Recent positive selection of a human androgen receptor/ectodysplasin A2 receptor haplotype and its relationship to male pattern baldness
- First Online:
- 398 Downloads
Genetic variants in the human androgen receptor gene (AR) are associated with male pattern baldness (androgenetic alopecia, AGA) in Europeans. Previous observations of long-range linkage disequilibrium at the AR locus are consistent with the hypothesis of recent positive selection. Here, we further investigate this signature and its relationship to the AGA risk haplotype. The haplotype homozygosity suggests that the AGA risk haplotype was driven to high frequency by positive selection in Europeans although a low meiotic recombination rate contributed to the high haplotype homozygosity. Further, we find high levels of population differentiation as measured by FST and a series of fixed derived alleles along an extended region centromeric to AR in the Asian HapMap sample. The predominant AGA risk haplotype also carries the putatively functional variant 57K in the flanking ectodysplasin A2 receptor gene (EDA2R). It is therefore probable that the AGA risk haplotype rose to high frequency in combination with this EDA2R variant, possibly by hitchhiking on a positively selected 57K haplotype.
Androgens mediate a wide range of developmental and physiological responses through the androgen receptor (AR), and play a crucial role in several stages of male development (Lee and Chang 2003). Mutations in the androgen receptor gene (AR) may cause androgen insensitivity syndrome (Ris-Stalpers et al. 1990) and spinobulbar muscular atrophy (La Spada et al. 1991) and are found in prostate and breast cancer tissues (reviewed by Gottlieb et al. 2004; Heinlein and Chang 2004; Weiss et al. 2005). Various studies have tested for association between variability in AR and a number of phenotypes including male infertility, breast cancer, and prostate cancer, but the results have been inconsistent (reviewed by Rajender et al. 2007). We recently found genetic variability in a 1 Mb region within and centomeric to AR to be associated with androgenetic alopecia (AGA) at a highly significant level (Hillmer et al. 2005, 2008a). The large size of the associated linkage disequilibrium (LD) block and the high frequency of the derived haplotype were suggestive of recent positive selection in Europeans (Hillmer et al. 2005). Since AR is linked to the male reproductive system, it is conceivable that variability in AR might have an impact on reproductive success and therefore be subject to natural selection. Consistent with this hypothesis are more recent studies which have reported significant signals of positive selection at the AR locus (Altshuler et al. 2005; Voight et al. 2006), although the major signals were located closer to the centromere.
In the present study, we have further analyzed the signature of positive selection in and around the AR locus and its relationship to the AGA susceptibility variants there. Our data demonstrate that the AR locus is a unique X-chromosomal region with striking differences between human populations. Interestingly, it is the AGA risk haplotype that shows evidence for positive selection in the European population. The AGA risk haplotype also carries a derived non-synonymous allele (57K) in the flanking ectodysplasin A2 receptor gene (EDA2R). The 57K allele may have been the target of positive selection in East Asians: it shows large allele frequency differences between populations (Sabeti et al. 2007), lies in a likely functional domain of the EDA2R protein, and the ancestral allele, 57R, is conserved from human to rat. Thus, the AGA risk haplotype may have hitchhiked to high frequency in Europeans as a result of positive selection on the linked 57K allele in EDA2R.
Materials and methods
Haplotype bifurcation and extended haplotype homozygosity (EHH)
Phased haplotype data of the non-pseudoautosomal region of the X chromosome of CEU (90 X chromosomes), CHB + JPT (134 X chromosomes), and YRI (90 X chromosomes) were downloaded from the International HapMap project home page (http://www.hapmap.org/, release 16) (Altshuler et al. 2005; Gibbs et al. 2003). The European Americans (CEU) are referred to as “European”, and the Yoruba (YRI) from Nigeria are referred to as “African”. The Han Chinese and Japanese (CHB + JPT) data were pooled and are referred to as “East Asian”. SNPs polymorphic in the respective populations were selected. Haplotype bifurcation, EHH, and relative EHH (REHH) were analyzed using Sweep software v1.1 (Sabeti et al. 2002), which was downloaded from (http://www.broad.mit.edu/mpg/sweep). The longest non-overlapping haplotype blocks with between 3 and 20 SNPs were determined using the block definition of Gabriel et al. (2002). The haplotype block which overlapped with the 5′ end of AR was defined as the core for bifurcation, EHH, and REHH analysis. The EHH approach proposed by Sabeti et al. (2002) compares rates of LD decay between haplotypes. The REHH corrects for local variation in the recombination rate by comparing the EHHs of different core haplotypes present at a locus. Only core haplotypes with frequencies >0.05 were selected for presentation. To determine the significance of EHH, the EHHs within a distance of 0.5 cM of the selected core were compared to the EHHs within a distance of 0.5 cM of all other cores of the non-pseudoautosomal X chromosome.
Haplotype block structure
Genotype data, one Mb upstream and downstream from the AR start site (chrX:65,546,894…67,546,894, build 35) of CEU, CHB + JPT, and YRI were downloaded from the International HapMap project home page (http://www.hapmap.org/, release 21a) (Altshuler et al. 2005) and haplotype structure was analyzed with Haploview software v4.0 (Barrett et al. 2005) using the definition of Gabriel et al. (2002).
Derived allele frequencies
The phased haplotype data from 1 Mb upstream and downstream of AR of those SNPs, which were polymorphic in at least one population were downloaded (HapMap release 21a), and chimpanzee alleles were defined as ancestral alleles (Mikkelsen et al. 2005). In addition, macaque alleles were used to identify the ancestral alleles and SNPs with discrepant predictions from chimpanzee and macaque were excluded.
Genotype data (build 35) were downloaded from the HapMap project home page (release 20). Only unrelated individuals were included: 60 European American, 45 Han Chinese, 45 Japanese, and 60 Yoruba. SNPs of the pseudoautosomal region of the X chromosome were not considered. A weighted average FST was calculated according to equation (10) of Weir and Cockerham (1984) for 759 non-overlapping 200 kb windows across the X chromosome. The windows were designed such that AR was centered inside a single window. For each FST comparison, P values were defined as the proportion of 200 kb windows with FST values equal to or greater than the FST value in the window containing AR. Similarly, the 95 and 99% thresholds represent the FST values above which 5 and 1% of the windows’ FST values lie, respectively. We define the value of the test statistic as “significant” if it lies within the most extreme 5% of the empirical distribution of that test statistic. We acknowledge that our P value only represents the position of a value relative to the empirical distribution and that this is not a statistical “significance test”.
Haplotype specific LD differed considerably between populations. In Africans, the diversity of haplotypes was higher and the bifurcations more complex than in the European population (Supplementary Fig. 1). Afr-H3 (which is closest to Eur-H1) has a frequency of only 12% and an EHH of 0.65 (P value = 2.3 × 10−4) centromeric to the core (Supplementary Fig. 1). However, the centromeric REHH of Afr-H3 is 5.5 (P value = 0.13). The variability centromeric to AR is markedly lower in the East Asian population compared to the European and African populations. Therefore, haplotype comparisons between the populations are not straightforward, and none of the East Asian core SNPs has been shown to be associated with AGA. Due to low variability within and centromeric to AR, bifurcations of the haplotypes toward the centromere are low (Supplementary Fig. 2). In the East Asian population, EHH for all three high-frequency haplotypes in both directions of the core have P values < 0.05, but the respective REHH values were not unusually high.
Thus, the EHH analysis shows that it is the AGA risk haplotype that displays extended haplotype homozygosity from AR toward EDA2R, but it does not provide a significant REHH signal.
Differentiation of common SNPs between the three major continental populations
To understand more clearly the overall LD structure at the AR locus, we compared HapMap genotype data of the three major continental populations. The LD and haplotype block structure differ markedly across the populations. Centromeric of AR is a region of high LD in the European population (Supplementary Fig. 3b). This region includes EDA2R which is located more than 900 kb centromeric to AR. The LD in this region is lower in the African sample, in agreement with previous studies which have consistently found lower LD in African than in non-African populations (Hinds et al. 2005; Altshuler et al. 2005) (Supplementary Fig. 3c). In contrast, only a few SNPs between EDA2R and AR are polymorphic in the East Asian population, and the distant polymorphic SNPs are in low LD (Supplementary Fig. 3d). This is in accordance with the low haplotype bifurcation in the East Asian population. Again, the most striking differences between the investigated populations were found centromeric to AR.
The observation of low variability in the East Asian population prompted us to investigate the distribution of ancestral and derived alleles in East Asia. Comparison of the derived allele frequencies of Europeans, Africans, and East Asians revealed remarkable differences. Most SNPs in and between EDA2R and AR which are polymorphic in at least one of the three populations were observed to be fixed for the derived allele in the East Asian sample (Supplementary Fig. 4), consistent with a recent selective sweep in East Asians at this locus.
Family based recombination rate at the AR locus
Although the earlier-reported haplotype homozygosity at the AR locus shows a pattern consistent with the influence of positive selection, it is important to note that the influence of recombination rate variation on extended haplotype homozygosity is still being explored (O’Reilly et al. 2008). Such research has pointed out that the existing methods for identifying loci with signatures of recent positive selection are prone to biases towards regions with low recombination rates (Myles et al. 2008; Nielsen et al. 2007). What is more, an increased population recombination rate is associated with those functional categories of genes that show faster and accelerated sequence evolution between species, an observation which led to the hypothesis that evolution could favor recombination at loci subject to stronger directional selection (Frazer et al. 2007; Freudenberg et al. 2007). For these reasons, it is important to consider the local recombination rate when hypothesizing about a recent history of non-neutral evolution at a genetic locus. A low recombination rate can be expected for the AR locus, since recombination is known to be lower towards the centromere, and the X chromosome only recombines in females. According to the Decode genetic map (Kong et al. 2002), the female recombination rate for the interval between the EDA2R and the AR (65.6–66.6 Mb on the NCBI 35 assembly) is 0.756 cM/Mb, which is below average. In order to interpret the possible influence of the local recombination rate on LD in a population, the female recombination rate must be translated into the sex-averaged rate (0.378 cM/Mb). This value is clearly at the lower end of the distribution of recombination rates for all 1 Mb windows across the human genome (Supplementary Fig. 5). A low recombination rate is thus one factor that is likely to have contributed to the extent of LD and haplotype homozygosity at the AR locus. The lack of formal significance for our REHH analysis, which corrects for local variation in the recombination rate, suggests that the high EHH might be influenced by the low recombination rate at this locus. Thus, low recombination rate certainly has contributed to the strong LD from EDA2R to AR, which applies in either the presence or the absence of recent positive selection.
The present study was initiated in order to understand the putative signatures of recent positive selection at the human AR/EDA2R locus and their relationship to haplotypes that have been shown to be the major genetic determinant of AGA (Ellis et al. 2001, 2007; Hayes et al. 2005; Hillmer et al. 2005, 2008a; Richards et al. 2008). In the European population, the AGA-associated alleles are on the most frequent haplotype Eur-H1, which showed an unusually high EHH toward EDA2R. Although REHH was not extreme, the levels of EHH and REHH are clearly higher compared to the levels of the less frequent haplotypes in the European population. The high haplotype specific LD, which is indicative of a young haplotype with a high frequency, suggests a rapid increase in the frequency of Eur-H1.
The chromosomal region extending from AR to EDA2R displays several unusual features: (1) The interval of more than 1 Mb displays high LD in the European population. It has been previously shown that the average pairwise LD of the AR locus was the highest of all X-chromosomal windows with the same recombination rate (Hillmer et al. 2005). (2) The LD of the AR/EDA2R locus differs between the three continental populations, with comparatively low LD observed in Africans. (3) In East Asians, genetic variability in this region is remarkably low due to fixation of the derived allele of most SNPs. Since a locus that has experienced positive selection will subsequently show low levels of genetic variation (Kim and Stephan 2002) and an excess of high-frequency derived variants (Fay and Wu 2000), the pattern of variation at the AR/EDA2R locus in East Asia may be best explained by a selective sweep. When those SNPs which are non-polymorphic in East Asia are included in an LD analysis, the region displays high LD and the pattern is more similar to that of Europeans.
In order to explain the pattern of genetic variability observed at the AR locus, genetic drift needs to be considered as the null hypothesis. In fact, population range expansions can lead to the geographic spread of alleles and thereby to strong differences in allele frequencies between continents (Hofer et al. 2009; Klopfstein et al. 2006). It has been reported that East Asians generally have fewer rare derived alleles than Europeans, which has been interpreted as an indication of increased genetic drift in East Asians (Keinan et al. 2007). Furthermore, significantly larger FST values have been observed between East Asian and Africans than between Europeans and Africans, with a more pronounced bottleneck in East Asians being the probable cause (Keinan et al. 2007). Compared to the pattern of the autosomes, the X chromosome shows a stronger genetic drift in North Europeans and East Asians and both populations harbor more high-frequency derived alleles on the X chromosome than expected (Keinan et al. 2009).
Association of rs1385699 and AR core haplotype SNPs with AGA and haplotype frequencies
198 German affected versus 157 German unaffected (Hillmer et al. 2005)
Association P with AGA
9.41 × 10−5
3.17 × 10−11
2.56 × 10−6
OR for AGA
OR (95% CI)
AGA risk allele
296 German affected versus 347 German controls (Hillmer et al. 2008b)
Association P with AGA
1.6 × 10−4
4.61 × 10−6
9.05 × 10−9
OR for AGA
OR (95% CI)
AGA risk allele
All four recently reported genome-wide studies have shown that the AR/EDA2R locus is the major AGA-susceptibility locus in the European population (Hillmer et al. 2008a, 2008b; Prodi et al. 2008; Richards et al. 2008). An interesting question thus arises from the observation that AGA prevalence differs between populations. Previous work has shown that AGA is about half as frequent in Asians as in Europeans (Hamilton 1951; Norwood 1975; Paik et al. 2001; Pathomvanich et al. 2002, Tang et al. 2000). Epidemiological data for African populations is sparse, but it has been reported that males without AGA are four times more frequent in Africa than in Europe (Setty 1970). The European AGA risk alleles, however, are fixed in the East Asian sample (including 57K) and their frequencies therefore do not correlate with the AGA frequencies. As AGA might be regarded as a secondary sexual characteristic, rather than a disorder, it is possible that the hair loss itself was under enhanced sexual selection by identifying the older male leader comparable to the silver-backed gorilla (Randall 2007). The reported lower prevalence of AGA in Africans might then be explained by the importance of scalp hair for the protection against the tropical sun which outweighed the enhanced sexual selection (Randall 2007). Since the causative variant has not yet been identified, the lower prevalence of AGA in Asia and presumably in Africa might indicate a higher frequency of the functional AGA allele in the European population, either due to its origin on a Europe-specific background or differences in population demography. Alleles of 21 SNPs between AR and EDA2R are more frequent in Europeans than in East Asians and Africans (Supplementary Table 1). Three of these have been tested for association with AGA in a German sample (Hillmer et al. 2008a) and the AGA risk allele of one SNP (rs471205) is more frequent in Europeans than in East Asians and Africans. Although the association of rs471205 was only moderate and a functional relevance for AGA development, therefore, seems unlikely, the other SNPs with this frequency pattern are good candidates for further investigation. Since the HapMap data does not cover all SNPs in this region, the presence of more SNPs showing the respective population differentiation is possible. Alternatively, it may be that the penetrance of the mutation is modified by genetic background or environmental factors. Genetic studies of AGA in non-European populations are warranted in order to address this question.
Importantly, the markers with the strongest population differences (as measured by average FST values in the European/African and East Asian/African comparison) are found in the region extending from AR toward the centromere, including EDA2R. It is tempting to speculate that the proposed positive selection on 57K (Sabeti et al. 2007) is a driving force behind the population haplotype distribution centromeric to AR and that hitchhiking of the AGA susceptibility mutation on the 57K background accounts for the high prevalence of the phenotype. In either case, it seems safe to say that the mutation which confers risk for AGA in the European population rose to high frequency in combination with the derived allele 57K in EDA2R. 57K being the causative variant for AGA might be regarded as less likely as this allele is fixed in East Asia whereby the prevalence of AGA is lower in this population compared to Europeans. As the signature of positive selection is less profound in Europeans compared to East Asians, but AGA prevalence is higher in Europeans, it seems likely that the AGA causative variant and the cause for positive selection are distinct from each other. Resequencing of the AGA risk haplotype of the AR/EDA2R locus in different populations and subsequent association analysis are needed to clarify whether a Europe specific mutation accounts for the higher prevalence of AGA in this population.
This work was supported by the European Research Grant for Research on Alopecia funded by Pfizer Consumer Healthcare, the Deutsche Forschungsgemeinschaft (FOR 423), and the Alfried Krupp von Bohlen und Halbach-Stiftung. We thank Christine Schmael for revision of the manuscript. The work by S.M., D.A.H., K.T., and M.S. was funded by the Max Plank Society.
This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.
- Frazer KA, Ballinger DG, Cox DR, Hinds DA, Stuve LL, Gibbs RA, Belmont JW, Boudreau A, Hardenbol P, Leal SM, Pasternak S, Wheeler DA, Willis TD, Yu F, Yang H, Zeng C, Gao Y, Hu H, Hu W, Li C, Lin W, Liu S, Pan H, Tang X, Wang J, Wang W, Yu J, Zhang B, Zhang Q, Zhao H, Zhou J, Gabriel SB, Barry R, Blumenstiel B, Camargo A, Defelice M, Faggart M, Goyette M, Gupta S, Moore J, Nguyen H, Onofrio RC, Parkin M, Roy J, Stahl E, Winchester E, Ziaugra L, Altshuler D, Shen Y, Yao Z, Huang W, Chu X, He Y, Jin L, Liu Y, Sun W, Wang H, Wang Y, Xiong X, Xu L, Waye MM, Tsui SK, Xue H, Wong JT, Galver LM, Fan JB, Gunderson K, Murray SS, Oliphant AR, Chee MS, Montpetit A, Chagnon F, Ferretti V, Leboeuf M, Olivier JF, Phillips MS, Roumy S, Sallee C, Verner A, Hudson TJ, Kwok PY, Cai D, Koboldt DC, Miller RD, Pawlikowska L, Taillon-Miller P, Xiao M, Tsui LC, Mak W, Song YQ, Tam PK, Nakamura Y, Kawaguchi T, Kitamoto T, Morizono T, Nagashima A, Ohnishi Y, Sekine A, Tanaka T, Tsunoda T et al (2007) A second generation human haplotype map of over 3.1 million SNPs. Nature 449:851–861PubMedCrossRefGoogle Scholar
- Fujimoto A, Kimura R, Ohashi J, Omi K, Yuliwulandari R, Batubara L, Mustofa MS, Samakkarn U, Settheetham-Ishida W, Ishida T, Morishita Y, Furusawa T, Nakazawa M, Ohtsuka R, Tokunaga K (2008) A scan for genetic determinants of human hair morphology: EDAR is associated with Asian hair thickness. Hum Mol Genet 17:835–843PubMedCrossRefGoogle Scholar
- 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, Altshuler D (2002) The structure of haplotype blocks in the human genome. Science 296:2225–2229PubMedCrossRefGoogle Scholar
- Gibbs RA, Belmont JW, Hardenbol P, Willis TD, Yu F, Yang H, Ch’ang LY, Huang W, Liu B, Shen Y, Tam PKH, Tsui LC, Waye MMY, Wong JTF, Zeng C, Zhang Q, Chee MS, Galver LM, Kruglyak S, Murray SS, Oliphant AR, Montpetit A, Hudson TJ, Chagnon F, Ferretti V, Leboeuf M, Phillips MS, Verner A, Kwok PY, Duan S, Lind DL, Miller RD, Rice JP, Saccone NL, Taillon-Miller P, Xiao M, Nakamura Y, Sekine A, Sorimachi K, Tanaka T, Tanaka Y, Tsunoda T, Yoshino E, Bentley DR, Deloukas P, Hunt S, Powell D, Altshuler D, Gabriel SB, Zhang H, Zeng C, Matsuda I, Fukushima Y, Macer DR, Suda E, Rotimi CN, Adebamowo CA, Aniagwu T, Marshall PA, Matthew O, Nkwodimmah C, Royal CDM, Leppert MF, Dixon M, Stein LD, Cunningham F, Kanani A, Thorisson GA, Chakravarti A, Chen PE, Cutler DJ, Kashuk CS, Donnelly P, Marchini J, McVean GAT, Myers SR, Cardon LR, Abecasis GR, Morris A, Weir BS, Mullikin JC, Sherry ST, Feolo M, Altshuler D, Daly MJ, Schaffner SF, Qiu R, Kent A, Dunston GM, Kato K, Niikawa N, Knoppers BM, Foster MW, Clayton EW, Wang VO, Watkin J, Gibbs RA, Belmont JW, Sodergren E, Weinstock GM et al (2003) The international HapMap project. Nature 426:789–796CrossRefGoogle Scholar
- Hayes VM, Severi G, Eggleton SA, Padilla EJ, Southey MC, Sutherland RL, Hopper JL, Giles GG (2005) The E211 G > A androgen receptor polymorphism is associated with a decreased risk of metastatic prostate cancer and androgenetic alopecia. Cancer Epidemiol Biomarkers Prev 14:993–996PubMedCrossRefGoogle Scholar
- Hillmer AM, Hanneken S, Ritzmann S, Becker T, Freudenberg J, Brockschmidt FF, Flaquer A, Freudenberg-Hua Y, Jamra RA, Metzen C, Heyn U, Schweiger N, Betz RC, Blaumeiser B, Hampe J, Schreiber S, Schulze TG, Hennies HC, Schumacher J, Propping P, Ruzicka T, Cichon S, Wienker TF, Kruse R, Nöthen MM (2005) Genetic variation in the human androgen receptor gene is the major determinant of common early-onset androgenetic alopecia. Am J Hum Genet 77:140–148PubMedCrossRefGoogle Scholar
- Hillmer AM, Brockschmidt FF, Hanneken S, Eigelshoven S, Steffens M, Flaquer A, Herms S, Becker T, Kortüm AK, Nyholt DR, Zhao ZZ, Montgomery GW, Martin NG, Mühleisen TW, Alblas MA, Moebus S, Jöckel KH, Bröcker-Preuss M, Erbel R, Reinartz R, Betz RC, Cichon S, Propping P, Baur MP, Wienker TF, Kruse R, Nöthen MM (2008a) Susceptibility variants for male-pattern baldness on chromosome 20p11. Nat Genet 40:1279–1281PubMedCrossRefGoogle Scholar
- Hillmer AM, Flaquer A, Hanneken S, Eigelshoven S, Kortüm AK, Brockschmidt FF, Golla A, Metzen C, Thiele H, Kolberg S, Reinartz R, Betz RC, Ruzicka T, Hennies HC, Kruse R, Nöthen MM (2008b) Genome-wide scan and fine-mapping linkage study of androgenetic alopecia reveals a locus on chromosome 3q26. Am J Hum Genet 82:737–743PubMedCrossRefGoogle Scholar
- Kong A, Gudbjartsson DF, Sainz J, Jonsdottir GM, Gudjonsson SA, Richardsson B, Sigurdardottir S, Barnard J, Hallbeck B, Masson G, Shlien A, Palsson ST, Frigge ML, Thorgeirsson TE, Gulcher JR, Stefansson K (2002) A high-resolution recombination map of the human genome. Nat Genet 31:241–247PubMedGoogle Scholar
- Mikkelsen TS, Hillier LW, Eichler EE, Zody MC, Jaffe DB, Yang SP, Enard W, Hellmann I, Lindblad-Toh K, Altheide TK, Archidiacono N, Bork P, Butler J, Chang JL, Cheng Z, Chinwalla AT, de Jong P, Delehaunty KD, Fronick CC, Fulton LL, Gilad Y, Glusman G, Gnerre S, Graves TA, Hayakawa T, Hayden KE, Huang X, Ji H, Kent WJ, King MC, Kulbokas EJIII, Lee MK, Liu G, Lopez-Otin C, Makova KD, Man O, Mardis ER, Mauceli E, Miner TL, Nash WE, Nelson JO, Pääbo S, Patterson NJ, Pohl CS, Pollard KS, Prüfer K, Puente XS, Reich D, Rocchi M, Rosenbloom K, Ruvolo M, Richter DJ, Schaffner SF, Smit AFA, Smith SM, Suyama M, Taylor J, Torrents D, Tuzun E, Varki A, Velasco G, Ventura M, Wallis JW, Wendl MC, Wilson RK, Lander ES, Waterston RH (2005) Initial sequence of the chimpanzee genome and comparison with the human genome. Nature 437:69–87CrossRefGoogle Scholar
- Richards JB, Yuan X, Geller F, Waterworth D, Bataille V, Glass D, Song K, Waeber G, Vollenweider P, Aben KK, Kiemeney LA, Walters B, Soranzo N, Thorsteinsdottir U, Kong A, Rafnar T, Deloukas P, Sulem P, Stefansson H, Stefansson K, Spector TD, Mooser V (2008) Male-pattern baldness susceptibility locus at 20p11. Nat Genet 40:1282–1284PubMedCrossRefGoogle Scholar
- Ris-Stalpers C, Kuiper GG, Faber PW, Schweikert HU, van Rooij HC, Zegers ND, Hodgins MB, Degenhart HJ, Trapman J, Brinkmann AO (1990) Aberrant splicing of androgen receptor mRNA results in synthesis of a nonfunctional receptor protein in a patient with androgen insensitivity. Proc Natl Acad Sci USA 87:7866–7870PubMedCrossRefGoogle Scholar
- Sabeti PC, Reich DE, Higgins JM, Levine HZ, Richter DJ, Schaffner SF, Gabriel SB, Platko JV, Patterson NJ, McDonald GJ, Ackerman HC, Campbell SJ, Altshuler D, Cooper R, Kwiatkowski D, Ward R, Lander ES (2002) Detecting recent positive selection in the human genome from haplotype structure. Nature 419:832–837PubMedCrossRefGoogle Scholar
- Sabeti PC, Varilly P, Fry B, Lohmueller J, Hostetter E, Cotsapas C, Xie X, Byrne EH, McCarroll SA, Gaudet R, Schaffner SF, Lander ES, Frazer KA, Ballinger DG, Cox DR, Hinds DA, Stuve LL, Gibbs RA, Belmont JW, Boudreau A, Hardenbol P, Leal SM, Pasternak S, Wheeler DA, Willis TD, Yu F, Yang H, Zeng C, Gao Y, Hu H, Hu W, Li C, Lin W, Liu S, Pan H, Tang X, Wang J, Wang W, Yu J, Zhang B, Zhang Q, Zhao H, Zhou J, Gabriel SB, Barry R, Blumenstiel B, Camargo A, Defelice M, Faggart M, Goyette M, Gupta S, Moore J, Nguyen H, Onofrio RC, Parkin M, Roy J, Stahl E, Winchester E, Ziaugra L, Altshuler D, Shen Y, Yao Z, Huang W, Chu X, He Y, Jin L, Liu Y, Sun W, Wang H, Wang Y, Xiong X, Xu L, Waye MM, Tsui SK, Xue H, Wong JT, Galver LM, Fan JB, Gunderson K, Murray SS, Oliphant AR, Chee MS, Montpetit A, Chagnon F, Ferretti V, Leboeuf M, Olivier JF, Phillips MS, Roumy S, Sallee C, Verner A, Hudson TJ, Kwok PY, Cai D, Koboldt DC, Miller RD, Pawlikowska L, Taillon-Miller P, Xiao M, Tsui LC et al (2007) Genome-wide detection and characterization of positive selection in human populations. Nature 449:913–918PubMedCrossRefGoogle Scholar