American Journal of Medical Genetics Part C (Semin. Med. Genet.) 142C:24 –32 (2006) A R T I C L E Current Perspectives on the Genetic Analysis of Autism HILARY COON* Although no definitive genetic mutations leading to autism susceptibility have been established, the field has many new resources to tackle this difficult problem. Numbers of families at many research sites are now large, and new collaborations among these groups will allow for collections of subjects with enough statistical power to detect relatively small gene effects. New technological advances in genotyping will allow for more fine-grained genetic analysis, and more sophisticated techniques have been developed to address the vast amounts of data acquired. Researchers have also begun to focus on intermediate phenotypes associated with autism, such as elevated serotonin, increased head circumference, social difficulties, or language impairment or delay. These phenotypes may help to stratify affected cases into more genetically homogeneous subgroups, and may also occur in their clinically unaffected relatives. The study of intermediate phenotypes may allow investigators to find common gene variants that lead to autism susceptibility. Here we discuss the major intermediate phenotypes, and give an overview of current genetic analysis techniques. ß 2006 Wiley-Liss, Inc. KEY WORDS: autism; intermediate phenotype; genetic linkage; genetic association INTRODUCTION Autism researchers now have an impressive array of methods available for the analysis of genetic data. These methods, together with dense marker data, detailed diagnostic data, and many intermediate phenotypes associated with autism combine to provide the autism genetics community with rich resources to search for autism susceptibility genes. Collaborative efforts have now resulted in large, informative collections of families. Results from multiple genome scans in different samples and with different subsets of the clinical phenotype are beginning to suggest some Hilary Coon, Ph.D., is Associate Professor of Psychiatry and Principal Investigator for the Genetics Component of the Utah Autism Research Program at the University of Utah. She is a statistical geneticist with interest in the analysis of complex traits. Grant sponsor: NICHD; Grant number: 5 U19 HD035476 and R01 MH069359; Grant sponsor: Utah Autism Foundation. *Correspondence to: Hilary Coon, Ph.D., Utah Neurodevelopmental Genetics Project, 421 Wakara Way, Suite 143, Salt Lake City, UT 84108. E-mail: Hilary@bach.med. utah.edu DOI 10.1002/ajmg.c.30079 ß 2006 Wiley-Liss, Inc. convergence of regions, and to direct new research. Within the next 5– 10 years, even larger consortium efforts will begin to reveal results on family collections larger than any yet studied. Collaborative efforts have now resulted in large, informative collections of families. Results from multiple genome scans in different samples and with different subsets of the clinical phenotype are beginning to suggest some convergence of regions, and to direct new research. Within the next 5–10 years, even larger consortium efforts will begin to reveal results on family collections larger than any yet studied. An additional benefit of large collaborative studies may be the ability to characterize genetic and/or phenotypic subsets of subjects. In addition to the availability of large collaborative samples, the autism genetics research community can now take advantage of approximately 9 million single nucleotide polymorphisms (SNPs; e.g., http://www.ncbi.nlm.nih.gov/SNP) and detailed data describing the nonrandom associations of alleles at close genetic markers among populations (linkage disequilibrium) across the genome (HapMap [Daly et al., 2001; Gabriel et al., 2002; Gibbs et al., 2003; Hinds et al., 2005]). Analysis methods have been hard pressed to keep up with the massive influx of genetic data. Linkage analysis has been extended to handle larger datasets [Gudbjartsson et al., 2000; Abecasis et al., 2002], and many new methods for family-based association analysis have been developed [Boehnke and Langefeld, 1998; Horvath and Laird, 1998; Spielman and Ewens, 1998; Clayton, 1999; Martin et al., 2000; Seltman et al., 2003]. Effects of different causal genes across different families (heterogeneity] can be tested ARTICLE AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMIN. MED. GENET.): DOI 10.1002/ajmg.c [Devlin and Roeder, 1999; Pritchard et al., 2000a,b], and techniques have been developed to investigate gene-bygene interactions [Devlin et al., 2003; Ritchie et al., 2003a,b]. The complexity of the phenotype of autism poses challenges for genetic analysis. Autism involves deficits in the three behavioral domains of social interaction, communication, and repetitive and stereotyped behaviors. Autism comprises a heterogeneous spectrum that also includes Asperger syndrome and PDD-Not Otherwise Specified (PDDNOS [American Psychiatric Association, 1994]). Additional phenotypes associated with autism will be described below. Recent studies have suggested that autism is more common than previously thought (from 0.0016 to 0.004) [Bertrand et al., 2001; Chakrabarti and Fombonne, 2001; Muhle et al., 2004]. There is strong evidence for the genetic etiology of autism [Folstein and Rutter, 1977; Steffenburg et al., 1989; Bailey et al., 1996, 1998; Smalley and Collins, 1996; Szatmari et al., 1998; Fombonne, 1999; Chakrabarti and Fombonne, 2001]. Heritability estimates are high, and twin studies suggest non-Mendelian inheritance, with the involvement of at least two and perhaps many interacting genes [Pickles et al., 1995; Risch et al., 1999; Pritchard, 2001]. The recurrence risk, twin concordance data, and extended familial clinical phenotypes support the hypothesis that autism results from a strong, though not absolute, genetic influence. It is likely that genetic and environmental risk factors play complex, interacting roles in causing autism [Jones and Szatmari, 2002]. Genetic studies of autism are well underway. There are now 12 published genome scans of autism spectrum disorders [IMGSAC, 1998; Barrett et al., 1999; Philippe et al., 1999; Buxbaum et al., 2001; IMGSAC, 2001; Liu et al., 2001; Auranen et al., 2002; Shao et al., 2002; Yonan et al., 2003; Buxbaum et al., 2004; Ylisaukkooja et al., 2004; McCauley et al., 2005]. Evidence of locations of possible susceptibility genes can be found in multiple locations across the genome from these studies, with most current efforts centering around candidate regions on chromosome 7q, chromosome 2q, and chromosome 17q in the region of the serotonin transporter locus SLC6A4. Evidence of locations of possible susceptibility genes can be found in multiple locations across the genome from these studies, with most current efforts centering around candidate regions on chromosome 7q, chromosome 2q, and chromosome 17q in the region of the serotonin transporter locus SLC6A4. METHODS Figure 1 gives an overview of the processes involved in the genetic analysis of autism. The sections that follow will take each step in more detail. New Phenotypes Autism genetics research has begun to take advantage of a wider array of phenotypes than clinical diagnosis. Phenotypes associated with autism, called intermediate phenotypes, may serve one of two purposes. First, intermediate phenotypes may serve to stratify affected cases into more genetically homogeneous subgroups. Second, phenotypes that occur in affected subjects may also occur in their clinically unaffected relatives. These phenotypes may be indicators of susceptibility genes, present singly in clinically unaffected family Define and choose phenotypes for analysis Rates and distribution in autism subjects and relatives (or control sample), sufficient variation, segregation in pedigrees, heritability Choose candidate gene(s) (association study) Choose marker set (pedigree linkage study) Positional (under linkage peaks), functional variants in known genes Traditional DNA markers, Single Nucleotide Polymorphisms (SNPs) Check data/ preliminary analysis Genotypes: misinheritance, map distance, adminxture, allele frequencies Phenotypes: outliers, distributions, effects of covariates Linkage analysis (pedigree study) Affection status with or without intermediate phenotypes Association analysis (case/control or family-based association study) Simple chi-square tests, transmission disequilibrium tests, family based association analysis Verify findings Secondary analyses. For linkage, check if results are robust to: allele frequencies, genetic models, maps, analysis methods. For candidate genes verify findings with, expression studies. For both, verify findings with independent replication studies. OPTION: Refine phenotype or define phenotype subset, then return to the top. Figure 1. 25 Schematic of steps in the genetic analysis of autism. 26 AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMIN. MED. GENET.): DOI 10.1002/ajmg.c members, but present together with other susceptibility genes or environmental factors in persons with autism. The study of an intermediate phenotype may therefore reveal one of many common interacting genes. The following brief review presents several promising intermediate phenotypes for autism. Broader Autism Phenotype (BAP) The BAP is characterized by impairments in the three core domains of autism (social behaviors, communication, repetitive behaviors) present in a milder form than required for diagnosis. Besides The Broader Autism Phenotype characterized by impairments in the three core domains of autism (social behaviors, communication, repetitive behaviors) present in a milder form than required for diagnosis. traditional autism, families with an autistic child show increased rates of the BAP (12%–25%) [Folstein and Piven, 1991; Bolton et al., 1994; Bailey et al., 1996]. Piven and Folstein have developed an algorithm that incorporates items measuring rigidity, friendship patterns, pragmatic language and other features to assign membership to the BAP. Using this algorithm, 50% of parents of autistic children manifest the broader phenotype, compared to 2% of parents with a Down syndrome child [Piven et al., 1997a,b]. Szatmari et al.  have found similar results for parents, but not in more extended family members. New measures are being developed specifically for genetic analysis of the BAP. For example, the Social Responsiveness Scale [Constantino et al., 2000, 2003] is a new, quantitative measure of social deficits characteristic of autism spectrum disorders that shows promise for future genetic studies of autism. Language function Language function is a specific aspect of the BAP that has been studied quite intensively, and may define the most heritable aspect of the BAP [Piven et al., 1997b; Folstein et al., 1999; Lord et al., 2001; Dawson et al., 2002]. Language has already proven to be a useful phenotype for refining genetic studies of autism [Alarcon et al., 2002; Bradford et al., 2001; Buxbaum et al., 2001]. Developmental language disorders, particularly specific language impairment, have shown strong genetic etiology [Rapin, 1996; Wassink and Piven, 2000; Bartlett et al., 2002; Newbury and Monaco, 2002; Silverman et al., 2002]. Language characteristics include deficits in phonologic processing, language development, speech, and reading. Familial deficits in these characteristics have been identified in parents and siblings of probands with autism [Plumet et al., 1995; Piven et al., 1997b; Wolk and Giesen, 2000; Silverman et al., 2002]. Language characteristics include deficits in phonologic processing, language development, speech, and reading. Familial deficits in these characteristics have been identified in parents and siblings of probands with autism. Serotonin and autism Elevated serotonin may identify families at increased risk for autism, as suggested by several lines of evidence. Studies of blood levels in probands find an elevation in approximately 20%–30% of cases [Anderson et al., 1987; Cook and Leventhal, 1996; Anderson, 2002]. Elevation of serotonin is present in family members of autistic subjects with hyperserotonemia [Kuperman et al., 1985; Abramson et al., 1989; Cook et al., 1990, 1994; Leventhal et al., 1990; Piven et al., 1991]. Pharmacologic studies implicate ARTICLE the serotonin system in autism [Cook et al., 1994; McDougle et al., 1996; Potenza and McDougle, 1997]. In addition to recent genome-wide linkage findings on 17q, there is specific evidence for the involvement of the serotonin transporter protein gene on 17q (HTT) [Cook et al., 1997; Klauck et al., 1997; Marazziti et al., 2000; Yirmiya et al., 2001, Tordjman et al., 2001; Kim et al., 2002]. Head circumference Abnormalities in head circumference in autism have been well documented. In particular, macrocephaly (head size over the 97th centile for age/sex) occurs in about 20% of autistic individuals [Fombonne, 1999], and is also common in autism spectrum disorders [Woodhouse et al., 1996]. Investigations Abnormalities in head circumference in autism have been well documented. In particular, macrocephaly (head size over the 97th centile for age/sex) occurs in about 20% of autistic individuals and is also common in autism spectrum disorders. of head circumference as a quantitative phenotype show increases in quantitative head circumference in autism [Woodhouse et al., 1996; Lainhart et al., 1997]. This increase in head size has been shown through neuroimaging studies to be due to abnormal enlargement of the brain. Head circumference may prove to be an interesting and easily measured stratification variable for association studies. Findings using this simple measure on large data sets could then be explored in interesting subsets using more expensive MRI techniques. Choosing intermediate phenotypes To date, intermediate phenotypes studied have been primarily language ARTICLE AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMIN. MED. GENET.): DOI 10.1002/ajmg.c delays, repetitive behaviors, and social behaviors. Additional phenotypic traits should be chosen with demonstrated genetic etiology (significant heritability). Heritability is the proportion of variance in a phenotype attributable to genes. Estimates of the heritability of quantitative traits on large pedigrees can be obtained using several analysis programs (e.g., SOLAR [Almasy and Blangero, 1998] or PAP [Hasstedt and Cartwright, 1979]). Intermediate phenotypes may be interdependent. The extent to which phenotypes cosegregate in families can be tested in the Pedigree Analysis Package (PAP [Hasstedt and Cartwright, 1979]). The PAP model assumes that variation in each phenotype is due to the sum of independent effects of shared genes and shared environment among family members, as well as random environmental effects specific to each individual. The program estimates separate genetic and environmental components of the covariance between two phenotypes. The genetic estimate indicates the extent to which the phenotypes are influenced by genes common to them both. Data Checking and Preliminary Analysis Before embarking on an analysis of genetic data, error checking should be done, both for phenotypes and genotypes. Such checking can potentially avoid false negative findings. Laboratories use internal genotyping controls, and have stringent criteria for accepting or deleting uncertain genotypes using lab checking protocols. Genotype checking Checking for remaining inconsistencies can be done using various genotype cleaning programs, such as PEDSYS (http://www. sfbr.org/sfbr/public/software/pedsys/pedsys.html). Software such as CRIMAP [Lander and Green, 1987] can be used to verify observed map distances compared to fixed genetic map distances. Data can sometimes indicate inflated map distances, and tools such as SimWalk2 can be used to identify genotypes that generate the double recombinants which inflate the map. Ancestry and allele frequencies Particularly with the creation of collaborative data sets, the issue of determining ancestry within samples is important. Ancestry can be estimated from genetic data using the STRUCTURE2 program [Pritchard et al., 2000a], or a mixture model based on principal component analyses, as described by Zhu et al. . The primary group within the sample can first be analyzed, with families of other ancestries subsequently tested for homogeneity of linkage. Heterogeneity can be evaluated specifically in linkage analysis by using one of several available techniques [Vieland, 1998; Devlin et al., 2002; Logue and Vieland, 2004; Wijsman and Yu, 2004]. Checking phenotypes: Distributions and effects of covariates Analysis methods are often sensitive to outliers and deviations from the normal distribution of quantitative traits being analyzed. Each trait should be tested for its distributional properties, and outliers should be investigated. If the data are non-normal, either a transformation should be applied, or internal corrections to the resulting scores should be used, as implemented in many of the current analysis programs. In addition, several of the traits of interest for autism may have strong effects of covariates that should be taken into account before genetic analysis. These effects may include age, sex, and medications, depending upon the trait in question. While some genetic analysis software has the capability of covariate adjustment, careful scrutiny of these effects should precede the final analysis. Linkage Analysis: Basic Methods and Refinements We will briefly review the general concepts of linkage. Linkage analysis is based on the recombination of DNA from parental gametes. When gametes form during meiosis, physical crossing over, breakage, and recombination of 27 chromosomes can occur, mixing up the parental genetic material. The variations in DNA sequence at a location (locus) on a chromosome are called alleles. If two loci are far enough apart, an odd or even number of crossovers will happen randomly between them and the recombinant and non-recombinant types of chromosomes will occur in the gametes with equal frequency. When this happens, the loci are unlinked, even though they may be on the same chromosome. In this case, the recombination frequency (theta) is ½, because half of the gametes have recombinant chromosomes and half have non-recombinant chromosomes. Recombination frequency is proportional to physical distance (map distance) on a chromosome, though correspondence between recombination and distance can be affected by different male and female recombination rates, recombination hot spots, and higher recombination rates at telomeres. Genes will be inherited together (linked) if they are close on the same chromosome because recombination within a smaller distance is less likely. In the case of linkage, recombinant chromosomes occur in the gametes less frequently (theta < ½ of the time) than non-recombinant chromosomes. The most basic method of genetic linkage involves assuming a trait is caused by one single gene in a family, then counting recombinations between each of many genetic markers and the trait. Fewer recombinations indicate the trait gene is more likely to be close to that marker. To quantify the degree of linkage in a family, one assumes various values of theta, then writes the probability of observing the marker and the trait in each family member given each degree of recombination. The joint probability of the whole family together given a particular assumption about recombination is called the likelihood. The likelihood is also computed assuming no linkage (theta ¼ ½). When each likelihood, assuming linkage (a value of theta < ½), is compared with the likelihood of the family assuming no linkage, the result is a likelihood ratio, sometimes called an odds ratio. The log of this odds ratio is the lod score, a 28 AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMIN. MED. GENET.): DOI 10.1002/ajmg.c measure of the likelihood of linkage versus no linkage. For any given family, one value of theta will give the highest odds ratio. If there is no linkage in the family, the highest odds ratio will be 1.0 (the log of this odds ratio ¼ lod score ¼ 0), and will occur when theta ¼ 1/2. Lod scores can be summed across multiple families to give a sample lod score. If some families within the sample are linked to a particular locus, while others are not, genetic heterogeneity has occurred. For a complex phenotype such as autism, genetic heterogeneity is likely. The use of intermediate phenotypes may help to classify subsets of families. In addition, lod scores can be computed taking into account estimates of heterogeneity. For a more complete discussion of genetic linkage, see Ott . Variance components linkage Quantitative phenotypes complicate linkage analysis. One approach to the analysis of these phenotypes is variance components linkage analysis, which assumes that a quantitative trait is influenced by genetic and environmental factors. The genetic factors may include specific loci in addition to polygenic influences. Tests for evidence for the specific loci compare the observed and expected covariance among the different classes of relatives in the family. The a priori correlation among two individuals in a family is the probability that they share a particular allele identical through a common ancestor (identical by descent, or IBD). This probability is called the coefficient of relationship, and is defined as twice their kinship coefficient. For example, the expected probability that full sibs share an allele IBD is ½. For first cousins, the probability is 1/ 8. The coefficient of relationship could be thought of as the average IBD probability. Given a particular map of genotyped markers, one can also compute a location-specific IBD probability at each location. Observed covariances among individuals are compared to expected covariances computed first assuming only the average IBD probability, then additionally with the pre- sence of a locus-specific component. The likelihood is computed when the additive genetic variance at a specific locus is fixed at zero, then compared to the likelihood at that location when the additive genetic variance is estimated. The log10 of the difference between these two likelihoods provides the lod score at that location. Methods are being developed to check for gene–gene interaction effects [e.g., Devlin et al., 2003]. Gene–gene interaction effects could be substantially larger than the main effects. Due to issues of multiple testing, tests should be confined to gene interactions that are either compelling biologically, have substantial main effects, or occur in excellent candidate genes. Methods are being developed to check for gene–gene interaction effects. Gene–gene interaction effects could be substantially larger than the main effects. Due to issues of multiple testing, tests should be confined to gene interactions that are either compelling biologically, have substantial main effects, or occur in excellent candidate genes. Linkage analysis of intermediate phenotypes Multivariate analysis may be helpful to investigate the possibility of common susceptibility genes underlying multiple intermediate phenotypes for autism. Preliminary cosegregation analysis should inform decisions about clustering phenotype data. This clustering could be done prior to genetic analysis using a number of standard clustering techniques. It may also be possible to specify an analysis of multiple quantitative phenotypes, such as BAP, repetitive behavior and language ability [Iturria and Blangero, 2000]. ARTICLE Linkage analysis of SNPs in pedigrees The explosion of SNP data in recent years presents a feast for the genetics researcher that is as yet difficult to digest. The critical analysis issue raised by the addition of the SNP data concerns the problem of linkage in the presence of possible linkage disequilibrium (LD) among markers within large pedigrees. Many consortium projects will have some combination of SNP data and microsatellite data, and will need to address this issue. Linkage analysis in extended pedigrees using markers in linkage disequilibrium will require some methodological development. While many current linkage analysis programs can handle arbitrary haplotype frequencies for a small number of loci, they cannot produce multipoint lod scores for a large number of markers. Current approaches usually involve suboptimal approximations, such as selecting subsets of loci that appear not to be associated. The problem is further complicated by the need to estimate the interlocus associations from the observed genotypes. One promising approach is the graphical modeling methods of Thomas and Camp . A graphical model [Lauritzen, 1996] describes the pattern of associations between loci using a Markov graph in which loci are represented by vertices with the property that given the states of its neighbors in the graph, a locus is conditionally independent of all other loci. Thomas and Camp  showed that such a graph can be estimated from a sample of reconstructed haplotypes by maximizing a penalized likelihood function. This work has recently been extended to handle diploid data [Thomas, 2005] using a two stage Markov chain Monte Carlo process that iterates between reconstructing haplotypes given the graphical model, and estimating the model given the reconstructed haplotypes. Thus, this process combines estimating a graphical model for linkage disequilibrium with haplotype reconstruction and haplotype frequency estimation. The method is currently being extended to handle family data by combining it with the Markov chain Monte Carlo linkage analysis ARTICLE AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMIN. MED. GENET.): DOI 10.1002/ajmg.c method developed by Thomas et al. . Verifying positive linkage findings The clusters of markers responsible for the linkage signal should be identified and checked for errors and mapping accuracy to determine the robustness of the linkage signal to markers. Allele frequencies should also be varied to assess sensitivity. If an analysis was done with a genetic model, these parameters should also be varied to see how sensitive the result is to the model assumptions. Different analysis methods and software should be used to determine if the result is dependent upon a quirk in the technique or analysis program. Finally, independent replication in a different sample, though difficult for complex traits, should be considered the gold standard of genetic linkage replication. Genetic Association The fundamental principle underlying genetic association is that of linkage disequilibrium (LD). LD is the nonrandom sharing alleles at close DNA markers across unrelated subjects within a population, as compared to allele sharing only within a family (linkage). The fundamental principle underlying genetic association is that of linkage disequilibrium (LD). LD is the non-random sharing alleles at close DNA markers across unrelated subjects within a population, as compared to allele sharing only within a family (linkage). LD can happen if two loci are very close, so that recombination across generations within a population has not yet occurred. If there is LD between a traitcausing gene and a marker, subjects with the trait in the population will share the same allele at the marker. Tests for genetic association look for significant differences in the frequency of an allele among unrelated subjects with a trait versus unrelated control subjects. A somewhat different type of molecular information can be obtained by using comparative genome hybridization (CGH), a microarray technique that will reveal indication of genome-wide chromosomal instability [see Oostlander et al., 2004 for review]. By assessing control populations, the frequency of specific deletions in cases can be compared to that in controls. In addition, the number of deletion/insertions found in autism subjects can be compared to the number found in the controls using a simple binomial test. While previous association studies of autism have been confined to categories of the clinical phenotype, future studies will have a number of other variables (e.g., language function, regressive onset, head circumference, blood assays) to explore. While previous association studies of autism have been confined to categories of the clinical phenotype, future studies will have a number of other variables (e.g., language function, regressive onset, head circumference, blood assays) to explore. For a candidate gene association study, the design is usually to screen the gene for variants using a subset of subjects, rather than to screen the entire sample. The extent of LD among the informative SNPs must then be estimated. Haplotype blocks (series of markers in LD) can be defined using software such as the HapBlockFinder program [Zhang et al., 2002]. The average haplotype block size has been estimated at about 35 kb [Daly et al., 2001; Gabriel et al., 2002]. Assuming an approximate gene size of 54 kb [Deloukas et al., 2001], an average gene 29 will be covered by about two haplotype blocks [Gabriel et al., 2002]. Once blocks are defined, one can then select two to three haplotype-tagging SNPs (htSNPs) per block. In the SNP typing phase, these htSNPs would then be typed in the entire sample. Case-control analysis methods To test for association between a trait and a single genetic variant, for example, a functional variant in a gene, chi-square methods can be used. Programs are available [e.g., Estimating Haplotypes; Xie and Ott, 1993; Zhao et al., 2000] to allow the test of multiple variants and haplotype blocks, and to specify the allele frequency of the putative traitcausing gene. These programs allow for simulation of the data given the analysis assumptions so that an empirical P-value can be reported. If a case-control design is to be used, controls should be carefully chosen. The finding of differences in allele frequencies between cases and controls can be due simply to population differences if the control sample is not well matched to the population of the case sample (admixture). Transmission disequilibrium tests Although case-control designs present advantages in ease of ascertainment, trio designs (affected subject and their two parents) avoid admixture, and may therefore be preferred. In the trio design, transmission disequilibrium is used to test for association [Terwilliger and Ott, 1992; Spielman et al., 1993]. This method counts the non-transmitted alleles from heterozygous parents as the control comparison. The classical TDT test [Spielman et al., 1993] has been extended to include covariates in a variety of frameworks: variance components/quantitative traits [Allison, 1997; Rabinowitz, 1997; Xiong et al., 1998; Abecasis et al., 2000]; extended families [Spielman and Ewens, 1998]; imprinting [Weinberg et al., 1998]; and X chromosome analysis [Ho and Bailey-Wilson, 2000]. Family-based association tests Tests of association using information from more extended relatives offer an 30 AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMIN. MED. GENET.): DOI 10.1002/ajmg.c interesting alternative to the classic casecontrol or trio designs. The Family Based Association Test (FBAT) and associated programs [Laird et al., 2000; Rabinowitz and Laird, 2000] can test for association between a qualitative or quantitative trait and genetic variants in a pedigree sample. The method employed by these programs tests for linkage as well as association, and again avoids false positive results due to admixture. The Family Based Association Test (FBAT) and associated programs can test for association between a qualitative or quantitative trait and genetic variants in a pedigree sample. The method employed by these programs tests for linkage as well as association, and again avoids false positive results due to admixture. functional change in linkage disequilibrium with the associated variant, further molecular studies can be designed to verify the finding. Such experiments may include knock out animal models or gene expression studies, but the methods involved in these experiments are beyond the scope of this paper. CONCLUSION The genetic analysis of autism stands poised at the threshold of discovery. Clinicians have turned their considerable talents toward the development of specific phenotype assessments that will perhaps identify genetically homogeneous subsets of subjects with autism. In addition, intermediate phenotypes associated with autism may allow the characterization of gene carriers within families, multiplying the information available to genetic research. Molecular information has exploded in recent years, offering an unprecedented level of detail for genetic analysis. Tools to dissect this information, while still in development, will facilitate the characterization of susceptibility genes for autism in the near future. ACKNOWLEDGMENTS FBAT is based on the original transmission disequilibrium test [Spielman et al., 1993], which quantifies the transmission of alleles of heterozygous parent to affected offspring. When true association exists, the transmission will deviate from the expected 50%. The FBAT test is an extension of this method that performs an adaptation of this test in pedigrees, rather than just parent-offspring pairs. FBAT also allows tests of different genetic models (additive, dominant, or recessive), and includes an option to test association with quantitative traits. The hbat command in FBATallows one to test for association with the haplotypes. Verifying association findings Positive association results should be verified in an independent sample. In addition, if the associated variant changes gene function, or if there is a This work was supported by 5 U19 HD035476, one of the NICHD Collaborative Programs of Excellence in Autism, by R01 MH069359, and by the Utah Autism Foundation. We thank those on our staff whose work have made this manuscript possible. REFERENCES Abecasis GR, Cardon LR, Cookson WO. 2000. A general test of association for quantitative traits in nuclear families. Am J Hum Genet 66:279–292. Abecasis G, Cherny S, Cookson W, Cardon L. 2002. Merlin-rapid analysis of dense genetic maps using sparse gene flow trees. Nat Genet 30:97–101. Abramson RK, Wright HH, Carpenter R, Brennan W, Lumpuy O, Cole E, Young SR. 1989. Elevated blood serotonin in autistic probands and their first-degree relatives. J Aut Dev Disord 19:397–407. Alarcon M, Cantor RM, Liu J, Gilliam C, Autism Genetic Research Exchange Consortium, Geschwind DH. 2002. Evidence for a language quantitative trait locus on chromosome 7q in multiplex autism families. Am J Hum Genet 70:60–71. ARTICLE Allison DB. 1997. Transmission disequilibrium tests for quantitative traits. Am J Hum Genet 60:676–690. Almasy L, Blangero J. 1998. Multipoint quantitative-trait linkage analysis in general pedigrees. Am J Hum Genet 62:1198–1211. American Psychiatric Association. 1994. Diagnostic and statistical manual of mental disorders. (4th edn.) Washington DC: American Psychiatric Association. Anderson GM. 2002. Genetics of childhood disorders: XLV. Autism, part 4: Serotonin and autism. J Am Acad Child Adolsec Psychiatry 41:1513–1516. Anderson GM, Freedman DX, Cohen DJ, Volkmar FR, Hoder EL, McPhedran P, Minderaa RB, Hansen CR, Young JG. 1987. Whole blood serotonin in autistic and normal subjects. J Child Psychol Psychiatry 28:885–900. Auranen M, Vanhala R, Varilo T, Ayers K, Kempas E, Ylisaukko-oja T, Sinsheimer JS, Peltonen L, Järvelä I. 2002. A genomewide screen for autism-spectrum disorders: Evidence for a major susceptibility locus on chromosome 3q25-27. Am J Hum Genet 71:777–790. Bailey A, Phillips W, Rutter M. 1996. Autism: Towards an integration of clinical, genetic, neuropsychological, and neurobiological perspectives. J Child Psychol Psychiatry 37:89–126. Bailey AP, Luthert P, Dean A, Harding B, Janota I, Montgomery M, Rutter M, Lantos P. 1998. A clinicopathological study of autism. Brain 121:889–905. Barrett S, Beck JC, Bernier R, Bisson E, Braun TA, Casavant TL, Childress D, Folstein SE, Garcia M, Gardiner MB, Gilman S, Haines JL, Hopkins K, Landa R, Meyer NH, Mullane JA, Nishimura DY, Palmer P, Piven J, Purdy J, Santangelo SL, Searby C, Sheffiled V, Singleton J, Slager S, Struchen T, Svenson S, Vieland V, Wang K, Winklosky B. 1999. An autosomal genomic screen for autism. Am J Med Genet 88:609–615. Bartlett CW, Flax JF, Logue MW, Vieland VJ, Bassett AS, Tallal P, Brzustowicz LM. 2002. A major susceptibility locus for specific language impairment is located on 13q21. Am J Hum Genet 71:45–55. Bertrand J, Mars A, Boyle C, Bove F, YearginAllsopp M, Decoufle P. 2001. Prevalence of autism in a United States population: The Brick Township, New Jersey, investigation. Pediatrics 108:1155–1161. Boehnke M, Langefeld CD. 1998. Genetic association mapping based on discordant sib pairs: The discordant-alleles test. Am J Hum Genet 62:950–961. Bolton P, Macdonald H, Pickles A, Rios P, Goode S, Crowson M, Bailey A, Rutter M. 1994. A case-control family history study of autism. J Child Psychol Psychiatry 35:877–900. Bradford Y, Haines J, Hutcheson H, Gardiner M, Braun T, Sheffield V, Cassavant T, Huang W, Wang K, Vieland V, Folstein S, Santangelo S, Piven J. 2001. Incorporating language phenotypes strengthens evidence of linkage to autism. Am J Med Genet 105:539–547. Buxbaum JD, Silverman JM, Smith CJ, Kilifarski M, Reichert J, Hollander E, Lawlor BA, Fitzgerald M, Greenberg DA, Davis KL. 2001. Evidence for a susceptibility gene for ARTICLE AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMIN. MED. GENET.): DOI 10.1002/ajmg.c autism on chromosome 2 and for genetic heterogeneity. Am J Hum Genet 68:1514– 1520. Buxbaum JD, Silverman J, Keddache M, Smith CJ, Hollander E, Ramoz N, Reichert JG. 2004. Linkage analysis for autism in a subset families with obsessive-compulsive behaviors: Evidence for an autism susceptibility gene on chromosome 1 and further support for susceptibility genes on chromosome 6 and 19. Mol Psychiatry 9:144–150. Chakrabarti S, Fombonne E. 2001. Pervasive developmental disorders in preschool children. JAMA 285:3093–3099. Clayton D. 1999. A generalization of the transmission/disequilibrium test for uncertain-haplotype transmission. Am J Hum Genet 65:1170–1177. Constantino JN, Przybeck T, Friesen D, Todd RD. 2000. Reciprocal social behavior in children with and without pervasive developmental disorders. J Dev Behav Pediatr 21:2–11. Constantino JN, Davis SA, Todd RD, Schindler MK, Gross MM, Brophy SL, Metzger LM, Shoushtari CS, Splinter R, Reich W. 2003. Validation of a brief quantitative measure of autistic traits: Comparison of the social responsiveness scale with the autism diagnostic interview-revised. J Aut Dev Disord 33:427–433. Cook EH, Leventhal BL. 1996. The serotonin system in autism. Curr Opin Pediatr 8:348– 354. Cook EHJr, Leventhal BL, Heller W, Metz J, Wainwright M, Freedman DX. 1990. Autistic children and their first-degree relatives: Relationships between serotonin and norepinephrine levels and intelligence. J Neuropsychiatry Clin Neurosci 2:268–274. Cook EHJr, Charak DA, Arida J, Spohn JA, Roizen NJ, Leventhal BL. 1994. Depressive and obsessive-compulsive symptoms in hyperserotonemic parents of children with autistic disorder. Psychiatry Res 52:25–33. Cook EH, Courchesne R, Lord C, Cox NJ, Yan S, Lincoln A, Haas R, Courchesne E, Levanthal BL. 1997. Evidence of linkage between the serotonin transporter and autistic disorder. Mol Psychiatry 2:247–250. Daly MJ, Rioux JD, Schaffner SF, Hudson TJ, Lander ES. 2001. High-resolution haplotype structure in the human genome. Nat Genet 29:229–232. Dawson G, Webb S, Schellenber GD, Dager S, Friedman S, Aylward E, Richards T. 2002. Defining the broader phenotype of autism: Genetic, brain, and behavioral perspectives. Dev Psychopathol 14:581–611. Deloukas P, Matthews LH, Ashurst J, Burton J, Gilbert JG, Jones M, Stavrides G, Almeida JP, Babbage AK, Bagguley CL, Bailey J, Barlow KF, Bates KN, Beard LM, Beare DM, Beasley OP, Bird CP, Blakey SE, Bridgeman AM, Brown AJ, Buck D, Burrill W, Butler AP, Carder C, Carter NP, Chapman JC, Clamp M, Clark G, Clark LN, Clark SY, Clee CM, Clegg S, Cobley VE, Collier RE, Connor R, Corby NR, Coulson A, Coville GJ, Deadman R, Dhami P, Dunn M, Ellington AG, Frankland JA, Fraser A, French L, Garner P, Grafham DV, Griffiths C, Griffiths MN, Gwilliam R, Hall RE, Hammond S, Harley JL, Heath PD, Ho S, Holden JL, Howden PJ, Huckle E, Hunt AR, Hunt SE, Jekosch K, Johnson CM, Johnson D, Kay MP, Kimberley AM, King A, Knights A, Laird GK, Lawlor S, Lehvaslaiho MH, Leversha M, Lloyd C, Lloyd DM, Lovell JD, Marsh VL, Martin SL, McConnachie LJ, McLay K, McMurray AA, Milne S, Mistry D, Moore MJ, Mullikin JC, Nickerson T, Oliver K, Parker A, Patel R, Pearce TA, Peck AI, Phillimore BJ, Prathalingam SR, Plumb RW, Ramsay H, Rice CM, Ross MT, Scott CE, Sehra HK, Shownkeen R, Sims S, Skuce CD, Smith ML, Soderlund C, Steward CA, Sulston JE, Swann M, Sycamore N, Taylor R, Tee L, Thomas DW, Thorpe A, Tracey A, Tromans AC, Vaudin M, Wall M, Wallis JM, Whitehead SL, Whittaker P, Willey DL, Williams L, Williams SA, Wilming L, Wray PW, Hubbard T, Durbin RM, Bentley DR, Beck S, Rogers J. 2001. The DNA sequence and comparative analysis of human chromosome 20. Nature 414:865–871. Devlin B, Roeder K. 1999. Genomic control for association studies. Biometrics 55:997–1004. Devlin B, Jones BL, Bacanu SA, Roeder K. 2002. Mixture models for linkage analysis of affected sibling pairs and covariates. Genet Epidemiol 22:52–65. Devlin D, Roeder K, Wasserman L. 2003. Analysis of multilocus models of association. Genet Epidemiol 25:36–47. Folstein SE, Piven J. 1991. Etiology of autism: Genetic influences. Pediatrics 87:767–773. Folstein SE, Rutter M. 1977. Infantile autism: A genetic study of 21 twin pairs. J Child Psychol Psychiatry 18:297–321. Folstein SE, Santangelo SL, Gilman SE, Piven J, Landa R, Lainhart J, Hein J, Wzorek M. 1999. Predictors of cognitive test patterns in autism families. J Child Psychol Psychiatry 40:1117–1128. Fombonne E. 1999. the epidemiology of autism: A review. Psychol Med 29:767–786. Gabriel SB, Schaffner SF, Nguyen H, Moore JM, Roy J, Blumenstiel B, Higgins J, DeFelice M, Lochner A, Faggart M, Liu-Cordero SN, Rotimi C, Adeyemo A, Cooper R, Ward R, Lander ES, Daly MJ. 2002. Altshuler D. The structure of haplotype blocks in the human genome. Science 296: 2225–2229. Gibbs RA, Belmont JW, Hardenbol P, Willis TD, Yu F, Yang H. 2003. The International HapMap Project. Nature 426:789–796. Gudbjartsson DF, Jonasson K, Frigge ML, Kong A. 2000. Allegro, a new computer program for multipoint linkage analysis. Nat Genet 25:12–13. Hasstedt S, Cartwright P. 1979. PAP: Pedigree Analysis Package, Computing and Techical Report 13. Salt Lake City: Department of Medical Biophysics and Computing, University of Utah. Hinds DA, Stuve LL, Nilsen GB, Halperin E, Eskin E, Ballinger DG, Frazer KA, Cox DR. 2005. Whole-genome patterns of common DNA variation in three human populations. Science 307:1072–1079. Ho GY, Bailey-Wilson JE. 2000. The transmission/disequilibrium test for linkage on the X chromosome. Am J Hum Genet 66:1158– 1160. Horvath S, Laird NM. 1998. A discordant-sibship test for disequilibrium and linkage: No need 31 for parental data. Am J Hum Genet 63: 1886–1897. IMGSAC. 1998. A full genome screen for autism with evidence for linkage to a region on chromosome 7q: International Molecular Genetic Study of Autism Consortium. Hum Mol Genet 7:571–578. IMGSAC. 2001. A genomewide screen for autism: Strong evidence for linkage to chromosomes 2q, 7q, and 16p. Am J Hum Genet 69:570–581. Iturria SJ, Blangero J. 2000. An EM algorithm for obtaining maximum likelihood estimates in the multi-phenotype variance components linkage model. Ann Hum Genet 64:349–362. Jones MB, Szatmari P. 2002. A risk-factor model of epistatic interaction, focusing on autism. Am J Med Genet 114:558–565. Kim S-J, Cox N, Courchesne R, Lord C, Corsello C, Akshoomoff N, Guter S, Leventhal BL, Courchesne E, Cook EH. 2002. Transmission disequilibrium mapping at the serotonin transporter gene (SLC6A4) region in autistic disorder. Mol Psychiatry 7:278–288. Klauck SM, Poustka F, Benner A, Lesch KP, Poustka A. 1997. Serotonin transporter (5HTT) gene variants associated with autism? Hum Mol Genet 6:2233–2238. Kuperman S, Beeghly JH, Burns TL, Tsai LY. 1985. Serotonin relationships of autistic probands and their first-degree relatives. J Am Acad Child Psychiatry 24:186–190. Lainhart JE, Piven J, Wzorek M, Landa R, Santangelo SL, Coon H, Folstein SE. 1997. Macrocephaly in children and adults with autism. J Am Acad Child Adolsec Psychiatry 36:282–290. Laird NM, Horvath S, Xu X. 2000. Implementing a unified approach to family based tests of association. Genet Epidemiol 19(suppl): 36–42. Lander ES, Green P. 1987. Construction of multilocus genetic linkage maps in humans. Proc Natl Acad Sci USA 84:2363–2367. Lauritzen SL. 1996. Graphical Models. Oxford: Clarendon Press. Leventhal BL, Cook EHJr, Morford M, Ravitz A, Freedman D. 1990. Relationships of whole blood serotonin and plasma norepinephrine within families. J Autism Dev Disord 20: 499–511. Liu JJ, Nyholt DR, Magnussen P, Parano E, Pavone P, Geschwind D, Lord C, Iversen P, Hoh J, Ott J, Gilliam TC. 2001. A genomewide screen for autism susceptibility loci. Am J Hum Genet 69:327–340. Logue MW, Vieland VJ. 2004. A new method for computing the multipoint posterior probability of linkage. Hum Hered 57:90–99. Lord C, Leventhal BL, Cook EHJr. 2001. Quantifying the phenotype in autism spectrum disorders. Am J Med Genet 105:36– 38. Marazziti D, Muratori F, Cesari A, Masala I, Baroni S, Giannaccini G, Dell’Osso L, Cosenza A, Pfanner P, Cassano GB. 2000. Increased density of the platelet serotonin transporter in autism. Pharmacopsychiatry 33:165–168. Martin ER, Monks SA, Warren LL, Kaplan NL. 2000. A test for linkage and association in general pedigrees: The pedigree disequilibrium test. Am J Hum Genet 67:146– 154. 32 AMERICAN JOURNAL OF MEDICAL GENETICS PART C (SEMIN. MED. GENET.): DOI 10.1002/ajmg.c McCauley JL, Li C, Jiang L, Olson LM, Crockett G, Gainer K, Folstein SE, Haines JL, Sutcliffe JS. 2005. Genome-wide and Ordered-Subset linkage analyses provide support for autism loci on 17q and 19p with evidence of phenotypic and interlocus genetic correlates. BMC Med Genet 6:1. McDougle CJ, Naylor ST, Cohen DJ, Aghajanian GK, Heninger GR, Price LH. 1996. Effects of tryptophan depletion in drug-free adults with autistic disorder. Arch Gen Psychiatry 53:993–1000. Muhle R, Trentacoste SV, Rapin I. 2004. The genetics of autism. Pediatrics 113:472–486. Newbury DF, Monaco AP. 2002. Molecular genetics of speech and language disorders. Curr Opin Pediatr 14:696–701. Oostlander AE, Meijer GA, Ylstra B. 2004. Microarray-based comparative genome hybridization and its applications in human genetics. Clin Genet 66:488–495. Ott J. 1999. Analysis of Human Genetic Linkage. Baltimore: Johns Hopkins University Press. Philippe A, Martinez M, Guilloudbataille M, Gillberg C, Rastam M, Sponheim E, Coleman M, Zappella M, Aschauer H, Van Maldergem L, Penet C, Feingold J, Brice A, Leboyer M. 1999. Genome-wide scan for autism susceptibility genes. Hum Mol Genet 8:805–812. Pickles A, Bolton P, MacDonald H, Bailey A, Le Couteur A, Sim CH, Rutter M. 1995. Latent-class analysis of recurrence risks for complex phenotypes with selection and measurement error: A twin and family history study of autism. Am J Hum Genet 57:717–726. Piven J, Chase GA, Landa R, Wzorek M, Gayle J, Cloud D, Folstein S. 1991. Psychiatric disorders in the parents of autistic individuals. J Am Acad Child Adolsec Psychiatry 30:471–478. Piven J, Palmer P, Jacobi D, Childress D, Arndt S. 1997a. Broader autism phenotype: Evidence from a family history study of multipleincidence autism families. Am J Psychiatry 154:185–190. Piven J, Palmer P, Landa R, Santangelo S, Jacobi D, Childress D. 1997b. Personality and language characteristics in parents from multiple-incidence autism families. Am J Med Genet 74:398–411. Plumet MH, Goldblum MC, Leboyer M. 1995. Verbal skills in relatives of autistic females. Cortex 31:723–733. Potenza MN, McDougle CJ. 1997. The Role of Serotonin in Autism-Spectrum Disorders. CNS Spectrums 2:25–42. Pritchard JK. 2001. Are rare variants responsible for susceptibility to complex diseases? Am J Hum Genet 69:124–137. Pritchard JK, Stephens M, Donnelly PJ. 2000a. Inference of population structure using multilocus genotype data. Genetics 155: 945–959. Pritchard JK, Stephens M, Rosenberg NA, Donnelly P. 2000b. Association mapping in structured populations. Am J Hum Genet 67:170–181. Rabinowitz D. 1997. A transmission disequilibrium test for quantitative trait loci. Hum Hered 47:342–350. Rabinowitz D, Laird NM. 2000. A unified approach to adjusting association tests for population admixture with arbitrary pedigree structure and arbitrary missing marker information. Hum Hered 504:227–233. Rapin I. 1996. Practitioner review: Developmental language disorders: A clinical update. J Child Psychol Psychiatry 37:643–655. Risch N, Spiker D, Lotspeich L, Nouri N, Hinds D, Hallmayer J, Kalaydjieva L, McCague P, Dimiceli S, Pitts T, Nguyen L, Yang J, Harper C, Thorpe D, Vermeer S, Young H, Hebert J, Lin A, Ferguson J, Chiotti C, Wiese-Slater S, Rogers T, Salmon B, Nicholas P, Myers RM. 1999. A genomic screen of autism: Evidence for a multilocus etiology. Am J Hum Genet 65:493–507. Ritchie MD, Hahn LW, Moore JH. 2003a. Power of multifactor dimensionality reduction for detecting gene-gene interactions in the presence of genotyping error, missing data, phenocopy, and genetic heterogeneity. Genet Epidemiol 24:150–157. Ritchie MD, White BC, Parker JS, Hahn LW, Moore JH. 2003b. Optimization of neural network architecture using genetic programming improves detection and modeling of gene-gene interactions in studies of human diseases. BMC Bioinformatics 4:28. Seltman H, Roeder K, Devlin B. 2003. Evolutionary-based association analysis using haplotype data. Genet Epidemiol 25:48–58. Shao YJ, Raiford KL, Wolpert CM, Cope HA, Ravan SA, Ashleykoch AA, Abramson RK, Wright HH, Delong RG, Gilbert JR, Cuccaro ML, Pericakvance MA. 2002. Phenotypic homogeneity provides increased support for linkage on chromosome 2 in autistic disorder. Am J Hum Genet 70: 1058–1061. Silverman JM, Smith CJ, Schmeidler J, Hollander E, Lawlor BA, Fitzgerald M, Buxbaum JD, Delaney K, Galvin P, Autism Genetic Research Exchange Consortium. 2002. Symptom domains in autism and related conditions: Evidence for familiality. Am J Med Genet 114:64–73. Smalley SL, Collins F. 1996. Brief report: Genetic, prenatal, and immunologic factors. J Autism Dev Disord 26:195–198. Spielman RS, Ewens WJ. 1998. A sibship test for linkage in the presence of association: The sib transmission/disequilibrium test. Am J HumGenet 62:450–458. Spielman RS, McGinnis RE, Ewens WJ. 1993. Transmission test for linkage disequilibrium: The insulin gene region and insulin-dependent diabetes mellitus (IDDM). Am J Hum Genet 52:506–516. Steffenburg S, Gillberg C, Hellgren L, Andersson L, Gillberg IC, Jakobsson G, Bohman M. 1989. A twin study of autism in Denmark, Finland, Iceland, Norway and Sweden. J Child Psychol Psychiatry 30:405–416. Szatmari P, Jones MB, Fisman S, Tuff L, Bartolucci G, Mahoney WJ, Bryson SE. 1995. Parents and collateral relatives of children with pervasive developmental disorders: A family history study. Am J Med Gen 60:60:282–289. Szatmari P, Jones MB, Zwaigenbaum L, MacLean JE. 1998. Genetics of autism: Overview and new directions. J Autism Dev Disord 28: 351–368. Terwilliger JD, Ott J. 1992. A haplotype-based ‘haplotype relative risk’ approach to detecting allelic associations. Hum Hered 42:337–346. ARTICLE Thomas A. 2005. Characterizing allelic associations from unphased diploid data by graphical modeling. Genet Epidemiol 29:23–35. Thomas A, Camp NJ. 2004. Graphical modeling of the joint distribution of alleles at associated loci. Am J Hum Genet 74:1088–1101. Thomas A, Gutin A, Abkevich V, Bansai A. 2000. Multipoint linkage analysis by blocked (Gibbs) sampling. Stats Comp 10:259–269. Tordjman S, Gutneckt L, Carlier M, Spitz E, Antoine C, Slama F, Cohen D, Ferrari P, Roubertoux P, Anderson G. 2001. Role of the serotonin transporter in the behavioral expression of autism. Mol Psychiatry 6: 434–439. Vieland VJ. 1998. Bayesian linkage analysis, or: How I learned to stop worrying and love the posterior probability of linkage. Am J Hum Genet 63:947–954. Wassink TH, Piven J. 2000. The molecular genetics of autism. Curr Psychiatry Rep 2:170–175. Weinberg CR, Wilcox AJ, Lie RT. 1998. A loglinear approach to case-parent-triad data: Assessing effects of disease genes that act either directly or through maternal effects and that may be subject to parental imprinting. Am J Hum Genet 62:969–978. Wijsman E, Yu D. 2004. Joint oligogenic segregation and linkage analysis using bayesian markov chain monte carlo methods. Mol Biotechnol 28:205–226. Wolk L, Giesen J. 2000. A phonological investigation of four siblings with childhood autism. J Commun Disord 33:371–389. Woodhouse W, Bailey A, Rutter M, Bolton P, Baird G, Le Couteur A. 1996. Head circumference in autism and other pervasive developmental disorders. J Child Psychol Psychiatry 37:665–671. Xie X, Ott J. 1993. Testing linkage disequilibrium between a disease gene and marker loci. Am J Hum Genet 53:1107 [abstr]. Xiong MM, Krushkal J, Boerwinkle E. 1998. TDT statistics for mapping quantitative trait loci. Ann Hum Genet 62:431–452. Yirmiya N, Pilowsky T, Nemanov L, Arbelle S, Feinsilver T, Fried I, Ebstein RP. 2001. Evidence for an association with the serotonin transporter promoter region polymorphism and autism. Am J Med Genet 105:381–386. Ylisaukkooja T, NieminenvonWendt T, Kempas E, Sarenius S, Varilo T, vonWendt L, Peltonen L, Jarvela I. 2004. Genome-wide scan for loci of Asperger syndrome. Mol Psychiatry 9:161–168. Yonan AL, Alarcon M, Cheng R, Magnusson PKE, Spence SJ, Palmer AA, Grunn A, Juo SHH, Terwilliger JD, Liu JJ, Cantor RM, Geschwind DH, Gilliam TC. 2003. A genomewide screen of 345 families for autism-susceptibility loci. Am J Hum Genet 73:886–897. Zhang K, Deng M, Chen T, Waterman MS, Sun F. 2002. A dynamic programming algorithm for haplotype block partitioning. Proc Natl Acad Sci USA 99:7335–7339. Zhao JH, Curtis D, Sham PC. 2000. Model-free analysis and permutation test for allelic associations. Hum Hered 50:133–139. Zhu X, Zhang S, Zhao H, Cooper RS. 2002. Association mapping, using a mixture model for complex traits. Genet Epidemiol 23:181–196.