.00004, based on 1,000,000 permutations). This association was independent of PSORS1. ADAM33 has been previously associated with asthma, which demonstrates that immune system diseases may be controlled by common susceptibility genes with general effects on dermal inflammation and immunity. The identification of ADAM33 as a psoriasis susceptibility gene identified by positional cloning in an outbred population should provide insights into the pathogenesis and natural history of this common disease.

The neuronal sortilin-related receptor SORL1 is genetically associated with Alzheimer disease

The recycling of the amyloid precursor protein (APP) from the cell surface via the endocytic pathways plays a key role in the generation of amyloid beta peptide (Abeta) in Alzheimer disease. We report here that inherited variants in the SORL1 neuronal sorting receptor are associated with late-onset Alzheimer disease. These variants, which occur in at least two different clusters of intronic sequences within the SORL1 gene (also known as LR11 or SORLA) may regulate tissue-specific expression of SORL1. We also show that SORL1 directs trafficking of APP into recycling pathways and that when SORL1 is underexpressed, APP is sorted into Abeta-generating compartments. These data suggest that inherited or acquired changes in SORL1 expression or function are mechanistically involved in causing Alzheimer disease.

Genetic polymorphisms of MMP1, MMP3 and MMP7 gene promoter and risk of colorectal adenoma

Background

Matrix metalloproteinases (MMP) have been shown to play a role in colorectal cancer (CRC). More recently, MMP1, MMP3 and MMP7 functional gene promoter polymorphisms have been found to be associated with CRC occurrence and prognosis. To document the role of MMP polymorphisms in the early step of colorectal carcinogenesis, we investigated their association with colorectal adenoma risk in a case-control study comprising 295 patients with large adenomas (LA), 302 patients with small adenomas (SA) and 568 polyp-free (PF) controls.

Methods

Patients were genotyped using automated fragment analysis for MMP1 -1607 ins/del G and MMP3 -1612 ins/delA (MMP3.1) polymorphisms and allelic discrimination assay for MMP3 -709 A/G (MMP3.2) and MMP7 -181 A/G polymorphisms. Association between MMP genotypes and colorectal adenomas was first tested for each polymorphism separately and then for combined genotypes using the combination test. Adjustment on relevant variables and estimation of odds ratios were performed using unconditional logistic regression.

Results

No association was observed between the polymorphisms and LA when compared to PF or SA. When comparing SA to PF controls, analysis revealed a significant association between MMP3 -1612 ins/delA polymorphism and SA with an increased risk associated with the 6A/6A genotype (OR = 1.67, 95%CI: 1.20–2.34). Using the combination test, the best association was found for MMP3.1-MMP1 (p = 0.001) with an OR of 1.88 (95%CI: 1.08–3.28) for the combined genotype 2G/2G-6A/6A estimated by logistic regression.

Conclusion

These data show a relation between MMP1 -1607 ins/del G and MMP3 -1612 ins/delA combined polymorphisms and risk of SA, suggesting their potential role in the early steps of colorectal carcinogenesis.

Detection of parent-of-origin effects in nuclear families using haplotype analysis

Despite the potential pitfalls of stratification, population-based association studies nowadays are being conducted more often than family-based association studies. However, the mechanism of genomic imprinting has lately been implicated in the etiology of genetic complex diseases and can be detected using statistics only in family-based designs. Powerful tests for association and imprinting have been proposed previously for case-parent trios and single markers. Since the power of association studies can be improved if multiple affected children and haplotypes are considered, we extended the parental asymmetry test (PAT) for imprinting to a test that is suited for both general nuclear families and haplotypes, called HAP-PAT. Significance of the HAP-PAT is determined via a Monte-Carlo simulation procedure. In addition to the HAP-PAT, we modified a haplotype-based association test, proposed by us before, in such a way that either only paternal or maternal transmissions contribute to the test statistic. The approaches were implemented in FAMHAP and we evaluated their performance under a variety of disease models. We were able to demonstrate the usefulness of our haplotype-based approaches to detect parent-of-origin effects. Furthermore, we showed that also in the presence of imprinting it is more reasonable to consider all affected children of a nuclear family, than to randomly select one affected child from each family and to conduct a trio study using the selected individuals.

Clustering of haplotypes based on phylogeny: how good a strategy for association testing?

Haplotypes are now widely used in association studies between markers and disease susceptibility locus. However, when a large number of markers are considered, the number of possible haplotypes increases leading to two problems: an increased number of degrees of freedom that may result in a lack of power and the existence of rare haplotypes that may be difficult to take into account in the statistical analysis. In a recent paper, Durrant et al proposed a method, CLADHC, to group haplotypes based on distance matrices and showed that this could considerably increase the power of the association test as compared to either single-locus analysis or haplotype analysis without prior grouping. Although the authors considered different one-disease-locus susceptibility models in their simulations, they did not study the impact of the linkage disequilibrium (LD) pattern and of the susceptibility allele frequency on their conclusions. Here, we show, using haplotype data from five regions of the genome of different lengths and with different LD patterns, that, when a single disease susceptibility locus is simulated, the prior grouping of haplotypes based on the algorithm of Durrant et al does not increase the power of association testing except in very particular situations of LD patterns and allele frequencies.

Use of pairwise marker combination and recursive partitioning in a pharmacogenetic genome-wide scan

The objective of pharmacogenetic research is to identify a genetic marker, or a set of genetic markers, that can predict how a given person will respond to a given medicine. To search for such marker combinations that are predictive of adverse drug events, we have developed and applied two complementary methods to a pharmacogenetic study of the hypersensitivity reaction (HSR) associated with treatment with abacavir, a medicine that is used to treat HIV-infected patients. Our results show that both of these methods can be used to uncover potentially useful predictive marker combinations. The pairwise marker combination method yielded a collection of marker pairs that featured a spectrum of sensitivities and specificities. Recursive partitioning results led to the genetic delineation of multiple risk categories, including those with extremely high and extremely low risk of HSR. These methods can be readily applied in pharmacogenetic candidate gene studies as well as in genome-wide scans.

Haplotype interaction analysis of unlinked regions

Genetically complex diseases are caused by interacting environmental factors and genes. As a consequence, statistical methods that consider multiple unlinked genomic regions simultaneously are desirable. Such consideration, however, may lead to a vast number of different high-dimensional tests whose appropriate analysis pose a problem. Here, we present a method to analyze case-control studies with multiple SNP data without phase information that considers gene-gene interaction effects while correcting appropriately for multiple testing. In particular, we allow for interactions of haplotypes that belong to different unlinked regions, as haplotype analysis often proves to be more powerful than single marker analysis. In addition, we consider different marker combinations at each unlinked region. The multiple testing issue is settled via the minP approach; the P value of the "best" marker/region configuration is corrected via Monte-Carlo simulations. Thus, we do not explicitly test for a specific pre-defined interaction model, but test for the global hypothesis that none of the considered haplotype interactions shows association with the disease. We carry out a simulation study for case-control data that confirms the validity of our approach. When simulating two-locus disease models, our test proves to be more powerful than association methods that analyze each linked region separately. In addition, when one of the tested regions is not involved in the etiology of the disease, only a small amount of power is lost with interaction analysis as compared to analysis without interaction. We successfully applied our method to a real case-control data set with markers from two genes controlling a common pathway. While classical analysis failed to reach significance, we obtained a significant result even after correction for multiple testing with our proposed haplotype interaction analysis. The method described here has been implemented in FAMHAP.

Contrasting multi-site genotypic distributions among discordant quantitative phenotypes: theAPOA1/C3/A4/A5 gene cluster and cardiovascular disease risk factors

Most tests of association between DNA sequence variation and quantitative phenotypes in samples of randomly chosen individuals rely on specification of genotypic strata followed by comparison of phenotypes across these strata. This strategy often succeeds when phenotypic differences are caused by one or two single nucleotide polymorphisms (SNPs) among the surveyed markers. However, when multiple-SNP haplotypes account for observed phenotypic variation, identification of the best partitioning requires examination of an inordinate number of SNP combinations. An alternative approach is to rank individuals by their phenotypic measures and ask whether attributes of the genotypic variation show a non-random distribution along this phenotypic ranking. One simple version of this strategy selects the top and bottom tails of the distribution, and then tests whether genotypes from these two samples are drawn from a single population. This framework does not require the recovery of phased haplotypes and allows contrasts between large numbers of sites at once. We use a method based on this approach to identify associations between plasma triglyceride level, a risk factor for cardiovascular disease, and multi-site genotypes located in the APOA1/C3/A4/A5 cluster of apolipoprotein genes in unrelated individuals (1,071 African-American females, 780 African-American males, 1,036 European-American females, and 930 European-American males) sampled from four US cities as part of the Coronary Artery Risk Development in Young Adults (CARDIA) study. Method performance is investigated using simulations that model genealogical variation and different genetic architectures. Results indicate that this multi-site test can identify genotype-phenotype associations with reasonable power, including those generated by some simple epistatic models.

Allele variations in the OCA2 gene (pink-eyed-dilution locus) are associated with genetic susceptibility to melanoma

The occuloalbinism 2 (OCA2) gene, localized at 15q11, encodes a melanosomal transmembrane protein that is involved in the most common form of human occulo-cutaneous albinism, a human genetic disorder characterized by fair pigmentation and susceptibility to skin cancer. We wondered whether allele variations at this locus could influence susceptibility to malignant melanoma (MM). In all, 10 intragenic single-nucleotide polymorphisms (SNPs) were genotyped in 113 patients with melanomas and in 105 Caucasian control subjects with no personal or family history of skin cancer. By comparing allelic distribution between cases and controls, we show that MM and OCA2 are associated (p value=0.030 after correction for multiple testing). Then, a recently developed strategy, the 'combination test' enabled us to show that a combination formed by two SNPs was most strongly associated to MM, suggesting a possible interaction between intragenic SNPs. In addition, the role of OCA2 on MM risk was also detected using a logistic model taking into account the presence of variants of the melanocortin 1 receptor gene (MC1R, a key pigmentation gene) and all pigmentation characteristics as melanoma risk factors. Our data demonstrate that a second pigmentation gene, in addition to MC1R, is involved in genetic susceptibility to melanoma.

Exhaustive allelic transmission disequilibrium tests as a new approach to genome-wide association studies

Genome-wide disease-association mapping has been heralded as the study design of the next generation, but the lack of analytical methods to use genotype data fully is a large stumbling block. Here we describe an algorithm and statistical method that efficiently and exhaustively exploits haplotype information by subjecting alleles (a marker or contiguous sets of markers) from sliding windows of all sizes to transmission disequilibrium tests. By applying our method to simulated data and to Hirschsprung disease, we show that it can detect both common and rare disease variants of small effect. These results show that the theoretical benefits of genome-wide association studies are at last realizable.

A powerful strategy to account for multiple testing in the context of haplotype analysis

Haplotypes--that is, linear arrangements of alleles on the same chromosome that were inherited as a unit--are expected to carry important information in the context of association fine mapping of complex diseases. In consideration of a set of tightly linked markers, there is an enormous number of different marker combinations that can be analyzed. Therefore, a severe multiple-testing problem is introduced. One method to deal with this problem is Bonferroni correction by the number of combinations that are considered. Bonferroni correction is appropriate for independent tests but will result in a loss of power in the presence of linkage disequilibrium in the region. A second method is to perform simulations. It is unfortunate that most methods of haplotype analysis already require simulations to obtain an uncorrected P value for a specific marker combination. Thus, it seems that nested simulations are necessary to obtain P values that are corrected for multiple testing, which, apparently, limits the applicability of this approach because of computer running-time restrictions. Here, an algorithm is described that avoids such nested simulations. We check the validity of our approach under two disease models for haplotype analysis of family data. The true type I error rate of our algorithm corresponds to the nominal significance level. Furthermore, we observe a strong gain in power with our method to obtain the global P value, compared with the Bonferroni procedure to calculate the global P value. The method described here has been implemented in the latest update of our program FAMHAP.

Association in Multifactorial Traits: How to Deal with Rare Observations?

To detect the role of a candidate gene for a trait in a sample of individuals, we may test SNP haplotype or diplotype effects. For a limited sample size, many haplotype or diplotype categories may contain few individuals. This involves a power decrease when testing the association between the trait and the haplotypes or diplotypes as these categories provide little additional information while increasing the degrees of freedom. The present paper proposes a new strategy to group rare categories based on a measure of similarity between haplotypes or diplotypes and compares it to two other possible strategies to deal with rare categories: a SNP selection strategy based on haplotype diversity, and a grouping strategy that pools all rare categories into a single baseline group. This comparison is performed by means of simulation under four scenarios. We show that this new strategy shows the largest increase in power irrespective of the model underlying the candidate gene in the studied trait. This strategy therefore provides a powerful alternative to currently used methods to reduce the number of rare categories.