Ancestral and recombinant 16-locus HLA haplotypes in the Hutterites

Immunogenetics (1999) 49: 491 497 Q Springer-Verlag 1999 ORIGINAL PAPER Lowell R. Weitkamp 7 Carole Ober Ancestral and recombinant 16-locus HLA haplotypes in the Hutterites Received: 1 May 1998 / Revised: 2 December 1998 Abstract Prior studies in the Schmiedeleut Hutterites of South Dakota have demonstrated associations between human leukocyte antigen (HLA) haplotype matching and fetal loss (Ober et al. 1992) and mate preferences (Ober et al. 1997), as well as deficiencies of homozygotes for HLA haplotypes (Kostyu et al. 1993). These studies were based on the serologically-defined five-locus HLA-A, -C, -B, -DR, -DQ haplotype. To further elucidate the effects of specific major histocompatibility (MHC) loci or regions on fetal loss and mate choice, we genotyped a sample of Hutterites for 14 MHC loci by DNA or biochemical methods. Typing for additional loci in the HLA-A to HLA-DPB1 region increased the number of recognized Hutterite MHC haplotypes to 67, and further localized the site of crossover in 9 of 15 recombinant haplotypes. Hutterite MHC haplotype sequences are similar to those observed in outbred Caucasians, suggesting that the influence of HLA haplotypes on fetal loss and mating structure may be general. Key words HLA 7 Ancestral haplotypes 7 Recombinant haplotypes 7 Hutterites Introduction L.R. Weitkamp Department of Psychiatry and Division of Genetics, University of Rochester Medical Center, Rochester, NY 14642, USA C. Ober (Y) Department of Human Genetics, The University of Chicago, 924 East 57th Street, Chicago, IL 60637, USA E-mail: carole6genetics.uchicago.edu, Tel: c1-773-8340735, Fax: c1-773-8340505 The Schmiedeleut (S-leut) subdivision of the Hutterites, a Caucasian religious isolate that migrated from Europe to the northern United States and Canada in the 1870s, has been the subject of studies of the relationship between genes in the MHC and reproductive performance (Hauck and Ober 1991; Ober et al. 1983, 1985, 1988, 1992, 1998). The Hutterites are one of the most fertile human populations, having stable marriages, relatively short interbirth intervals, large completed family sizes (medianp8), and relatively few (2%) childless couples (Ober 1995; Sheps 1965). Their communal lifestyle ensures that all Hutterites are exposed to similar environments, including a traditional diet, proscriptions on smoking and birth control measures, and only occasional alcohol consumption. Extensive genealogical records indicate moderate remote inbreeding. Although first-cousin marriages are very rare, the average kinship coefficient of mates is 0.0369 (Ober et al. 1992); the S-leut can trace their ancestry to only 68 founders (Mange 1964; Martin 1970). The small number of founders, who may themselves have been related, suggested that the Hutterites would have limited genetic variability as compared with outbred Caucasian populations especially for highly polymorphic loci or haplotypes. Indeed, in a sample of 852 S-leut adults who were serologically-typed for five HLA loci, HLA-A, -B, -C, -DR, and -DQ, there were only 48 unique ancestral and 11 contemporary recombinant haplotypes defined through pedigree analysis (Kostyu et al. 1993). Four additional recombinant haplotypes were found in a larger sample of 1082 Hutterites that included children (Dawson et al. 1995). The small number of HLA haplotypes indicates that the Hutterites, as compared with outbred populations, have an increased opportunity to choose a mate with an MHC haplotype identical at all loci between HLA-A and HLA-DQ by virtue of the haplotypes being identical-by-descent (IBD). Thus, this population provides an unusual opportunity to assess the effect of matching for MHC genes and haplotypes on the choice of a mate and of MHC mating types on the reproductive performance of couples. Based on serologically-defined five-locus haplotypes, we reported decreased fertility among spouses matching for HLA antigens (Ober et al. 1992), fewer than expected spouses matching for a ha-

492 plotype (Ober et al. 1997), and deficiencies of homozygotes (Kostyu et al. 1993). Here we report a re-analysis of the MHC haplotype structure of the Hutterite population based on (1) molecular typing of three of the five loci previously typed serologically (HLA-C, HLA-DR, HLA-DQ) and (2) typing for eleven additional loci in the HLA region. Importantly, three of the Hutterite five-locus HLA haplotypes were split by allelic differences within the HLA-A to HLA-DQ region; two more were split by typing HLA loci centromeric to this region. We report the complete description of the 16-locus composition of 52 ancestral and 15 recombinant MHC haplotypes in the S-leut Hutterites and note that the MHC sequences common in the Hutterites represent approximately half of the MHC sequences common in outbred Caucasians. The relationships between the Hutterite high-resolution haplotypes and fetal loss (Ober et al. 1998), mate choice (Ober et al. 1999), and homozygous deficiencies (Robertson et al. 1999) are described elsewhere. Materials and methods The Hutterite population and sample composition A description of the Hutterite population is provided in Ober and co-workers (1997). HLA types were determined by serology for a total of 1343 persons. From this group, 962 HLA-typed, married persons were used as the foundation for constructing a genealogy. The genealogy contained 1891 members and up to 12 generations, ending with 62 progenitors who were born in the early 1700s to early 1800s. Within the genealogy, 180 persons who had not been HLA-typed by serology were assigned deduced HLA haplotypes based on family relationships. Genetic typing HLA haplotypes were initially defined by five serologically-typed loci, HLA-A, -B, -C, -DR, and -DQ (Kostyu et al. 1989, 1993), resulting in 48 ancestral and 15 contemporary recombinant haplotypes (Dawson et al. 1995). In order to re-evaluate these results, we typed 14 HLA region loci in 80 or more Hutterites who had been previously typed by serological methods. HLA-C serological types were re-evaluated and extended by typing with sequence-specific oligonucleotide probes (SSOPs) and by sequencing(bunce and Welsh 1994; Bunce et al. 1994). DR serological types were re-evaluated and extended by typing for DRB1 by sequencing (Tilanus and Eliaou 1995); allele assignments were made by SBTyper (Pharmacia Biotech, Piscataway, N.J.). DQ serological types were re-evaluated and extended by typing for DQA1 and DQB1. DQA1 was typed by restriction fragment length polymorphism (RFLP) and sequence-specific polymerase chain reaction (SS-PCR) (Inoko and Masao 1993; Olerup et al. 1993); DQB1 was typed by sequencing (Tilanus and Eliaou 1995) with allele assignments using SBTyper. Five new loci between HLA-A and HLA-DQ were typed according to the following procedures: HLA-E (Grimsley and Ober 1997), TNFa (Udalova et al. 1993), properdin factor B (Alper et al. 1972), and complement components C4A and C4B (Zhang et al. 1988). In addition, the MHC haplotype was extended telomeric by sequencing HLA-G alleles (Ober 1997; Ober et al. 1996) and centromeric by typing LMP2 and LMP7 using RFLPs (Deng et al. 1995), TAP1 using microsatellites (Carrington and Dean 1994), and DPB1 by sequencing, with allele assignments by SBTyper, and by SSOPs (Bugawan et al. 1990; Tilanus and Eliaou 1995). Blood samples were not available for all of the Hutterites who had been previously typed serologically. Whenever possible, at least three distantly related individuals with the same five-locus haplotype were typed for the DNA-based markers, at least six persons for plasma complement components, C4A and C4B, and most persons for DQA1 and BF. When samples were available, individuals with a recombinant haplotype were typed for as many loci as possible. An attempt was also made to type all persons with a haplotype that had been split and who were not known by pedigree relationships to have one or the other of the splits. In cases where this was not possible the pre-split designation for the haplotype was retained. Results Allelic composition of haplotypes A map of the 16 MHC loci is shown in Fig. 1. The numbers of alleles at each locus observed in the Hutterites are shown in Table 1. The 52 different ancestral haplotypes (designated by a 4-digit haplotype number beginning with 1) that were in fact defined by alleles at these Fig. 1 Map of the MHC on chromosome 6p21 showing the 16 loci used to define haplotypes in this study. HLA loci are shown below the line and non-hla loci are shown above the line. Modified from Ober (1998)

493 Table 1 The number of alleles at 16 MHC loci in the Hutterites Locus G 6 A 12 E 2 C 13 B 15 TNFa 11 Bf 3 C4A 7 C4B 5 DRB1 14 DQA1 7 DQB1 9 LMP2 2 TAP1 4 LMP7 2 DBP1 9 No. of alleles 16 loci are shown in Fig. 2, ordered according to complement and then DR-DQ haplotypes. Shaded areas indicate regions of allelic identity. Two of the previously described five-locus haplotypes have been eliminated (1010 and 1033) and a new haplotype (1080) was observed. The designation of haplotypes 1010 (A2, Cx, B50, DR3, DQ2) and 1045 (A2, C6, B50, DR3, DQ2) as separate was, as expected (Dawson et al. 1995), incorrect; individuals with both haplotypes were typed as HLA-C * 0602. Thus, the five individuals initially assigned 1010 based on absent reactivity with weak antisera for Cw6 have been reassigned haplotype 1045. The designation of haplotypes 1033 (A26, Cx, B38, DR1, DQ1) and 1034 (A26, Cx, B38, DR2, DQ1) as distinct haplotypes (Dawson et al. 1995; Kostyu et al. 1993) was in error because individuals with both haplotypes in fact had the DRB1 allele 15011. Fig. 2 Allelic composition of 16-locus ancestral haplotypes in the Hutterites. Haplotype frequencies were based on the 962 HLA-typed, married persons in the genealogy. Haplotypes are ordered by Bf to DQB1 region genes. Regions of identity have been shaded for haplotypes that were identical from Bf through DQB1. Serologically defined haplotypes 1007, 1016, and 1020 (shown on the bottom of the figure), were split into two distinct haplotypes on the basis of alleles at one or more loci. For a few individuals in our sample with these haplotypes, the split haplotype could not be determined. In these cases, the pre-split haplotype was retained (see text for details). The allele HLA- G * 01014 was officially assigned by the WHO Nomenclature Committee in February 1998

494 Accordingly the 1033 haplotype was eliminated and the eight individuals originally assigned this haplotype were reassigned the 1034 haplotype. Individuals with the new 1080 ancestral haplotype (A2, C3, B60, DR1, DQ1) were originally typed as having the 1014 haplotype (A2, C3, B60, DR2, DQ1). Typing of DRB1 in these individuals showed the allele 0101, corresponding to DR1; alleles at seven other loci (cf. Fig. 2) confirm the 1080 haplotype to be different from 1014. Thus, the total number of five-locus ancestral haplotypes is now 47 rather than 48. Three HLA-A to HLA-DQ haplotypes were split by typing complement C4B (1007 split into 1084 and 1085), HLA-C, C4A, C4B, and HLA-DRB1 (1016 split into 1082 and 1083), and TNFa, C4B, and HLA-DRB1 (1074 split into 1086 and 1087). Each of these splits was confirmed by the results of typing the most centromeric of the MHC loci, HLA-DPB1. Two additional haplotypes (1002 split into 1090 and 1091; 1020 split into 1088 and 1089) were identical at fifteen of the sixteen loci but had different alleles at HLA-DPB1. The frequency of each haplotype in the 962 married-person cohort is shown in Fig. 2. The frequencies shown for 1007 (0.003), 1016 (0.009), and 1020 (0.001) are for individuals whose split type is unknown because no sample was available for typing and that individual could not be shown by pedigree relationship to have a haplotype IBD with either defined split. Prior to splitting, 1007, 1016, and 1020 were the three most common haplotypes (frequencies of 0.068, 0.068, and 0.075, respectively, from Fig. 2; see also Dawson et al. 1995). The frequency of the splits of 1016 were about equal: 0.030 for 1082 and 0.029 for 1083. The frequencies of the splits for 1007 were 0.052 for 1084 and 0.013 for 1085; for 1020 the frequencies of the splits were 0.018 for 1088 and 0.056 for 1089. The haplotypes 1002 and 1074 were uncommon and consequently their splits were rare (^0.006). The 15 recombinant haplotypes (designated by a haplotype number beginning with 2) are listed in Fig. 3. Only recombinants within the HLA-A to HLA-DQ region would have been detectable, since typing for loci centromeric to HLA-DQ was limited to a subsample of individuals (see Materials and methods). Furthermore, recombinants within the HLA-A to HLA-DQ region would only have been recognized if the recombination generated haplotypes with new allelic configurations. There were two changes to the list of 15 recombinant haplotypes published by Dawson and co-workers (1995). The 2017 recombinant haplotype (A2-Cw3-B62- DR1-DQ1), found in two persons (Kostyu et al. 1993), was determined, based on HLA-C molecular typing, to be 0401 and therefore was identical to their parental 1017 ancestral haplotype (A2-Cw4-B62-DR1-DQ1). Haplotype 2030, a recombinant of 1041 and 1012, was uncertain at the time of earlier studies and therefore not reported. This recombinant haplotype has now been observed in a daughter of the proband and confirmed by TNFa, DRB1, and DQA1 allelic differences between 1012 and 1041. Thus, the number of different recombinant chromosomes remains 15. Seven of the recombinants (2001, 2003, 2005, 2010, 2011, 2016, and 2030) were observed in more than one related person. All of these recombinants were confirmed by typing for additional loci on either side of the breakpoint (Fig. 3). Although eight of the recombinants have each been observed in only one person, three of these (2002, 2014, and 2022) were confirmed by typing for additional loci on either side of the breakpoint. One (2007) was confirmed by complement typing (on one side of the breakpoint) and by more than one serologically-typed locus on the other side of the breakpoint. The other four were previously declared recombinant based solely on the presence or absence of a single HLA-A antigen (without the necessity of recognizing the presence or absence of an allelic antigen). We regard these recombinants (2004, absence of A2; 2006, presence of A26; 2012, presence of A31; 2013, presence of A2) as provisional. Eight of the 15 recombinations occurred between HLA-A and HLA-B, and, of these, two were demonstrated to have occurred between HLA-A and HLA-E (2001 and 2003), four between HLA-A and HLA-C (2004, 2011, 2012, and 2016), and two between HLA-A and HLA-B (2006 and 2013). Of the seven recombinations between HLA-B and HLA-DR, two occurred between HLA-B and TNFa (2002 and 2005), one between TNFa and BF (2010), one between HLA-B and BF (2007), one between TNFa and C4A (2014), one between C4A and HLA-DR (2022), and one between TNFa and HLA-DR (2030). Four of the recombinants (2002, 2012, 2014, and 2022) were found in persons not included in the 1891 member genealogy (i.e., they were in unmarried children of persons in the genealogy). Thus, at least eight and perhaps as many as eleven recombinants were observed in the contemporary part of the genealogy (among the 1142 persons with known or inferred genotypes). Haplotype structure of the genealogy There have only been 15 surnames in the Hutterites since the early 1700s, suggesting that the Hutterite population had 15 male founders. There may have been a few more female founders, since it was not uncommon for females to die before the male reproductive period was completed. However, some female founders may have been related to each other, and indeed some male founders may have been related to each other or to female founders through female lines. From this historical information we previously estimated that the number of unrelated HLA haplotypes in the 62 progenitors to be fewer than 80, perhaps about 60 (Ober et al. 1997). On the assumption that no recombination within the tightly linked complement complex (BF, C4A, C4B) oc-

495 Fig. 3 Fifteen recombinant MHC haplotypes in the Hutterites. Asterisks denote recombinants that are considered provisional (see text for details) curred in 687 meioses that linked the progenitors to Hutterites with known or inferred MHC types, 16 different complement haplotypes would have been introduced by the 62 progenitors. For the same reason, the 21 different DR-DQ haplotypes are also likely to have been introduced by the progenitors. Recombination between the complement loci and the DR-DQ loci is infrequent. Among the seven known HLA-B/HLA-DR recombinants, six were localized with respect to C4A: five were telomeric and only one was centromeric to C4A. Thus, most and possibly all of the 34 different ancestral haplotypes defined by alleles at loci from BF through HLA-DQB1 were likely to have been introduced by the 62 progenitors. Indeed, there are only two ancestral haplotypes that are identical from HLA-G through C4B and different at HLA-DRB1 (1031 and 1032), creating only one possibility for a C4/DR recombinant among the ancestral haplotypes. Eighteen ancestral haplotypes shared BF/HLA- DQB1 haplotypes with one of 14 other ancestral haplotypes. The alleles in common for each of the 14 groups of ancestral haplotypes are highlighted in Fig. 2. Haplotypes 1088 and 1089 may have arisen from each other by a recombination telomeric to HLA-DPB1 or through gene conversion. Based on the relative frequencies of 1088 and 1089 and the infrequency of DPB1 * 0402 among other haplotypes, 1088 is very likely the derived haplotype. On the other hand, 1017 and 1089 are both likely to have been progenitor haplotypes since there is no other source among the 52 an-

496 cestral haplotypes of the HLA-G to HLA-B segment of either 1017 or 1089. Haplotype 1090 may have been derived from 1091, because at the only locus in which they differ, HLA-DPB1, 1090 has an allele that is much less frequent in this population than the allele in 1091. Haplotype 1043, although identical to 1090 from HLA-B through DPB1, is likely to have been introduced by a progenitor, since the HLA-G through HLA-C segment in 1043 is unique in the Hutterite pedigree. There were four pairs of haplotypes that were identical for the segment from TNFa through DPB1 (1037 1045; 1019 1080; 1038 1018; 1014 1034; Fig. 2). Both members of the first pair were frequent (0.018; 0.019) and had haplotypes for HLA-G through HLA-B that were not found in the other 52 ancestral haplotypes. They are both likely to have been introduced by progenitors. Similarly, the 1019 haplotype was frequent (0.035) and also has a unique HLA-G through HLA-B haplotype. However, 1080 was a rare haplotype (0.001) that was identical from HLA-G through HLA-B with 1014 (frequency 0.027). It may well have arisen by a recent recombination between 1019 and 1014. Both 1038 and 1018 (frequencies of 0.008 and 0.011) had unique HLA-G through HLA-B haplotypes. Haplotype 1034 (frequency 0.055) shared the HLA-G through HLA-B segment with haplotype 1072 (frequency 0.001). The latter haplotype, although rare, was the only ancestral haplotype with the complement C4A * 2 allele. Thus, it was present in a progenitor at least as a partial haplotype. Three pairs of haplotypes were identical for the segment from BF through DPB1 (1039 1012; 1082 1042; 1023 1026). All six of these haplotypes had unique HLA-G through TNFa haplotypes. The rare haplotype 1076 (frequency 0.001) was not typed for all loci; it may be identical with the uncommon 1073 haplotype (frequency 0.004) from HLA-G through LMP7. Either could have been derived from the other by recombination near DPB1. Other ancestral haplotypes could have been derived from progenitor haplotypes by more than one recombination in the course of 12 generations and 687 meioses. Four candidates for recombination within the ancestral part of the pedigree (1088, 1090, 1080, and 1076) correspond reasonably well in terms of recombination rate with the observation of 8 to 11 recombinants among members of the pedigree with known haplotypes. Discussion Molecular typing of MHC loci in a Hutterite cohort previously typed by serological methods resulted in changes in five locus (HLA-A, B, C, DR, DQ) haplotypes. One new low frequency haplotype was recognized and two low frequency haplotypes were eliminated. More importantly, three of the 47 five-locus ancestral MHC haplotypes were split by typing for six additional loci within the HLA-A to HLA-DQ region. Two of the split haplotypes, 1007 and 1016, were among the most frequent Hutterite five-locus haplotypes. Two of the 50 HLA-A to HLA-DQ ancestral haplotypes were further distinguished by typing for HLA- DPB1, a locus centromeric to HLA-DQ. Fifteen recombinant haplotypes were observed in the contemporary Hutterite population. The localization of the breakpoint was refined for nine of the 15 recombinant haplotypes. In two of the eight HLA-A/ HLA-B recombinants, the breakpoint occurred between HLA-A and HLA-E (2001, 2003). Of the seven recombinations that had been observed between HLA- B and HLA-DR, five occurred between HLA-B and C4A (2002, 2005, 2007, 2010, 2014), one between C4A and HLA-DR (2022), and one between TNFa and HLA-DR (2030). These data further support the observations of others that recombination between the complement loci and the HLA-DR/HLA-DQ loci is rare (Degli-Esposti et al. 1992), and allow extrapolations about the number of unique ancestral haplotypes present in the Hutterite population and the relationship between these haplotypes and those observed in outbred Caucasian populations. Despite the enormous number of different haplotypes theoretically possible in outbred populations, some haplotypes are much more frequent than expected from the product of the frequencies of the alleles at each locus. Degli-Esposti et al. (Degli-Esposti and co-workers (1992) found that 21 different HLA-B to HLA-DR haplotypes accounted for 133 of 348 (38%) of the haplotypes in an Australian Caucasian population. The same HLA-B to C4 and C4 to HLA- DR segments present in these 133 haplotypes were found in another 41 and 81 haplotypes, respectively. Thus 255/348 (73%) of the haplotypes in an outbred population could have been derived by a single recombination from one of 21 haplotype sequences. Most of the Hutterite MHC haplotype sequences were the same as those reported for this outbred Caucasian population. Ten of the 21 Australian HLA-B to HLA-DR haplotype sequences were found in 17 of the 50 Hutterite haplotypes, accounting for 32% of the haplotypes in the Hutterite genealogy. Another 21 Hutterite haplotypes had HLA-B to C4 or C4 to HLA-DR segments identical to those in the above 17 haplotypes, accounting for an additional 34% of the Hutterite haplotypes. Thus, roughly half of the MHC sequences common in outbred Caucasians were the common sequences in the Hutterites. The limited number of haplotypes in the Hutterites may permit the detection of haplotype-specific effects on fertility and mating structure(ober et al. 1992, 1997, 1998, 1999; Weitkamp and Ober 1998) but the effects of MHC haplotypes on mate choice and fertility in the Hutterites are not due to the presence of unusual haplotypes in this population. Rather, the HLA haplotype effect on fertility and mate choice may apply to outbred populations, either currently or historically.

497 Acknowledgments The authors acknowledge Dr. Ann Begovich for reagents provided for HLA-DPB1 typing by SSOPs, and Sally Guttormsen, Chris Billstrand, Xi-ling He, Barbara Rosinsky, Andrea Robertson, Carrie Aldrich, and Amy Eklund for technical assistance. This work was supported by PHS grant HD21244. References Alper CA, Boenish T, Watson L (1972) Genetic polymorphism in human glycine-rich beta-glycoprotein. J Exp Med 135: 68 80 Bugawan TL, Begovich AB, Erlich HA (1990) Rapid HLA-DPB typing using enzymatically amplified DNA and nonradioactive sequence-specific oligonucleotide probes. Immunogenetics 32: 231 241 Bunce M, Barnardo MDNM, Welsh KI (1994) Improvements in HLA-C typing using sequence-specific primers (PCR-SSP) including definition of HLA-Cw9 and Cw10 and a new allele HLA- Cwy/8v. Tissue Antigens 44 : 200 203 Bunce M, Welsh KI (1994) Rapid DNA typing for HLA-C using sequence-specific primers (PCR-SSP): Identification of serological and non-serologically defined HLA-C alleles including several new alleles. Tissue Antigens 43: 7 17 Carrington M, Dean M (1994) A polymorphic dinucleotide repeat in the third intron of TAP1. Hum Mol Genet 3 :218 Dawson DV, Ober C, Kostyu DD (1995) Extended HLA profile of an inbred isolate: the Schmiedeleut Hutterites of South Dakota. Genet Epidemiol 12 : 47 62 Degli-Esposti MA, Leaver AL, Christiansen FT, Witt CS, Abraham LJ, Dawkins RL (1992) Ancestral haplotypes: conserved population MHC haplotypes. Hum Immunol 34: 242 252 Deng GY, Muir A, Maclaren NK, She J-X (1995) Association of LMP2 and LMP7 genes within the major histocompatibility complex with insulin-dependent diabetes mellitus: Population and family studies. Am J Hum Genet 56:528 534 Grimsley C, Ober C (1997) Population genetic studies of HLA-E: Evidence for selection. Human Immunol 52: 33 40 Hauck WW, Ober C (1991) Statistical analysis of outcomes from repeated pregnancies: effects of HLA sharing on fetal loss. Genet Epidemiol 8: 187 197 Inoko H, Masao O (1993) PCR-RFLP. In: Hui KM, Bidwell JL (eds) Handbook of HLA Typing Techniques, pp 1 8, CRC Press, Ann Arbor Kostyu DD, Dawson DV, Elias S, Ober C (1993) Deficit of homozygotes in a Caucasian isolate. Hum Immunol 37 : 135 142 Kostyu DD, Ober CL, Dawson DV, Ghanayem M, Elias S, Martin AO (1989) Genetic analysis of HLA in the U.S. Schmiedeleut Hutterites. Am J Hum Genet 45 :261 269 Mange AP (1964) Growth and inbreeding of a human isolate. Hum Biol 36: 104 133 Martin AO (1970) The founder effect in a human isolate: Evolutionary implications. Am J Phys Anthropol 32: 351 368 Ober C (1995) HLA and reproduction: lessons from studies in Hutterites. Placenta 16: 569 577 Ober C (1997) HLA-G alleles in Hutterites (Corrigendum). J Reprod Immunol 33:89 90 Ober C, Elias S, Kostyu DD, Hauck WW (1992) Decreased fecundability in Hutterite couples sharing HLA-DR. Am J Hum Genet 50: 6 14 Ober C, Elias S, O Brien E, Kostyu DD, Hauck WW, Bombard A (1988) HLA sharing and fertility in Hutterite couples: evidence for prenatal selection against compatible fetuses. Am J Reprod Immunol Microb 18: 111 115 Ober C, Hauck WW, Kostyu DD, O Brien E, Elias S, Simpson JL, Martin AO (1985) Adverse effects of HLA-DR sharing on fertility: a cohort study in a human isolate. Fertil Steril 44: 227 232 Ober C, Hyslop T, Elias S, Weitkamp LR, Hauck WW (1998) HLA matching and fetal loss: Results of a 10-year prospective study. Hum Reprod 13: 101 106 Ober C, Martin AO, Simpson JL, Hauck WW, Amos DB, Kostyu DD, Fotino M, Allen FH (1983) Shared HLA antigens and reproductive performance in the Hutterites. Am J Hum Genetics 35: 990 1004 Ober C, Rosinsky B, Grimsley C, van der Ven K, Robertson A, Runge A (1996) Population genetics studies of HLA-G: allele frequencies and linkage disequilibrium with HLA-A. J Reprod Immunol 32: 111 123 Ober C, Weitkamp LR, Cox N (1999) HLA and Mate Choice. In: Johnston R, Müller-Schwarz D, Sorensen P (eds) Chemical Signals in Vertebrates 8, Plenum Press, New York Ober C, Weitkamp LR, Cox N, Dytch H, Kostyu D, Elias S (1997) HLA and mate choice in humans. Am J Hum Genet 61: 497 504 Olerup O, Aldener A, Fogdell A (1993) HLA-DQB1 and -DQA1 typing by PCR amplification with sequence-specific primers (PCR-SSP) in 2 h. Tissue Antigens 41: 119 134 Robertson A, Charlesworth D, Ober C (1999) The effect of inbreeding avoidance on Hardy-Weinberg equilibrium: examples of neutral and selected loci. Genet Epid, in press Sheps MC (1965) An analysis of reproductive patterns in an American isolate. Popul Stud 19: 65 80 Tilanus MGJ, Eliaou JF (1995) Technical Handbook, Twelfth International Histocompatibility Workshop and Conference. Utrect Udalova IA, Nedospasov SA, Webb GC, Chaplin DD, Turetskaya RL (1993) Highly informative typing of the human TNF locus using six adjacent polymorphic markers. Genomics 16: 180 186 Weitkamp LR, Ober C (1998) HLA and mate choice. Am J Hum Genet 62: 986 987 Zhang WJ, Kay PH, Cobain TW, Dawkins RL (1988) C4 allotyping on plasma or serum: application to routine laboratories. Hum Immunol 21:165 171