Please download to get full document.

View again

All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
  In the large genetic survey ‘Provinces Françaises’ the recombination fractions in the HLA system have been estimated by a family analysis programme (FAP). A total of 1332 families were analysed and in general the findings were in agreement with
  European Journal of Immunogenetics (1994), 21,3343 RECOMBINATION FRACTIONS IN THE HLA SYSTEM BASED ON THE DATA SET ‘PROVINCES FRANCAISES’: INDICATIONS OF HAPLOTYPE- SPECIFIC RECOMBINATION RATES M. THOMSEN,* . NEUGEBAUER,+ J ARNAUD,* N OROT,’ A. SEVIN,‘ M. AUR+ & A. CAMBON-THOMSEN*.* *INSERM U. 100, CHU Purpan, Toulouse, France, +lnstitut ur medizinkche Statistik, Dokumentation und Datenverarbeitung der Universitat Bonn, Bonn, Germany, and *CR PG, CNRS UPR 8291, CHU Purpan, Toulouse, France (Received 30 August 1993; revised 17 September 1993; accepted 28 September 1993) SUMMARY In the large genetic survey ‘Provinces Francaises’ the recombination fractions in the HLA system have been estimated by a family analysis programme FAP). A total of 1332 families were analysed and in general the findings were in agreement with recombination fractions reported previously. The maternal recombination rates were on average 1.8 times higher than the corresponding ones for males. The comparison of the recombination fractions with the corresponding physical distances suggests the existence of hot spots of recombination. The analysis did not show deviations from expected values for HLA-A and B alleles on HLA-A/B recombinant haplotypes. However, analysis of HLA-B/DR recombinant haplotypes showed a skewed distribu- tion of B and DR alleles. The significance of the findings is difficult to evaluate as all results are estimated numbers and frequencies but a manual analysis of the recombinant families confirmed the observations. HLA-B/DR recombinant haplo- types carried often HLA-DR3 and DRll whereas DR2 and DR7 were more rarely present on recombinant haplotypes. DR4 had an increased incidence on BF/DR recombinant haplotypes but not on A/B or B/BF recombinant haplotypes. Some of the haplotypes with the strongest linkage disequilibria as Al 88, DR3 and A3 B7, DR2 seem to be less frequently involved in recombinations than other haplotypes. Variations of recombination rates depending on certain alleles or haplotypes might partially explain the conservation of some haplotypes or part of haplotypes in Caucasoids. INTRODUCTION The genetic survey ‘Provinces Francaises’ PF) is a collaborative study where a large panel of genetic markers of nearly 1400 families from different regions of France and from Quebec have Correspondence: M. Thomsen, INSERM U 100, Allte H Serres, CHU Purpan, 31052 Toulouse Cedex, France. 33  34 M. homsen et al. been determined. A considerable number of laboratories have participated and a central database has been established in Toulouse (Ohayon & Cambon-Thomsen, 1986; Cambon-Thom- sen & Ohayon, 1988; Cambon-Thomsen et al., 1989). The selection of the families was based upon their srcin (more than three generations in the region studied), their willingness to give blood and the presence of at least two children. As most of the regions included were rural without history of recent admixture of foreign populations, the results cannot be extrapolated directly to the present population in France. One of the most informative marker systems is the major histocompatibility complex (MHC), in humans the HLA system, where the PF study has shown striking differences in allele frequencies between the regions studied. On the basis of family phenotype data, HLA haplotypes have been constructed. However, several solutions of the haplotype composition may exist when homozygosity or sharing of alleles in the parents occur. The probability that a given solution is correct depends upon the allele frequencies and the a-priori assumptions of linkage disequilibria between alleles in a given population. Recombinations within the HLA region are useful for mapping the relative position of the polymorphic loci. Exact estimates of the recombination fractions are difficult to obtain because a large sample is required and when pooling families from different sources it is difficult to avoid bias. The PF study is less biased because the families were selected without prior knowledge of cross-overs and thus more valid calculations of recombination fractions are possible. A substantial number of intra-HLA recombination events has been found and preliminary analyses on this material have been presented previously (Thomsen et al., 1986, 1989a). In this study, different assumptions were made on the strength of linkage disequilibria in estimating the recombination fractions on maternal and paternal haplotypes and an attempt was made to evaluate whether certain alleles or haplotypes might be implicated in recombination events more often than others. MATERIALS AND METHODS The selection of families, blood drawings and typing for most of the genetic markers were done in each of the collaborating laboratories (Ohayon & Cambon-Thomsen, 1986). Each of the families included the parents and a minimum of two children and the family should have been in the region for at least three generations. The families were typed according to standardized methods and with common reagents. Most of the families analysed have been typed for the following markers of the HLA region: HLA-A, B, C and DR, BF, C4A, C4B and Glyoxalase I (GLO). The principal part of the GLO typing was made in Toulouse according to classical methods with slight modifications (Amaud eral., 1988). Recently C2 typing was performed on he material, primarily on samples from the parents and when at least one of the parents had an allele differing from the common one (C), the other family members were also typed (M. Abbal er al., personal communication). The coded data were entered in Toulouse and the data verification was done centrally in collaboration with the participating laboratories. The original HLA types of cells assigned by the different laboratories were the basis of a computer cell typing performed after computer serum analysis of the common set of sera (Sierp et al., 1986). This gave rise to a re-evaluated cell typing by computer which was checked for family segregation both by FAP analysis and by manual verification. This new set of homogeneously assigned HLA types was used in this study. The family analysis package (FAP) created for the 9th International Histocompatibility Workshop (Neugebauer er al., 1984) has been used to construct the HLA haplotypes. It estimates the  Recombination fractions in the HLA system 35 TABLE . Recombination fractions 0). minimum estimate (without taking linkage disequilibrium into account) Loci e (yo) e (yo) Femalehale paternal maternal LOD score ratio A-C 0.8 C-B 0.0 B-BF 0.6 C4B-DR 1 1 DR-GLO 6.0 1 o 0.2 0.8 1.2 11.0 406 504 232 216 70 1.25 1.33 1.09 1.83 No definite paternal recombinations were found between C and B loci. The complement loci BF, C4A and C4B) have been considered together, the possible recombinations between these loci have not been repeatedly tested. probabilities of the different haplotype solutions by an iterative procedure where the linkage disequilibria are estimated after every run and the values obtained are used for the subsequent run. As a result estimations of haplotype frequencies are given, which can be characterized by allele frequencies and linkage disequilibria values. The original programme has been modified and extended in order to estimate recombination fractions (Neugebauer, 1989) assuming no linkage disequilibrium (procedure one) vs. linkage disequilibrium for the alleles of different loci (procedure two). For example in testing the recombination fraction of HLA-A I, B8 haplotype the product of the allele frequency f(AI ) * f(B8) would be used in procedure one and the estimated haplotype frequency f(AI, B8) which includes the well known delta value by use of f(AI, B8) = f(AI) * f(B8) + A(AI, 88) would be used in procedure two. When FAP is used to generate the number of HLA alleles or haplotypes involved in recombinations (see Tables 3 and 4) it is difficult to evaluate the statistical significance of differences between these estimated numbers. A manual analysis has been made of all families with an unambiguous recombination, not taking into account families with probable but unconfirmed recombination (see Tables and 6). In each family, two alleles from each locus on he two parental haplotypes involved in the recombination have been counted. In the case of possible homozygosity for a locus (the recombination being shown by neighbouring loci) the allele has been counted twice as the blank frequency for most of the typed loci is small. Chi-squared statistics have been used to evaluate the significance of the deviations. RESULTS Of the total of 1382 families studied in PF 1332 families were included in the current analysis. Reasons for exclusion were possible extra paternity, lack of typing results in some family members, only one child typed or uninterpretable serological results. Table 1 shows estimated cross-over frequencies without accounting for linkage disequilibria, i.e. whenever a solution without cross-over was possible, it was chosen even if the deduced haplotypes were very rare. If linkage disequilibria are considered, the estimates change for some combinations of loci (Table 2). In general, maternal recombinations are more frequent than paternal recombinations and on average the ratio between female and male recombinations is 1.8 (distance of A-B B-DR and DR-GLO).  36 M. Thomsen et al. 4 bcM 1 1 0.6 0 8 * 6.0 GLO DPDQDuc4BF BC E A ur: 1 I I 1 BP 0 1000 2000 3000 4000 FIG. 1. Comparison of genetic centiMorgan cM]) and physical kilobasepairs [kbp]) distances in a simplified pma of the HLA region on chromosome 6. The physical distances are taken from Campbell & Trowsdale 1993), the genetic distances from Table 1 based on the paternal recombination fractions in pairwise comparisons without taking into account the linkage disequilibrium. The number of genes in the DR and class-I11 region depends on the haplotype. For the sake of clarity, the individual a nd B-chain genes in the class-I1 region have not been indicated. In the B-C interval no definite paternal recombinations were found, and three maternal recombinations were found by manual analysis. The estimates of the male HLA recombination fractions have been expressed as genetic distances (Fig. l , which for small distances are equivalent to the recombination fractions. When they are compared with the physical distances (Campbell & Trowsdale, 1993), it appears that some chromosomal regions (e.g. BF/DR) are involved in recombinations more often than others (e.g. NB). This study aimed to analyse whether certain HLA alleles were more frequent on recombinant than on non-recombinant haplotypes. All the recombinant haplotypes with cross-over between A and DR were selected and the distribution of allele frequencies compared with that on non-recombinant haplotypes. Table 3 shows that some DR alleles (DR3 and DRII) are involved frequently in recombinations. No gross differences were found for HLA-A, B, BF or C4 alleles (data not shown). When A, B, DR haplotypes are considered it appears that several of the most frequent haplotypes are under-represented among the recombinant haplotypes (Table 4). The three most common haplotypes: Al, B8, DR3; A3, B7, DR2; and A29, B44, DR7; represented about 26 of the recombinant haplotypes listed in Table 4 and about 58 of the non-recombi- nant haplotypes. In contrast, the two haplotypes carrying 818 and DR3 (A30,B18,DR3 and A2,B18,DR3) constitute 26 of the recombinant and only 6 of the non-recombinant haplotypes listed in Table 4. TABLE . Recombination fractions, minimum estimate with linkage disequilibrium Femalehale Loci 8 paternal) 8 maternal) LOD score ratio A-C 0.7 1.1 514 1.57 C-B 0.0 0.3 978 B-BF 0.7 0.6 34 1 0.85 C4B-DR 1.1 1.3 277 1.18 DR-GLO 6.0 11.0 82 1.83 A-B 0.7 1.2 818 1.71 B-DR 1.1 1.7 734 1.55 A-GLO 6.0 12.0 82 2.0 The B-BFandor the C4B-DR recombination fractions might be overestimated as the recombination fraction of the B-DR interval is less than the sum of the two former intervals. However, some of this discrepancy may be explained by the occurrence of a B-BF and C4B-DR recombination in the same family.  Recombination fractions in the HLA system 37 TABLE . DR alleles on recombinant and non-recombinant haplotypes Estimated numbers DR recombinant non-recombinant Total 1 2 3 4 7 8 9 I1 12 13 14 Blank 11 (2.0 )' 21 (2.8 ) 15 (1.8 ) 3 (2.1Yo) 19 (2.5 ) 38 (4.2 ) 2 (4.8 ) 37 (5.3 ) 2 (4.2 ) 5 (4.3 ) 7 (2.5 ) 6 (1.6 ) 1 730 860 74 1 815 137 40 667 46 110 362 268 ~ 562 749 898 762 830 140 42 704 48 115 368 275 Total number 166 (3.0 ) 5327 5493 The numbers have been rounded. As the numbers are estimated, the increase of DR3 and DRll has not been evaluated by statistical tests see Table 6). The percentage represents the number of recombinant haplotypes with a given DR allele out of the total number of haplotypes with this allele. Manual evaluation of the recombinant families was performed and Table 5 shows that B35 is over-represented on recombinant haplotypes when the whole A/DR interval is considered (uncorrected P < 0.001). No deviations were found for HLA-A alleles. The manual analysis of DR alleles (Table 6; Fig. 2) is in agreement with the estimations by FAP: an increased frequency of DR3 nd DRll is found on B/DR recombinant haplotypes (uncorrected P < 0.O001) and this is not involving A/B ecombinant haplotypes. An interesting pattern is observed for DR4 which has TABLE . Relative frequencies of recombinant and non-recombinant haplotypes ~ ~ Estimated number ~ ~ A cw B BF DR recombinant non-recombinant Total ~ 1 2 1 3 29 23/24 3 2 2 W3 2 2 ~ 617 3 710 710 0 410 4 514 110 510 7 7 17 62 8 7 44 44 35 44 51 18 7 18 S S S S F F F S S FI S S 7 4 3 2 7 7 1 4 2 3 2 3 0 (0 ) 0 (0 ) 2 (1.8 ) 1(2.2 ) 2 (O.g0/o) 1 (0.9Yo) 1 (2.9 ) 3 (4.1 ) 1(4.2 ) 3 (6.1 ) 3 (6.3 ) 2 SOYO) 36 24 223 111 111 45 33 71 23 46 48 2 36 24 225 112 113 46 34 74 24 49 51 4 Total number 19 (2.4 ) 773 792 The 19 recombinant haplotypes n this Table constitute 11 of the total number of recombinant haplotypes and the 773 non-recombinant haplotypes constitute 15 of the total number of non-recombinant haplotypes. As the numbers are estimated, no significance levels are given.
Related Search
Similar documents
View more
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks