Testing the robustness of the likelihood-ratio test in a variance-component quantitative-trait loci-mapping procedure

Heritability of R2* iron in the basal ganglia and cortex

BACKGROUND

While iron is essential for normal brain functioning, elevated concentrations are commonly found in neurodegenerative diseases and are associated with impaired cognition and neurological deficits. Currently, only little is known about genetic and environmental factors that influence brain iron concentrations.

METHODS

Heritability and bivariate heritability of regional brain iron concentrations, assessed by R2* relaxometry at 3 Tesla MRI, were estimated with variance components models in 130 middle-aged to elderly participants of the Austrian Stroke Prevention Family Study.

RESULTS

Heritability of R2* iron ranged from 0.46 to 0.82 in basal ganglia and from 0.65 to 0.76 in cortical lobes. Age and BMI explained up to 12% and 9% of the variance of R2* iron, while APOE ε4 carrier status, hypertension, diabetes, hypercholesterolemia, sex and smoking explained 5% or less. The genetic correlation of R2* iron among basal ganglionic nuclei and among cortical lobes ranged from 0.78 to 0.87 and from 0.65 to 0.97, respectively. R2* rates in basal ganglia and cortex were not genetically correlated.

CONCLUSIONS

Regional brain iron concentrations are mainly driven by genetic factors while environmental factors contribute to a certain extent. Brain iron levels in the basal ganglia and cortex are controlled by distinct sets of genes.

Heritable anisotropy associated with cognitive impairments among patients with schizophrenia and their non-psychotic relatives in multiplex families

BACKGROUND

To test the functional implications of impaired white matter (WM) connectivity among patients with schizophrenia and their relatives, we examined the heritability of fractional anisotropy (FA) measured on diffusion tensor imaging data acquired in Pittsburgh and Philadelphia, and its association with cognitive performance in a unique sample of 175 multigenerational non-psychotic relatives of 23 multiplex schizophrenia families and 240 unrelated controls (total = 438).

METHODS

We examined polygenic inheritance (h2r) of FA in 24 WM tracts bilaterally, and also pleiotropy to test whether heritability of FA in multiple WM tracts is secondary to genetic correlation among tracts using the Sequential Oligogenic Linkage Analysis Routines. Partial correlation tests examined the correlation of FA with performance on eight cognitive domains on the Penn Computerized Neurocognitive Battery, controlling for age, sex, site and mother's education, followed by multiple comparison corrections.

RESULTS

Significant total additive genetic heritability of FA was observed in all three-categories of WM tracts (association, commissural and projection fibers), in total 33/48 tracts. There were significant genetic correlations in 40% of tracts. Diagnostic group main effects were observed only in tracts with significantly heritable FA. Correlation of FA with neurocognitive impairments was observed mainly in heritable tracts.

CONCLUSIONS

Our data show significant heritability of all three-types of tracts among relatives of schizophrenia. Significant heritability of FA of multiple tracts was not entirely due to genetic correlations among the tracts. Diagnostic group main effect and correlation with neurocognitive performance were mainly restricted to tracts with heritable FA suggesting shared genetic effects on these traits.

A quantitative trait rare variant nonparametric linkage method with application to age-at-onset of Alzheimer's disease

To analyze pedigrees with quantitative trait (QT) and sequence data, we developed a rare variant (RV) quantitative nonparametric linkage (QNPL) method, which evaluates sharing of minor alleles. RV-QNPL has greater power than the traditional QNPL that tests for excess sharing of minor and major alleles. RV-QNPL is robust to population substructure and admixture, locus heterogeneity, and inclusion of nonpathogenic variants and can be readily applied outside of coding regions. When QNPL was used to analyze common variants, it often led to loci mapping to large intervals, e.g., >40 Mb. In contrast, when RVs are analyzed, regions are well defined, e.g., a gene. Using simulation studies, we demonstrate that RV-QNPL is substantially more powerful than applying traditional QNPL methods to analyze RVs. RV-QNPL was also applied to analyze age-at-onset (AAO) data for 107 late-onset Alzheimer's disease (LOAD) pedigrees of Caribbean Hispanic and European ancestry with whole-genome sequence data. When AAO of AD was analyzed regardless of APOE ε4 status, suggestive linkage (LOD = 2.4) was observed with RVs in KNDC1 and nominally significant linkage (p < 0.05) was observed with RVs in LOAD genes ABCA7 and IQCK. When AAO of AD was analyzed for APOE ε4 positive family members, nominally significant linkage was observed with RVs in APOE, while when AAO of AD was analyzed for APOE ε4 negative family members, nominal significance was observed for IQCK and ADAMTS1. RV-QNPL provides a powerful resource to analyze QTs in families to elucidate their genetic etiology.

Fast and powerful genome wide association of dense genetic data with high dimensional imaging phenotypes

Genome wide association (GWA) analysis of brain imaging phenotypes can advance our understanding of the genetic basis of normal and disorder-related variation in the brain. GWA approaches typically use linear mixed effect models to account for non-independence amongst subjects due to factors, such as family relatedness and population structure. The use of these models with high-dimensional imaging phenotypes presents enormous challenges in terms of computational intensity and the need to account multiple testing in both the imaging and genetic domain. Here we present a method that makes mixed models practical with high-dimensional traits by a combination of a transformation applied to the data and model, and the use of a non-iterative variance component estimator. With such speed enhancements permutation tests are feasible, which allows inference on powerful spatial tests like the cluster size statistic.

Genetic Influence on Accessory Navicular Bone in the Foot: A Korean Twin and Family Study

An accessory navicular bone (AN) is the most common accessory ossicle in the foot. The presence of an AN bone can trigger various foot problems, such as posterior tibial tendon pathology, flattening of the medial longitudinal arch, and medial foot pain. Despite the clinical influence of presence of an AN in foot disease, the research regarding its inheritance is still insufficient. A total of 135 pairs of monozygotic (MZ) twins, 25 pairs of dizygotic (DZ) twins, and 676 singletons from families were enrolled in order to estimate genetic influences on AN. After confirmation of zygosity and family relationship with a tandem repeat marker kit and questionnaires, the presence and type of the AN was classified through bilateral feet radiographic examination. The heritability of an AN was estimated using quantitative genetic analysis based on a variance decomposition model considering various types of family relationships: father–offspring pair, mother–offspring pair, and pooled DZ twin and sibling pairs. As a result, approximately 40.96% of the participants in this study had an AN in either foot, with type II being the most common type. The heritability for the presence of any type of an AN in any foot was estimated as 0.88 (95% CI [0.82, 0.94]) after adjusting for age and sex. Specifically, type II AN showed the highest heritability of 0.82 (95% CI [0.71–0.93]). The high heritability of an AN found in this large twin and family study suggests that an AN is determined by the substantial influence of genetic factor.

Sequential recruitment of study participants may inflate genetic heritability estimates

After the success of genome-wide association studies to uncover complex trait loci, attempts to explain the remaining genetic heritability (h ²) are mainly focused on unraveling rare variant associations and gene-gene or gene-environment interactions. Little attention is paid to the possibility that h ² estimates are inflated as a consequence of the epidemiological study design. We studied the time series of 54 biochemical traits in 4373 individuals from the Cooperative Health Research In South Tyrol (CHRIS) study, a pedigree-based study enrolling ten participants/day over several years, with close relatives preferentially invited within the same day. We observed distributional changes of measured traits over time. We hypothesized that the combination of such changes with the pedigree structure might generate a shared-environment component with consequent h ² inflation. We performed variance components (VC) h ² estimation for all traits after accounting for the enrollment period in a linear mixed model (two-stage approach). Accounting for the enrollment period caused a median h ² reduction of 4%. For 9 traits, the reduction was of >20%. Results were confirmed by a Bayesian Markov chain Monte Carlo analysis with all VCs included at the same time (one-stage approach). The electrolytes were the traits most affected by the enrollment period. The h ² inflation was independent of the h ² magnitude, laboratory protocol changes, and length of the enrollment period. The enrollment process may induce shared-environment effects even under very stringent and standardized operating procedures, causing h ² inflation. Including the day of participation as a random effect is a sensitive way to avoid overestimation.

Model-Free Linkage Analysis of a Quantitative Trait

Model-free methods of linkage analysis for quantitative traits are a class of easily implemented, computationally efficient and statistically robust approaches to searching for linkage to a quantitative trait. By "model-free" we refer to methods of linkage analysis that do not fully specify a genetic model (i.e., the causal allele frequency, and penetrance functions). In this chapter we briefly survey the methods that are available, and then we discuss the necessary steps to implement an analysis using the programs GENIBD, SIBPAL and RELPAL in the S.A.G.E. (Statistical Analysis for Genetic Epidemiology) software suite.

Genetic effects on serum testosterone and sex hormone-binding globulin in men: a Korean twin and family study

We conducted a community-based cross-sectional study to evaluate the role of genetics in determining the individual difference in total testosterone and sex hormone-binding globulin levels. Study participants comprised 730 Korean men consisting of 142 pairs of monozygotic twins, 191 pairs of siblings, and 259 father-offspring pairs from 270 families who participated in the Healthy Twin study. Serum concentration of total testosterone and sex hormone-binding globulin were measured by chemiluminescence immunoassay, and free testosterone and bioavailable testosterone were calculated using Vermeulen's method. Quantitative genetic analysis based on a variance decomposition model showed that the heritability of total testosterone, free testosterone, bioavailable testosterone, and sex hormone-binding globulin were 0.56, 0.45, 0.44, and 0.69, respectively after accounting for age and body mass index. Proportions of variance explained by age and body mass index varied across different traits, from 8% for total testosterone to 31% for sex hormone-binding globulin. Bivariate analysis showed a high degree of additive genetic correlation (ρG = 0.67) and a moderate degree of individual-specific environmental correlation (ρE = 0.42) between total testosterone and sex hormone-binding globulin. The findings confirmed the important role of genetics in determining the individually different levels of testosterone and sex hormone-binding globulin during adulthood in Korean men as found in non-Asian populations, which may suggest that common biologic control for determining testosterone level directly or indirectly through binding protein are largely shared among different populations.

Prognostic and Predictive Values and Statistical Interactions in the Era of Targeted Treatment

The current era of targeted treatment has accelerated the interest in studying gene-treatment, gene-gene, and gene-environment interactions using statistical models in the health sciences. Interactions are incorporated into models as product terms of risk factors. The statistical significance of interactions is traditionally examined using a likelihood ratio test (LRT). Epidemiological and clinical studies also evaluate interactions in order to understand the prognostic and predictive values of genetic factors. However, it is not clear how different types and magnitudes of interaction effects are related to prognostic and predictive values. The contribution of interaction to prognostic values can be examined via improvements in the area under the receiver operating characteristic curve due to the inclusion of interaction terms in the model (ΔAUC). We develop a resampling based approach to test the significance of this improvement and show that it is equivalent to LRT. Predictive values provide insights into whether carriers of genetic factors benefit from specific treatment or preventive interventions relative to noncarriers, under some definition of treatment benefit. However, there is no unique definition of the term treatment benefit. We show that ΔAUC and relative excess risk due to interaction (RERI) measure predictive values under two specific definitions of treatment benefit. We investigate the properties of LRT, ΔAUC, and RERI using simulations. We illustrate these approaches using published melanoma data to understand the benefits of possible intervention on sun exposure in relation to the MC1R gene. The goal is to evaluate possible interventions on sun exposure in relation to MC1R.

Fast and powerful heritability inference for family-based neuroimaging studies

Heritability estimation has become an important tool for imaging genetics studies. The large number of voxel- and vertex-wise measurements in imaging genetics studies presents a challenge both in terms of computational intensity and the need to account for elevated false positive risk because of the multiple testing problem. There is a gap in existing tools, as standard neuroimaging software cannot estimate heritability, and yet standard quantitative genetics tools cannot provide essential neuroimaging inferences, like family-wise error corrected voxel-wise or cluster-wise P-values. Moreover, available heritability tools rely on P-values that can be inaccurate with usual parametric inference methods. In this work we develop fast estimation and inference procedures for voxel-wise heritability, drawing on recent methodological results that simplify heritability likelihood computations (Blangero et al., 2013). We review the family of score and Wald tests and propose novel inference methods based on explained sum of squares of an auxiliary linear model. To address problems with inaccuracies with the standard results used to find P-values, we propose four different permutation schemes to allow semi-parametric inference (parametric likelihood-based estimation, non-parametric sampling distribution). In total, we evaluate 5 different significance tests for heritability, with either asymptotic parametric or permutation-based P-value computations. We identify a number of tests that are both computationally efficient and powerful, making them ideal candidates for heritability studies in the massive data setting. We illustrate our method on fractional anisotropy measures in 859 subjects from the Genetics of Brain Structure study.

Genome wide association and linkage analyses identified three loci-4q25, 17q23.2, and 10q11.21-associated with variation in leukocyte telomere length: the Long Life Family Study

Leukocyte telomere length is believed to measure cellular aging in humans, and short leukocyte telomere length is associated with increased risks of late onset diseases, including cardiovascular disease, dementia, etc. Many studies have shown that leukocyte telomere length is a heritable trait, and several candidate genes have been identified, including TERT, TERC, OBFC1, and CTC1. Unlike most studies that have focused on genetic causes of chronic diseases such as heart disease and diabetes in relation to leukocyte telomere length, the present study examined the genome to identify variants that may contribute to variation in leukocyte telomere length among families with exceptional longevity. From the genome wide association analysis in 4,289 LLFS participants, we identified a novel intergenic SNP rs7680468 located near PAPSS1 and DKK2 on 4q25 (p = 4.7E-8). From our linkage analysis, we identified two additional novel loci with HLOD scores exceeding three, including 4.77 for 17q23.2, and 4.36 for 10q11.21. These two loci harbor a number of novel candidate genes with SNPs, and our gene-wise association analysis identified multiple genes, including DCAF7, POLG2, CEP95, and SMURF2 at 17q23.2; and RASGEF1A, HNRNPF, ANF487, CSTF2T, and PRKG1 at 10q11.21. Among these genes, multiple SNPs were associated with leukocyte telomere length, but the strongest association was observed with one contiguous haplotype in CEP95 and SMURF2. We also show that three previously reported genes-TERC, MYNN, and OBFC1-were significantly associated with leukocyte telomere length at p empirical < 0.05.

A generalized Defries-Fulker regression framework for the analysis of twin data

Twin studies compare the similarity between monozygotic twins to that between dizygotic twins in order to investigate the relative contributions of latent genetic and environmental factors influencing a phenotype. Statistical methods for twin data include likelihood estimation and Defries-Fulker regression. We propose a new generalization of the Defries-Fulker model that fully incorporates the effects of observed covariates on both members of a twin pair and is robust to violations of the Normality assumption. A simulation study demonstrates that the method is competitive with likelihood analysis. The Defries-Fulker strategy yields new insight into the parameter space of the twin model and provides a novel, prediction-based interpretation of twin study results that unifies continuous and binary traits. Due to the simplicity of its structure, extensions of the model have the potential to encompass generalized linear models, censored and truncated data; and gene by environment interactions.

Evaluation of a bayesian model integration-based method for censored data

OBJECTIVE

Non-random missing data can adversely affect family-based linkage detection through loss of power and possible introduction of bias depending on how censoring is modeled. We examined the statistical properties of a previously proposed quantitative trait threshold (QTT) model developed for when censored data can be reasonably inferred to be beyond an unknown threshold.

METHODS

The QTT model is a Bayesian model integration approach implemented in the PPL framework that requires neither specification of the threshold nor imputation of the missing data. This model was evaluated under a range of simulated data sets and compared to other methods with missing data imputed.

RESULTS

Across the simulated conditions, the addition of a threshold parameter did not change the PPL's properties relative to quantitative trait analysis on non-censored data except for a slight reduction in the average PPL as a reflection of the lowered information content due to censoring. This remained the case for non-normally distributed data and extreme sampling of pedigrees.

CONCLUSIONS

Overall, the QTT model showed the smallest loss of linkage information relative to alternative approaches and therefore provides a unique analysis tool that obviates the need for ad hoc imputation of censored data in gene mapping studies.

Bivariate Linkage between Acylation-Stimulating Protein and BMI and High-Density Lipoproteins

OBJECTIVE

Given the importance of visceral adiposity in the metabolic syndrome, whether levels of adipokines have shared genetic effects (pleiotropy) with aspects of the metabolic syndrome should be addressed. Acylation-stimulating protein (ASP), an adipose-derived protein, influences lipid metabolism, obesity, and glucose use. Therefore, our objective was to examine the genetic regulation of ASP and associated pleiotropic effects.

RESEARCH METHODS AND PROCEDURES

We assayed serum ASP levels in 435 Mexican Americans participating in the San Antonio Family Heart Study and performed univariate and bivariate variance components analysis.

RESULTS

Additive genetic heritability of ASP was 26% (p = 0.0004). Bivariate genetic analysis detected significant genetic correlations between ASP and several lipid measures but not between ASP and adiposity or diabetes measures. We detected two potential quantitative trait loci influencing ASP levels. The strongest signal was on chromosome 17 near marker D17S1303 [log of the odds ratio (LOD) = 2.7]. The signal on chromosome 15 reached its peak near marker D15S641 (LOD = 2.1). Both signals localize in regions reported to harbor quantitative trait loci influencing obesity and lipid phenotypes in this population. Bivariate linkage analysis yielded LODs of 4.7 for ASP and BMI on chromosome 17 and 3.2 for ASP and high-density lipoprotein2a on chromosome 15.

DISCUSSION

Given these findings, there seems to be a significant genetic contribution to variation in circulating levels of ASP and an interesting pattern of genetic correlation (i.e., pleiotropy) with other risk factors associated with the metabolic syndrome.

Sex-specific regulation of mitochondrial DNA levels: genome-wide linkage analysis to identify quantitative trait loci

Altered mitochondrial DNA (mtDNA) levels have been associated with common diseases in humans. We investigated the genetic mechanism that controls mtDNA levels using genome-wide linkage analyses in families from the Genetic Analysis of Idiopathic Thrombophilia Project (GAIT). We measure mtDNA levels by quantitative real-time PCR in 386 subjects from 21 extended Spanish families. A variance component linkage method using 485 microsatellites was conducted to evaluate linkage and to detect quantitative trait loci (QTLs) involved in the control of mtDNA levels. The heritalibility of mtDNA levels was 0.33 (p=1.82e-05). We identified a QTL on Chromosome 2 (LOD=2.21) using all of the subjects, independently on their sex. When females and males were analysed separately, three QTLs were identified. Females showed the same QTL on Chromosome 2 (LOD=3.09), indicating that the QTL identified in the analysis using all of the subjects was a strong female QTL, and another one on Chromosome 3 (LOD=2.67), whereas in males a QTL was identified on Chromosome 1 (LOD=2.81). These QTLs were fine-mapped to find associations with mtDNA levels. The most significant SNP association was for the rs10888838 on Chromosome 1 in males. This SNP mapped to the gene MRPL37, involved in mitochondrial protein translation. The rs2140855 on Chromosome 2 showed association in the analysis using all of the subjects. It was near the gene CMPK2, which encodes a mitochondrial enzyme of the salvage pathway of deoxyribonucleotide synthesis. Our results provide evidence of a sex-specific genetic mechanism for the control of mtDNA levels and provide a framework to identify new genes that influence mtDNA levels.

QTL mapping reveals the genetic architecture of loci affecting pre- and post-zygotic isolating barriers in Louisiana Iris

BACKGROUND

Hybridization among Louisiana Irises has been well established and the genetic architecture of reproductive isolation is known to affect the potential for and the directionality of introgression between taxa. Here we use co-dominant markers to identify regions where QTL are located both within and between backcross maps to compare the genetic architecture of reproductive isolation and fitness traits across treatments and years.

RESULTS

QTL mapping was used to elucidate the genetic architecture of reproductive isolation between Iris fulva and Iris brevicaulis. Homologous co-dominant EST-SSR markers scored in two backcross populations between I. fulva and I. brevicaulis were used to generate genetic linkage maps. These were used as the framework for mapping QTL associated with variation in 11 phenotypic traits likely responsible for reproductive isolation and fitness. QTL were dispersed throughout the genome, with the exception of one region of a single linkage group (LG) where QTL for flowering time, sterility, and fruit production clustered. In most cases, homologous QTL were not identified in both backcross populations, however, homologous QTL for flowering time, number of growth points per rhizome, number of nodes per inflorescence, and number of flowers per node were identified on several linkage groups.

CONCLUSIONS

Two different traits affecting reproductive isolation, flowering time and sterility, exhibit different genetic architectures, with numerous QTL across the Iris genome controlling flowering time and fewer, less distributed QTL affecting sterility. QTL for traits affecting fitness are largely distributed across the genome with occasional overlap, especially on LG 4, where several QTL increasing fitness and decreasing sterility cluster. Given the distribution and effect direction of QTL affecting reproductive isolation and fitness, we have predicted genomic regions where introgression may be more likely to occur (those regions associated with an increase in fitness and unlinked to loci controlling reproductive isolation) and those that are less likely to exhibit introgression (those regions linked to traits decreasing fitness and reproductive isolation).

Heritability of cognitive functions in families of successful cognitive aging probands from the Central Valley of Costa Rica

We sought to identify cognitive phenotypes for family/genetic studies of successful cognitive aging (SCA; maintaining intact cognitive functioning while living to late old age). We administered a battery of neuropsychological tests to nondemented nonagenarians (n = 65; mean age = 93.4 ± 3.0) and their offspring (n = 188; mean age = 66.4 ± 5.0) from the Central Valley of Costa Rica. After covarying for age, gender, and years of education, as necessary, heritability was calculated for cognitive functions at three pre-defined levels of complexity: specific neuropsychological functions (e.g., delayed recall, sequencing), three higher level cognitive domains (memory, executive functions, attention), and an overall neuropsychological summary. The highest heritability was for delayed recall (h² = 0.74, se = 0.14, p < 0.0001) but significant heritabilities involving memory were also observed for immediate recall (h² = 0.50), memory as a cognitive domain (h² = 0.53), and the overall neuropsychological summary (h² = 0.42). Heritabilities for sequencing (h² = 0.42), fluency (h² = 0.39), abstraction (h² = 0.36), and the executive functions cognitive domain (h² = 0.35) were also significant. In contrast, the attention domain and memory recognition were not significantly heritable in these families. Among the heritable specific cognitive functions, a strong pleiotropic effect (i.e., evidence that these may be influenced by the same gene or set of genes) for delayed and immediate recall was identified (bivariate statistic = 0.934, p < 0.0001) and more modest but significant effects were found for four additional bivariate relationships. The results support the heritability of good cognitive function in old age and the utilization of several levels of phenotypes, and they suggest that several measures involving memory may be especially useful for family/genetic studies of SCA.

Genetic susceptibility to dental caries on pit and fissure and smooth surfaces

Carious lesions are distributed nonuniformly across tooth surfaces of the complete dentition, suggesting that the effects of risk factors may be surface-specific. Whether genes differentially affect caries risk across tooth surfaces is unknown. We investigated the role of genetics on two classes of tooth surfaces, pit and fissure surfaces (PFS) and smooth surfaces (SMS), in more than 2,600 subjects from 740 families. Participants were examined for surface-level evidence of dental caries, and caries scores for permanent and/or primary teeth were generated separately for PFS and SMS. Heritability estimates (h(2), i.e. the proportion of trait variation due to genes) of PFS and SMS caries scores were obtained using likelihood methods. The genetic correlations between PFS and SMS caries scores were calculated to assess the degree to which traits covary due to common genetic effects. Overall, the heritability of caries scores was similar for PFS (h(2) = 19-53%; p < 0.001) and SMS (h(2) = 17-42%; p < 0.001). Heritability of caries scores for both PFS and SMS in the primary dentition was greater than in the permanent dentition and total dentition. With one exception, the genetic correlation between PFS and SMS caries scores was not significantly different from 100%, indicating that (mostly) common genes are involved in the risk of caries for both surface types. Genetic correlation for the primary dentition dfs (decay + filled surfaces) was significantly less than 100% (p < 0.001), indicating that genetic factors may exert differential effects on caries risk in PFS versus SMS in the primary dentition.

Genetic mapping of quantitative trait loci for disease-related phenotypes

Quantitative variation underlies normal as well as pathological traits, and large part of this variability is under the control of genetic loci. Thanks to a better understanding of the extent and nature of human genetic variability and the subsequent availability of an increasing number of genetic markers, genetic mapping of several such quantitative trait loci, or QTLs, has been accomplished in the past 20 years or so using linkage and association analysis in family-based and population-based studies. Rather than alternative, such methods are complementary as each has optimal power of detecting genetic variants underlying variability of quantitative traits under different scenarios defined by the QTL allele frequencies and magnitude of genetic effects. We describe how to apply such analyses to whole-genome or candidate-gene genetic marker data to correlate genetic variability to quantitative trait variability for the purpose of gene mapping and identification.

Model-free linkage analysis of a quantitative trait

Model-free methods of linkage analysis for quantitative traits are a class of easily implemented, computationally efficient, and statistically robust approaches to searching for linkage to a quantitative trait. By "model-free" we refer to methods of linkage analysis that do not fully specify a genetic model (i.e., the causal allele frequency and penetrance functions). In this chapter, we briefly survey the methods that are available, and then we discuss the necessary steps to implement an analysis using the programs GENIBD, SIBPAL, and RELPAL in the Statistical Analysis for Genetic Epidemiology (S.A.G.E.) software suite.

A genome-wide linkage study of mammographic density, a risk factor for breast cancer

INTRODUCTION

Mammographic breast density is a highly heritable (h2 > 0.6) and strong risk factor for breast cancer. We conducted a genome-wide linkage study to identify loci influencing mammographic breast density (MD).

METHODS

Epidemiological data were assembled on 1,415 families from the Australia, Northern California and Ontario sites of the Breast Cancer Family Registry, and additional families recruited in Australia and Ontario. Families consisted of sister pairs with age-matched mammograms and data on factors known to influence MD. Single nucleotide polymorphism (SNP) genotyping was performed on 3,952 individuals using the Illumina Infinium 6K linkage panel.

RESULTS

Using a variance components method, genome-wide linkage analysis was performed using quantitative traits obtained by adjusting MD measurements for known covariates. Our primary trait was formed by fitting a linear model to the square root of the percentage of the breast area that was dense (PMD), adjusting for age at mammogram, number of live births, menopausal status, weight, height, weight squared, and menopausal hormone therapy. The maximum logarithm of odds (LOD) score from the genome-wide scan was on chromosome 7p14.1-p13 (LOD = 2.69; 63.5 cM) for covariate-adjusted PMD, with a 1-LOD interval spanning 8.6 cM. A similar signal was seen for the covariate adjusted area of the breast that was dense (DA) phenotype. Simulations showed that the complete sample had adequate power to detect LOD scores of 3 or 3.5 for a locus accounting for 20% of phenotypic variance. A modest peak initially seen on chromosome 7q32.3-q34 increased in strength when only the 513 families with at least two sisters below 50 years of age were included in the analysis (LOD 3.2; 140.7 cM, 1-LOD interval spanning 9.6 cM). In a subgroup analysis, we also found a LOD score of 3.3 for DA phenotype on chromosome 12.11.22-q13.11 (60.8 cM, 1-LOD interval spanning 9.3 cM), overlapping a region identified in a previous study.

CONCLUSIONS

The suggestive peaks and the larger linkage signal seen in the subset of pedigrees with younger participants highlight regions of interest for further study to identify genes that determine MD, with the goal of understanding mammographic density and its involvement in susceptibility to breast cancer.

Genetic covariation underlying reading, language and related measures in a sample selected for specific language impairment

Specific language impairment is a developmental language disorder characterized by failure to develop language normally in the absence of a specific cause. Previous twin studies have documented the heritability of reading and language measures as well as the genetic correlation between those measures. This paper presents results from an alternative to the classical twin designs by estimating heritability from extended pedigrees. These pedigrees were previously studied as part of series of molecular genetic studies of specific language impairment where the strongest genetic findings were with reading phenotypes rather than language despite selecting pedigrees based on language impairments. To explore the relationship between reading and language in these pedigrees, variance components estimates of heritability of reading and language measures were conducted showing general agreement with the twin literature, as were genetics correlations between reading and language. Phonological short-term memory, phonological awareness and auditory processing were evaluated as candidate mediators of the reading-language genetic correlations. Only phonological awareness showed significant genetic correlations with all reading measures and several language measures while phonological short-term memory and auditory processing did not.

The Maximum-Likelihood-Binomial method revisited: a robust approach for model-free linkage analysis of quantitative traits in large sibships

Model-free linkage analysis methods, based on identity-by-descent allele sharing, are commonly used for complex trait analysis. The Maximum-Likelihood-Binomial (MLB) approach, which is based on the hypothesis that parental alleles are binomially distributed among affected sibs, is particularly popular. An extension of this method to quantitative traits (QT) has been proposed (MLB-QTL), based on the introduction of a latent binary variable capturing information about the linkage between the QT and the marker. Interestingly, the MLB-QTL method does not require the decomposition of sibships into constituent sibpairs and requires no prior assumption about the distribution of the QT. We propose a new formulation of the MLB method for quantitative traits (nMLB-QTL) that explicitly takes advantage of the independence of paternal and maternal allele transmission under the null hypothesis of no linkage. Simulation studies under H₀ showed that the nMLB-QTL method generated very consistent type I errors. Furthermore, simulations under the alternative hypothesis showed that the nMLB-QTL method was slightly, but systematically more powerful than the MLB-QTL method, whatever the genetic model, residual correlation, ascertainment strategy and sibship size considered. Finally, the power of the nMLB-QTL method is illustrated by a chromosome-wide linkage scan for a quantitative endophenotype of leprosy infection. Overall, the nMLB-QTL method is a robust, powerful, and flexible approach for detecting linkage with quantitative phenotypes, particularly in studies of non Gaussian phenotypes in large sibships.

Genome-wide linkage and association scans for quantitative trait Loci of serum lactate dehydrogenase-the framingham heart study

Serum lactate dehydrogenase (LDH) is used in diagnosing many diseases and is significantly determined by genetic factors. Three genes coding for LDH isoenzymes were mapped to chromosome 11q15 and 12p12. We used 330 Framingham Heart Study largest families for microsatellite linkage scan and 100K SNPs association scan to determine quantitative trait loci of LDH level. We estimated the heritability at 41%. Our genome-wide linkage analysis yielded several chromosomal regions, other than 11q and 12p, with LOD scores between 1 and 2.5. None of the 100K SNPs with a P-value <10(-4) in our genome-wide association study was close to the chromosomal regions where the LDH genes reside. Our study demonstrated a strong genetic effect on the variation of LDH levels. There may not be a single gene with a large effect, instead may be several genes with small effects in controlling the variation of serum LDH. Those genes may be located on chromosomal regions that differ from where the genes encoding LDH isoenzymes reside.

Heritability of acoustic startle magnitude, prepulse inhibition, and startle latency in schizophrenia and control families

Prepulse inhibition (PPI) is an acoustic startle paradigm that has been used as an operational measure of sensorimotor gating. Many patients with schizophrenia have impaired PPI, and several lines of evidence suggest that PPI may represent a heritable endophenotype in this disease. We examined startle magnitude and latencies in 40 schizophrenia patients, 58 first-degree relatives of these patients, and 100 healthy controls. After removing low-startlers, we investigated PPI and startle habituation in 34 schizophrenia patients, 43 relatives, and 86 control subjects. Heritability analyses were conducted using a variance-component approach. We found significant heritability of 45% for PPI at the 60-ms interval and 67% for startle magnitude. Onset latency heritability estimates ranged between 39% and 90% across trial types, and those for peak latency ranged from 29% to 68%. Heritability of startle habituation trended toward significance at 31%. We did not detect differences between controls and either schizophrenia patients or their family members for PPI, startle magnitude, or habituation. Startle latencies were generally longer in schizophrenia patients than controls. The heritability findings give impetus to applying genetic analyses to PPI variables, and suggest that startle latency may also be a useful measure in the study of potential endophenotypes for schizophrenia.

Localization of genes for V+LDL plasma cholesterol levels on two diets in the opossum Monodelphis domestica

Plasma cholesterol levels among individuals vary considerably in response to diet. However, the genes that influence this response are largely unknown. Non-HDL (V+LDL) cholesterol levels vary dramatically among gray, short-tailed opossums fed an atherogenic diet, and we previously reported that two quantitative trait loci (QTLs) influenced V+LDL cholesterol on two diets. We used hypothesis-free, genome-wide linkage analyses on data from 325 pedigreed opossums and located one QTL for V+LDL cholesterol on the basal diet on opossum chromosome 1q [logarithm of the odds (LOD) = 3.11, genomic P = 0.019] and another QTL for V+LDL on the atherogenic diet (i.e., high levels of cholesterol and fat) on chromosome 8 (LOD = 9.88, genomic P = 5 x 10(-9)). We then employed a novel strategy involving combined analyses of genomic resources, expression analysis, sequencing, and genotyping to identify candidate genes for the chromosome 8 QTL. A polymorphism in ABCB4 was strongly associated (P = 9 x 10(-14)) with the plasma V+LDL cholesterol concentrations on the high-cholesterol, high-fat diet. The results of this study indicate that genetic variation in ABCB4, or closely linked genes, is responsible for the dramatic differences among opossums in their V+LDL cholesterol response to an atherogenic diet.

Power of competing strategies of linkage analysis for complex traits

Variance components (VC) and the Bayesian Markov chain Monte Carlo (MCMC) analysis are two of the widely used linkage analysis approaches to mapping genes for complex quantitative traits. Both approaches can handle extended pedigrees and multiple markers and do not require a prespecified genetic model. In this study, we used simulated data to compare the performance of these two approaches with the traditional parametric linkage analysis. Using simulated data sets without linkage between a quantitative trait and the markers, we estimated a critical value for various test scores used in VC or MCMC and the location (LOC) score at a fixed level of significance (5%). These critical values were then used to determine the power for the three methods for simulated data sets with linkage. We found that both the VC and MCMC approaches worked well, compared with the LOC score, when there was only one gene underlying the quantitative trait; however, VC had higher power than the other methods in a simulation study of a complex phenotype influenced by more than one gene. We also compared two implementations of MCMC analysis, finding interpretation of results using the log of placement score was more accurate for linkage inference than the Bayes factor but required much more intensive simulation studies.

Bivariate heritability of total and regional brain volumes: the Framingham Study

Heritability and genetic and environmental correlations of total and regional brain volumes were estimated from a large, generally healthy, community-based sample, to determine if there are common elements to the genetic influence of brain volumes and white matter hyperintensity (WMH) volume. There were 1538 Framingham Heart Study participants with brain volume measures from quantitative magnetic resonance imaging who were free of stroke and other neurologic disorders that might influence brain volumes and who were members of families with at least 2 Framingham Heart Study participants. Heritability was estimated using variance component methodology and adjusting for the components of the Framingham stroke risk profile. Genetic and environmental correlations between traits were obtained from bivariate analysis. Heritability estimates ranging from 0.46 to 0.60 were observed for total brain, WMH, hippocampal, temporal lobe, and lateral ventricular volumes. Moderate, yet significant, heritability was observed for the other measures. Bivariate analyses demonstrated that relationships between brain volume measures, except for WMH, reflected both moderate to strong shared genetic and shared environmental influences. This study confirms strong genetic effects on brain and WMH volumes. These data extend current knowledge by showing that these 2 different types of magnetic resonance imaging measures do not share underlying genetic or environmental influences.

Mapping quantitative traits in unselected families: algorithms and examples

Linkage analysis has been widely used to identify from family data genetic variants influencing quantitative traits. Common approaches have both strengths and limitations. Likelihood ratio tests typically computed in variance component analysis can accommodate large families but are highly sensitive to departure from normality assumptions. Regression-based approaches are more robust but their use has primarily been restricted to nuclear families. In this paper, we develop methods for mapping quantitative traits in moderately large pedigrees. Our methods are based on the score statistic, which in contrast to the likelihood ratio statistic can use nonparametric estimators of variability to achieve robustness of the false-positive rate against departures from the hypothesized phenotypic model. Because the score statistic is easier to calculate than the likelihood ratio statistic, our basic mapping methods utilize relatively simple computer code that performs statistical analysis on output from any program that computes estimates of identity by descent. This simplicity also permits development and evaluation of methods to deal with multivariate and ordinal phenotypes, and with gene-gene and gene-environment interaction. We demonstrate our methods on simulated data and on fasting insulin, a quantitative trait measured in the Framingham Heart Study.

Cortical thickness or grey matter volume? The importance of selecting the phenotype for imaging genetics studies

Choosing the appropriate neuroimaging phenotype is critical to successfully identify genes that influence brain structure or function. While neuroimaging methods provide numerous potential phenotypes, their role for imaging genetics studies is unclear. Here we examine the relationship between brain volume, grey matter volume, cortical thickness and surface area, from a genetic standpoint. Four hundred and eighty-six individuals from randomly ascertained extended pedigrees with high-quality T1-weighted neuroanatomic MRI images participated in the study. Surface-based and voxel-based representations of brain structure were derived, using automated methods, and these measurements were analysed using a variance-components method to identify the heritability of these traits and their genetic correlations. All neuroanatomic traits were significantly influenced by genetic factors. Cortical thickness and surface area measurements were found to be genetically and phenotypically independent. While both thickness and area influenced volume measurements of cortical grey matter, volume was more closely related to surface area than cortical thickness. This trend was observed for both the volume-based and surface-based techniques. The results suggest that surface area and cortical thickness measurements should be considered separately and preferred over gray matter volumes for imaging genetic studies.

Genome-wide association and linkage analysis of quantitative traits: comparison of likelihood-ratio test and conditional score statistic

Over the past decade, genetic analysis has shifted from linkage studies, which identify broad regions containing putative trait loci, to genome-wide association studies, which detect the association of a marker with a specific phenotype. Because linkage and association analysis provide complementary information, developing a method to combine these analyses may increase the power to detect a true association. In this paper we compare a linkage score and association score test as well as a newly proposed combination of these two scores with traditional linkage and association methods.

A Variance-Component Framework for Pedigree Analysis of Continuous and Categorical Outcomes

Variance-component methods are popular and flexible analytic tools for elucidating the genetic mechanisms of complex quantitative traits from pedigree data. However, variance-component methods typically assume that the trait of interest follows a multivariate normal distribution within a pedigree. Studies have shown that violation of this normality assumption can lead to biased parameter estimates and inflations in type-I error. This limits the application of variance-component methods to more general trait outcomes, whether continuous or categorical in nature. In this paper, we develop and apply a general variance-component framework for pedigree analysis of continuous and categorical outcomes. We develop appropriate models using generalized-linear mixed model theory and fit such models using approximate maximum-likelihood procedures. Using our proposed method, we demonstrate that one can perform variance-component pedigree analysis on outcomes that follow any exponential-family distribution. Additionally, we also show how one can modify the method to perform pedigree analysis of ordinal outcomes. We also discuss extensions of our variance-component framework to accommodate pedigrees ascertained based on trait outcome. We demonstrate the feasibility of our method using both simulated data and data from a genetic study of ovarian insufficiency.

Suggestive linkage detected for blood pressure related traits on 2q and 22q in the population on the Samoan islands

Background

High blood pressure or hypertension is a major risk factor involved in the development of cardiovascular diseases. We conducted genome-wide variance component linkage analyses to search for loci influencing five blood pressure related traits including the quantitative traits systolic blood pressure (SBP), diastolic blood pressure (DBP) and pulse pressure (PP), the dichotomous trait hypertension (HT) and the bivariate quantitative trait SBP-DBP in families residing in American Samoa and Samoa, as well as in the combined sample from the two polities. We adjusted the traits for a number of environmental covariates such as smoking, alcohol consumption, physical activity and material life style.

Results

We found suggestive univariate linkage for SBP on chromosome 2q35-q37 (LOD 2.4) and for PP on chromosome 22q13 (LOD 2.2), two chromosomal regions that recently have been associated with SBP and PP, respectively.

Conclusion

We have detected additional evidence for a recently reported locus associated with SBP on chromosome 2q and a susceptibility locus for PP on chromosome 22q. However, differences observed between the results from our three partly overlapping genetically homogenous study samples from the Samoan islands suggest that additional studies should be performed in order to verify these results.

Calculating asymptotic significance levels of the constrained likelihood ratio test with application to multivariate genetic linkage analysis

The asymptotic distribution of the multivariate variance component linkage analysis likelihood ratio test has provoked some contradictory accounts in the literature. In this paper we confirm that some previous results are not correct by deriving the asymptotic distribution in one special case. It is shown that this special case is a good approximation to the distribution in many situations. We also introduce a new approach to simulating from the asymptotic distribution of the likelihood ratio test statistic in constrained testing problems. It is shown that this method is very efficient for small p-values, and is applicable even when the constraints are not convex. The method is related to a multivariate integration problem. We illustrate how the approach can be applied to multivariate linkage analysis in a simulation study. Some more philosophical issues relating to one-sided tests in variance components linkage analysis are discussed.

Association of AKT1 with verbal learning, verbal memory, and regional cortical gray matter density in twins

AKT1, encoding the protein kinase B, has been associated with the genetic etiology of schizophrenia and bipolar disorder. However, minuscule data exist on the role of different alleles of AKT1 in measurable quantitative endophenotypes, such as cognitive abilities and neuroanatomical features, showing deviations in schizophrenia and bipolar disorder. We evaluated the contribution of AKT1 to quantitative cognitive traits and 3D high-resolution neuroanatomical images in a Finnish twin sample consisting of 298 twins: 61 pairs with schizophrenia (8 concordant), 31 pairs with bipolar disorder (5 concordant) and 65 control pairs matched for age, sex and demographics. An AKT1 allele defined by the SNP rs1130214 located in the UTR of the gene revealed association with cognitive traits related to verbal learning and memory (P = 0.0005 for a composite index). This association was further fortified by a higher degree of resemblance of verbal memory capacity in pairs sharing the rs1130214 genotype compared to pairs not sharing the genotype. Furthermore, the same allele was also associated with decreased gray matter density in medial and dorsolateral prefrontal cortex (P < 0.05). Our findings support the role of AKT1 in the genetic background of cognitive and anatomical features, known to be affected by psychotic disorders. The established association of the same allelic variant of AKT1 with both cognitive and neuroanatomical aberrations could suggest that AKT1 exerts its effect on verbal learning and memory via neural networks involving prefrontal cortex.

Rank-based inverse normal transformations are increasingly used, but are they merited?

Many complex traits studied in genetics have markedly non-normal distributions. This often implies that the assumption of normally distributed residuals has been violated. Recently, inverse normal transformations (INTs) have gained popularity among genetics researchers and are implemented as an option in several software packages. Despite this increasing use, we are unaware of extensive simulations or mathematical proofs showing that INTs have desirable statistical properties in the context of genetic studies. We show that INTs do not necessarily maintain proper Type 1 error control and can also reduce statistical power in some circumstances. Many alternatives to INTs exist. Therefore, we contend that there is a lack of justification for performing parametric statistical procedures on INTs with the exceptions of simple designs with moderate to large sample sizes, which makes permutation testing computationally infeasible and where maximum likelihood testing is used. Rigorous research evaluating the utility of INTs seems warranted.

Heritability of nonalcoholic fatty liver disease

BACKGROUND & AIMS

Nonalcoholic fatty liver disease (NAFLD) is the most common chronic liver disease in the United States. The etiology is believed to be multifactorial with a substantial genetic component; however, the heritability of NAFLD is undetermined. Therefore, a familial aggregation study was performed to test the hypothesis that NAFLD is highly heritable.

METHODS

Overweight children with biopsy-proven NAFLD and overweight children without NAFLD served as probands. Family members were studied, including the use of magnetic resonance imaging to quantify liver fat fraction. Fatty liver was defined as a liver fat fraction of 5% or higher. Etiologies for fatty liver other than NAFLD were excluded. Narrow-sense heritability estimates for fatty liver (dichotomous) and fat fraction (continuous) were calculated using variance components analysis adjusted for covariate effects.

RESULTS

Fatty liver was present in 17% of siblings and 37% of parents of overweight children without NAFLD. Fatty liver was significantly more common in siblings (59%) and parents (78%) of children with NAFLD. Liver fat fraction was correlated with body mass index, although the correlation was significantly stronger for families of children with NAFLD than those without NAFLD. Adjusted for age, sex, race, and body mass index, the heritability of fatty liver was 1.000 and of liver fat fraction was 0.386.

CONCLUSIONS

Family members of children with NAFLD should be considered at high risk for NAFLD. These data suggest that familial factors are a major determinant of whether an individual has NAFLD. Studies examining the complex relations between genes and environment in the development and progression of NAFLD are warranted.

Bivariate whole-genome linkage scan for bone geometry and total body fat mass

To quantify the genetic correlations between total body fat mass (TBFM) and femoral neck geometric parameters (FNGPs) and, if possible, to detect the specific genomic regions shared by them, bivariate genetic analysis and bivariate whole-genome linkage scan were carried out in a large Caucasian population. All the phenotypes studied were significantly controlled by genetic factors (P < 0.001) with the heritabilities ranging from 0.45 to 0.68. Significantly genetic correlations were found between TBFM and CSA (cross-section area), W (sub-periosteal diameter), Z (section modulus) and CT (cortical thickness) except between TBFM and BR (buckling ratio). The peak bivariate LOD scores were 3.23 (20q12), 2.47 (20p11), 3.19 (6q27), 1.68 (20p12), and 2.47 (7q11) for the five pairs of TBFM and BR, CSA, CT, W, and Z in the entire sample, respectively. Gender-specific bivariate linkage evidences were also found for the five pairs. 6p25 had complete pleiotropic effects on the variations of TBFM & Z in the female sub-population, and 6q27 and 17q11 had coincident linkages for TBFM & CSA and TBFM & Z in the entire population. We identified moderate genetic correlations and several shared genomic regions between TBFM and FNGPs in a large Caucasian population.

Genetic overlap among intelligence and other candidate endophenotypes for schizophrenia

BACKGROUND

A strategy to improve genetic studies of schizophrenia involves the use of endophenotypes. Information on overlapping genetic contributions among endophenotypes may provide additional power, reveal biological pathways, and have practical implications for genetic research. Several cognitive endophenotypes, including intelligence, are likely to be modulated by overlapping genetic influences.

METHODS

We quantified potential genetic and environmental correlations among endophenotypes for schizophrenia, including sensorimotor gating, openness, verbal fluency, early visual perception, spatial working memory, and intelligence, using variance component models in 35 patients and 145 relatives from 25 multigenerational Dutch families multiply affected with schizophrenia.

RESULTS

Significant correlations were found between spatial working memory and intelligence (.45), verbal fluency and intelligence (.36), verbal fluency and spatial working memory (.20), and early visual perception and spatial working memory (.19). A strong genetic correlation (.75) accounted for 76% of the variance shared between spatial working memory and intelligence. Significant environmental correlations were found between verbal fluency and openness (.50) and between verbal fluency and spatial working memory (.58). Sensorimotor gating and openness showed few genetic or environmental correlations with other endophenotypes.

CONCLUSIONS

Our results suggest that intelligence strongly overlaps genetically with a known cognitive endophenotype for schizophrenia. Intelligence may thus be a promising endophenotype for genetic research in schizophrenia, even though the underlying genetic mechanism may still be complex. In contrast, sensorimotor gating and openness appear to represent separate genetic entities with simpler inheritance patterns and may therefore augment the detection of separate genetic pathways contributing to schizophrenia.

A Statistical Variance Components Framework for Mapping Imprinted Quantitative Trait Locus in Experimental Crosses

Current methods for mapping imprinted quantitative trait locus (iQTL) with inbred line crosses assume fixed QTL effects. When an iQTL segregates in experimental line crosses, combining different line crosses with similar genetic background can improve the accuracy of iQTLs inference. In this article, we develop a general interval-based statistical variance components framework to map iQTLs underlying complex traits by combining different backcross line crosses. We propose a new iQTL variance partition method based on the nature of marker alleles shared identical-by-decent (IBD) in inbred lines. Maternal effect is adjusted when testing imprinting. Efficient estimation methods with the maximum likelihood and the restricted maximum likelihood are derived and compared. Statistical properties of the proposed mapping strategy are evaluated through extensive simulations under different sampling designs. An extension to multiple QTL analysis is given. The proposed method will greatly facilitate genetic dissection of imprinted complex traits in inbred line crosses.

Impact of epistasis and QTL x environment interaction on the accumulation of seed mass of soybean (Glycine max L. Merr.)

The accumulation of seed mass in soybean is affected by both genotype and environment. The aim of the present study was to measure additive, epistatic and quantitative trait locus (QTL) x environment (QE) interaction effects of QTLs on the development of 100-seed weight in a population of 143 F5 derived recombinant inbred lines (RILs) developed from the cross between the soybean cultivars 'Charleston' and 'Dong Nong 594'. Broad-sense heritability of 100-seed weight from 30 days (30D) to 80D stages was 0.58, 0.52, 0.62, 0.60, 0.66 and 0.57, respectively. A total of 17 QTLs with conditional additive (a) effect and/or conditional additive x environment interaction (ae) effect at specific stages were identified in ten linkage groups by conditional mapping. Of them, only 4 QTLs had significant a effect or ae effect at different stages of seed development. Among QTLs with significant a effect, five acted positively and six acted negatively on seed development. A total of 35 epistatic pairwise QTLs of 100-seed weight were identified by conditional mapping at different developmental stages. Five pairs of QTL showed the additive x additive epistatic (aa) effect and 16 QTLs showed the aa x environment interaction (aae) effect at the different developmental stages. QTLs with aa effect as well with their environmental interaction effect appeared to vary at different developmental stages. Overall, the results indicated that 100-seed weight in soybean is under developmental, genetic and environmental control.

Selection of the most informative individuals from families with multiple siblings for association studies

Association analyses may follow an initial linkage analysis for mapping and identifying genes underlying complex quantitative traits and may be conducted on unrelated subsets of individuals where only one member of a family is included. We evaluate two methods to select one sibling per sibship when multiple siblings are available: (1) one sibling with the most extreme trait value; and (2) one sibling using a combination score statistic based on extreme trait values and identity-by-descent sharing information. We compare the type I error and power. Furthermore, we compare these selection strategies with a strategy that randomly selects one sibling per sibship and with an approach that includes all siblings, using both simulation study and an application to fasting blood glucose in the Framingham Heart Study. When genetic effect is homogeneous, we find that using the combination score can increase power by 30-40% compared to a random selection strategy, and loses only 8-13% of power compared to the full sibship analysis, across all additive models considered, but offers at least 50% genotyping cost saving. In the presence of genetic heterogeneity, the score offers a 50% increase in power over a random selection strategy, but there is substantial loss compared to the full sibship analysis. In application to fasting blood sample, two SNPs are found in common for the selection strategies and the full sample among the 10 highest ranked single nucleotide polymorphisms. The EV strategy tends to agree with the IBD-EV strategy and the analysis of the full sample.

Bivariate genome-wide linkage analysis for traits BMD and AAM: effect of menopause on linkage signals

Osteoporosis is an age-related systemic skeletal disease, characterized by low bone mineral density (BMD). Low BMD is closely associated with late age at menarche (AAM). Our previous bivariate genome-wide linkage analyses (GWLAs) between BMD and AAM identified two shared genomic regions in 2584 Caucasian females including both pre- and post-menopausal females. However, menopause often causes dramatic bone loss in post-menopausal females; this may introduce some confounding effects on the bivariate GWLA for BMD and AAM. To address the effect of menopause on the identification of genetic factors shared by BMD and AAM, we segregated the previously studied population of 2584 females into two separate subgroups consisting of 1462 pre-menopause subjects and 1122 post-menopausal subjects, and performed further bivariate GWLAs. The BMD was measured by Hologic Dual-energy X-ray (DXA) scanners (Hologic Inc., Bedford, MA, USA). Based on the genome-wide thresholds corrected for multiple testing, we found more significant genomic regions in the pre-menopausal group than in total group (including pre- and post-menopausal women), e.g., we found 4, 1, and 2 shared by spine BMD and AAM, femoral neck (FNK) BMD and AAM and ultra distal (UD) BMD and AAM, respectively. We did not found any significant linkage signals in the post-menopausal group. Importantly, the linkage signals at all significant regions were much stronger in pre-menopausal group than in the other groups: post-menopausal females and total females. For example, the linkage LOD score for FNK BMD and AAM is as high as 4.88 in pre-menopausal females, but only 0.24 and 0.31 in post-menopausal and total females, respectively. These results suggest that menopause introduces some noise signals into GWLAs when estimating the shared genetic factors by BMD and AAM. Therefore, it is very important to classify female subjects properly according to their menopause stage when performing such studies.