A genome-wide linkage and association study of musical aptitude identifies loci containing genes related to inner ear development and neurocognitive functions.

Humans have developed the perception, production and processing of sounds into the art of music. A genetic contribution to these skills of musical aptitude has long been suggested. We performed a genome-wide scan in 76 pedigrees (767 individuals) characterized for the ability to discriminate pitch (SP), duration (ST) and sound patterns (KMT), which are primary capacities for music perception. Using the Bayesian linkage and association approach implemented in program package KELVIN, especially designed for complex pedigrees, several single nucleotide polymorphisms (SNPs) near genes affecting the functions of the auditory pathway and neurocognitive processes were identified. The strongest association was found at 3q21.3 (rs9854612) with combined SP, ST and KMT test scores (COMB). This region is located a few dozen kilobases upstream of the GATA binding protein 2 (GATA2) gene. GATA2 regulates the development of cochlear hair cells and the inferior colliculus (IC), which are important in tonotopic mapping. The highest probability of linkage was obtained for phenotype SP at 4p14, located next to the region harboring the protocadherin 7 gene, PCDH7. Two SNPs rs13146789 and rs13109270 of PCDH7 showed strong association. PCDH7 has been suggested to play a role in cochlear and amygdaloid complexes. Functional class analysis showed that inner ear and schizophrenia-related genes were enriched inside the linked regions. This study is the first to show the importance of auditory pathway genes in musical aptitude.

Arginase I gene single-nucleotide polymorphism is associated with decreased risk of pulmonary hypertension in bronchopulmonary dysplasia.

AIM: To test the hypothesis that there are single-nucleotide polymorphisms (SNPs) in genes of the l-arginine/nitric oxide pathway associated with pulmonary hypertension (PH) in neonates with bronchopulmonary dysplasia (BPD). METHODS: Neonates with BPD were enrolled (n = 140) and clinical characteristics compared between case (BPD + PH) and control (BPD) groups. DNA was isolated from blood leucocytes and assayed for 17 SNPs in l-arginine/nitric oxide pathway genes by Sequenom massarray. Genes included carbamoyl-phosphate synthetase, ornithine transcarbamylase, argininosuccinate synthase, nitric oxide synthase and arginase. SNPs were selected from the National Center for Biotechnology Information database for their putative functionality. Calculated minor allele frequencies (MAF) of cases and controls were compared using χ2 and logistic regression. RESULTS: Of the 140 patients with BPD, 26% had echocardiographic evidence of PH. Ventilation days were longer for cases than controls (mean 31 vs. 15 days, p < 0.05). Of the 17 SNPs, rs2781666 in arginase I gene was less common in cases (MAF = 0.23) than controls (MAF = 0.37, p = 0.04). The odds of PH decreased by 43% (p = 0.047) for each copy of the SNP minor allele in arginase I gene in patients with BPD. CONCLUSION: Arginase I SNP (rs2781666) may be associated with protection against pulmonary hypertension in preterm neonates with BPD.

Measurement of statistical evidence on an absolute scale following thermodynamic principles.

Statistical analysis is used throughout biomedical research and elsewhere to assess strength of evidence. We have previously argued that typical outcome statistics (including p values and maximum likelihood ratios) have poor measure-theoretic properties: they can erroneously indicate decreasing evidence as data supporting an hypothesis accumulate; and they are not amenable to calibration, necessary for meaningful comparison of evidence across different study designs, data types, and levels of analysis. We have also previously proposed that thermodynamic theory, which allowed for the first time derivation of an absolute measurement scale for temperature (T), could be used to derive an absolute scale for evidence (E). Here we present a novel thermodynamically based framework in which measurement of E on an absolute scale, for which "one degree" always means the same thing, becomes possible for the first time. The new framework invites us to think about statistical analyses in terms of the flow of (evidential) information, placing this work in the context of a growing literature on connections among physics, information theory, and statistics.

Practical considerations for dividing data into subsets prior to PPL analysis.

OBJECTIVE: The PPL, a class of statistics for complex trait genetic mapping in humans, utilizes Bayesian sequential updating to accumulate evidence for or against linkage across potentially heterogeneous data (sub)sets. Here, we systematically explore the relative efficacy of alternative subsetting approaches for purposes of PPL calculation. METHODS: We simulated genotypes for three pedigree sets (sib pairs; 2-3 generations; >or=4 generations) based on families from an ongoing study. For each pedigree set, 100 replicates were generated under different levels of heterogeneity (1000 under 'no linkage'). Within each replicate, updating was performed across subsets defined randomly (RAND2, RAND4), by true (TRUE) linkage status, with a realistic (REAL) classification, by individual pedigree (PED), or without any subsetting (NONE). RESULTS: Under 'linkage', REAL yields larger PPLs compared to NONE, RAND2, RAND4, or PED. Under 'no linkage', RAND2, RAND4 and PED yield PPLs close to NONE. CONCLUSIONS: We have examined the impact of different subsetting strategies on the sampling behavior of the PPL. Our results underscore the utility of finding variables that can help delineate more homogeneous data subsets and demonstrate that, once such variables are found, sequential updating can be highly beneficial in the presence of appreciable heterogeneity at a linked locus, without inflation at an unlinked locus.

A posterior probability of linkage-based re-analysis of schizophrenia data yields evidence of linkage to chromosomes 1 and 17.

OBJECTIVE: Linkage analysis using 22 Canadian pedigrees identified a promising schizophrenia candidate region on 1q23 with a maximum 2-point HLOD under a recessive model of 5.8 [Brzustowicz et al. 2000]. In the current study, we revisited this data set using a Bayesian linkage analysis technique, namely the posterior probability of linkage (PPL). METHODS: The PPL has been developed as an alternative to traditional linkage analysis. It differs from both LOD scores and 'non-parametric' methods in that it directly measures the probability of linkage given the data, and incorporates prior genomic information. RESULTS: As expected, PPL results for 1q23 supported the previously observed linkage, with an estimated multipoint PPL of 99.7%. However, the PPL supported two further results: a second peak on chromosome 1 at 1p13 with a multipoint with PPL of 70% and a chromosome 17 marker (D17S784 at 17q25) with a multipoint PPL of 44%. CONCLUSIONS: The PPL-based analysis presented has the advantage over other likelihood-based linkage methods in that it avoids maximization and produces a less complex view of the strength of evidence for linkage.

Examination of AVPR1a as an autism susceptibility gene.

Impaired reciprocal social interaction is one of the core features of autism. While its determinants are complex, one biomolecular pathway that clearly influences social behavior is the arginine-vasopressin (AVP) system. The behavioral effects of AVP are mediated through the AVP receptor 1a (AVPR1a), making the AVPR1a gene a reasonable candidate for autism susceptibility. We tested the gene's contribution to autism by screening its exons in 125 independent autistic probands and genotyping two promoter polymorphisms in 65 autism affected sibling pair (ASP) families. While we found no nonconservative coding sequence changes, we did identify evidence of linkage and of linkage disequilibrium. These results were most pronounced in a subset of the ASP families with relatively less severe impairment of language. Thus, though we did not demonstrate a disease-causing variant in the coding sequence, numerous nontraditional disease-causing genetic abnormalities are known to exist that would escape detection by traditional gene screening methods. Given the emerging biological, animal model, and now genetic data, AVPR1a and genes in the AVP system remain strong candidates for involvement in autism susceptibility and deserve continued scrutiny.

The null distribution of the heterogeneity lod score does depend on the assumed genetic model for the trait.

It is well known that the asymptotic null distribution of the homogeneity lod score (LOD) does not depend on the genetic model specified in the analysis. When appropriately rescaled, the LOD is asymptotically distributed as 0.5 chi(2)(0) + 0.5 chi(2)(1), regardless of the assumed trait model. However, because locus heterogeneity is a common phenomenon, the heterogeneity lod score (HLOD), rather than the LOD itself, is often used in gene mapping studies. We show here that, in contrast with the LOD, the asymptotic null distribution of the HLOD does depend upon the genetic model assumed in the analysis. In affected sib pair (ASP) data, this distribution can be worked out explicitly as (0.5 - c)chi(2)(0) + 0.5chi(2)(1) + cchi(2)(2), where c depends on the assumed trait model. E.g., for a simple dominant model (HLOD/D), c is a function of the disease allele frequency p: for p = 0.01, c = 0.0006; while for p = 0.1, c = 0.059. For a simple recessive model (HLOD/R), c = 0.098 independently of p. This latter (recessive) distribution turns out to be the same as the asymptotic distribution of the MLS statistic under the possible triangle constraint, which is asymptotically equivalent to the HLOD/R. The null distribution of the HLOD/D is close to that of the LOD, because the weight c on the chi(2)(2) component is small. These results mean that the cutoff value for a test of size alpha will tend to be smaller for the HLOD/D than the HLOD/R. For example, the alpha = 0.0001 cutoff (on the lod scale) for the HLOD/D with p = 0.05 is 3.01, while for the LOD it is 3.00, and for the HLOD/R it is 3.27. For general pedigrees, explicit analytical expression of the null HLOD distribution does not appear possible, but it will still depend on the assumed genetic model.

Evidence supporting WNT2 as an autism susceptibility gene.

We examined WNT2 as a candidate disease gene for autism for the following reasons. First, the WNT family of genes influences the development of numerous organs and systems, including the central nervous system. Second, WNT2 is located in the region of chromosome 7q31-33 linked to autism and is adjacent to a chromosomal breakpoint in an individual with autism. Third, a mouse knockout of Dvl1, a member of a gene family essential for the function of the WNT pathway, exhibits a behavioral phenotype characterized primarily by diminished social interaction. We screened the WNT2 coding sequence for mutations in a large number of autistic probands and found two families containing nonconservative coding sequence variants that segregated with autism in those families. We also identified linkage disequilibrium (LD) between a WNT2 3'UTR SNP and our sample of autism-affected sibling pair (ASP) families and trios. The LD arose almost exclusively from a subgroup of our ASP families defined by the presence of severe language abnormalities and was also found to be associated with the evidence for linkage to 7q from our previously published genomewide linkage screen. Furthermore, expression analysis demonstrated WNT2 expression in the human thalamus. Based on these findings, we hypothesize that rare mutations occur in the WNT2 gene that significantly increase susceptibility to autism even when present in single copies, while a more common WNT2 allele (or alleles) not yet identified may exist that contributes to the disorder to a lesser degree.

Power to detect linkage based on multiple sets of data in the presence of locus heterogeneity: comparative evaluation of model-based linkage methods for affected sib pair data.

The development of rigorous methods for evaluating the overall strength of evidence for genetic linkage based on multiple sets of data is becoming increasingly important in connection with genomic screens for complex disorders. We consider here what happens when we attempt to increase power to detect linkage by pooling multiple independently collected sets of families under conditions of variable levels of locus heterogeneity across samples. We show that power can be substantially reduced in pooled samples when compared to the most informative constituent subsamples considered alone, in spite of the increased sample size afforded by pooling. We demonstrate that for affected sib pair data, a simple adaptation of the lod score (which we call the compound lod), which allows for intersample admixture differences can afford appreciably higher power than the ordinary heterogeneity lod; and also, that a statistic we have proposed elsewhere, the posterior probability of linkage, performs at least as well as the compound lod while having considerable computational advantages. The companion paper (this issue, pp 217-225) shows further that in application to multiple data sets, familiar model-free methods are in some sense equivalent to ordinary lod scores based on data pooling, and that they therefore will also suffer dramatic losses in power for pooled data in the presence of locus heterogeneity and other complicating factors.

Comparison of 'model-free' and 'model-based' linkage statistics in the presence of locus heterogeneity: single data set and multiple data set applications.

Earlier work [Knapp et al.: Hum Hered 1994;44:44-51] focusing on affected sib pair (ASP) data established the equivalence between the mean test and a test based on a simple recessive lod score, as well as equivalences between certain forms of the maximum likelihood score (MLS) statistic [Risch: Am J Hum Genet 1990;46:242-253] and particular forms of the lod score. Here we extend the results of Knapp et al. [1994] by reconsidering these equivalences for ASP data, but in the presence of locus heterogeneity. We show that Risch's MLS statistic under the possible triangle constraints [Holmans: Am J Hum Genet 1993;52:362-374] is locally equivalent to the ordinary heterogeneity lod score assuming a simple recessive model (HLOD/R); while the one-parameter MLS assuming no dominance variance is locally equivalent to the (homogeneity) recessive lod. The companion paper (this issue, pp 199-208) showed that when considering multiple data sets in the presence of locus heterogeneity, the HLOD can suffer appreciable losses in power. We show here that in ASP data, these equivalences ensure that this same loss in power is incurred by both forms of the MLS statistic as well. The companion paper also introduced an adaptation of the lod, the compound lod score (HLOD/C). We confirm that the HLOD/C maintains higher power than these 'model-free' methods when applied to multiple heterogeneous data sets, even when it is calculated assuming the wrong genetic model.

Power comparisons between the TDT and two likelihood-based methods.

We compare the statistical power of the transmission disequilibrium test (TDT) with that of two likelihood-based linkage tests, the classical LOD score and a modified LOD score in which a linkage disequilibrium (LD) parameter is incorporated into the likelihood (LD-LOD). We hypothesize that, when LD is present, the LD-LOD will have the greatest power of the three tests because the TDT breaks a multiplex pedigree into triads, and the LOD score has previously been shown to have lower power when LD is present but not accounted for. We test this hypothesis using a simulation study in which we generate affected sib-pair (ASP) pedigrees under a range of genetic models, varying the genotypic relative risk (GRR) from 6 to 16. Because the likelihood-based tests require that a genetic model be specified, we compare the tests under two scenarios. First, we assume the true genetic model in the analysis, and second, we compare the tests when the LD-LOD (LOD) is maximized over two wrong genetic models. For the generating models we considered, we find that the LD-LOD has greater power than the TDT even when the genetic models is mis-specified and the results corrected for multiple tests. Extreme differences occur under the multiplicative and dominant models, for which the difference in power is as high as 40% at complete LD. The LOD score provides the lowest power in the presence of LD for the range of GRR considered here.

Drawbacks of GENEHUNTER for larger pedigrees: application to panic disorder.

Large pedigrees can pose a problem for GENEHUNTER linkage analysis software. Differences in two-point and multipoint lodscores were observed when comparing GENEHUNTER to other linkage software. Careful consideration must be given when selecting linkage analysis programs. Am. J. Med. Genet. (Neuropsychiatr. Genet.) 96:781-783, 2000.

Effect of allelic heterogeneity on the power of the transmission disequilibrium test.

Due in part to an influential paper by Risch and Merikangas [(1996) Science 273:1516-1517], which suggested that disequilibrium tests would have greater power to detect genes of small effect than would linkage tests, interest in the use of the Transmission Disequilibrium Test (TDT) as an analysis tool for genomewide studies is steadily growing. However, the paper by Risch and Merikangas made several simplifying assumptions. One such assumption was that the underlying gene showed allelic homogeneity, and another was that the allele being measured was the actual susceptibility allele. Here we investigate the effect of allelic heterogeneity on the power of the TDT using multiplicative, additive, dominant, and recessive modes of inheritances in the context of a genomewide study. We further distinguish two cases: first, that the marker alleles are the actual susceptibility alleles, and second, that alleles are measured at a marker linked to the disease gene with zero recombination. We consider two family structures, either a single affected offspring (SAO) and two parents, or an affected sib-pair (ASP) and two parents. We find that, as expected, the power of the TDT declines as the number of susceptibility alleles at the locus being tested increases and the effect on power can be substantial. When a linked marker is measured rather than a susceptibility allele itself, sample sizes reach unattainable levels when as few as two susceptibility alleles are present. Across all the models we consider, the required number of families for a TDT with ASP sampling varies from 19 to over a million families. Thus, the TDT may not be an optimal test in the context of genomic screens under more biologically realistic assumptions.

The consistency of the posterior probability of linkage.

When searching for trait loci along the genome, properly incorporating prior genomic information into the analysis will almost certainly increase the chance of success. Recently, we devised a method that utilizes such prior information in the mapping of trait genes for complex disorders (Vieland, 1998; Wang el al. 1999; Vieland et al. 2000). This method uses the posterior probability of linkage (PPL) based on the admixture model as a measure of linkage information. In this paper, we study the consistency of the PPL. It is shown that, as the number of pedigrees increases, the PPL converges in probability to 1 when there is linkage between the marker and a trait locus, and converges to 0 otherwise. This conclusion is shown to be true for general pedigrees and trait models, even, when the likelihood functions are based on misspecified trait models. As part of the effort to prove this conclusion, it is shown that when there is no linkage, the maximum likelihood estimator of the recombination fraction in the admixture model is asymptotically 0.5, even when the admixture model misrepresents the true model.

Further evidence for the increased power of LOD scores compared with nonparametric methods.

In genetic analysis of diseases in which the underlying model is unknown, "model free" methods-such as affected sib pair (ASP) tests-are often preferred over LOD-score methods, although LOD-score methods under the correct or even approximately correct model are more powerful than ASP tests. However, there might be circumstances in which nonparametric methods will outperform LOD-score methods. Recently, Dizier et al. reported that, in some complex two-locus (2L) models, LOD-score methods with segregation analysis-derived parameters had less power to detect linkage than ASP tests. We investigated whether these particular models, in fact, represent a situation that ASP tests are more powerful than LOD scores. We simulated data according to the parameters specified by Dizier et al. and analyzed the data by using a (a) single locus (SL) LOD-score analysis performed twice, under a simple dominant and a recessive mode of inheritance (MOI), (b) ASP methods, and (c) nonparametric linkage (NPL) analysis. We show that SL analysis performed twice and corrected for the type I-error increase due to multiple testing yields almost as much linkage information as does an analysis under the correct 2L model and is more powerful than either the ASP method or the NPL method. We demonstrate that, even for complex genetic models, the most important condition for linkage analysis is that the assumed MOI at the disease locus being tested is approximately correct, not that the inheritance of the disease per se is correctly specified. In the analysis by Dizier et al., segregation analysis led to estimates of dominance parameters that were grossly misspecified for the locus tested in those models in which ASP tests appeared to be more powerful than LOD-score analyses.

Results of a genome-wide genetic screen for panic disorder.

Panic disorder is characterized by spontaneous and recurrent panic attacks, often accompanied by agoraphobia. The results of family, twin, and segregation studies suggest a genetic role in the etiology of the illness. We have genotyped up to 23 families that have a high density of panic disorder with 540 microsatellite DNA markers in a first-pass genomic screen. The thirteen best families (ELOD > 6.0 under the dominant genetic model) have been genotyped with an ordered set of markers encompassing all the autosomes, at an average marker density of 11 cM. Over 110,000 genotypes have been generated on the whole set of families, and the data have been analyzed under both a dominant and a recessive model, and with the program SIBPAIR. No lod scores exceed 2.0 for either parametric model. Two markers give lod scores over 1.0 under the dominant model (chromosomes 1p and 20p), and four do under the recessive model (7p, 17p, 20q, and X/Y). One of these (20p) may be particularly promising. Analysis with SIBPAIR yielded P values equivalent to a lod score of 1.0 or greater (i.e., P < .016, one-sided, uncorrected for multiple tests) for 11 marker loci (2, 7p, 8p, 8q, 9p, 11q, 12q, 16p, 20p and 20q).

Statistical evaluation of age-at-onset anticipation: a new test and evaluation of its behavior in realistic applications.

The discovery that microsatellite repeat expansions can cause clinical disease has fostered renewed interest in testing for age-at-onset anticipation (AOA). A commonly used procedure is to sample affected parent-child pairs (APCPs) from available data sets and to test for a difference in mean age at onset between the parents and the children. However, standard statistical methods fail to take into account the right truncation of both the parent and child age-at-onset distributions under this design, with the result that type I error rates can be inflated substantially. Previously, we had introduced a new test, based on the correct, bivariate right-truncated, age-at-onset distribution. We showed that this test has the correct type I error rate for random APCPs, even for quite small samples. However, in that paper, we did not consider two key statistical complications that arise when the test is applied to realistic data. First, affected pairs usually are sampled from pedigrees preferentially selected for the presence of multiple affected individuals. In this paper, we show that this will tend to inflate the type I error rate of the test. Second, we consider the appropriate probability model under the alternative hypothesis of true AOA due to an expanding microsatellite mechanism, and we show that there is good reason to believe that the power to detect AOA may be quite small, even for substantial effect sizes. When the type I error rate of the test is high relative to the power, interpretation of test results becomes problematic. We conclude that, in many applications, AOA tests based on APCPs may not yield meaningful results.