Proc. Natl. Acad. Sci. USAVol. 83, pp. 1016-1020, February 1986Genetics

Genetic analysis of the fragile-X mental retardation syndrome withtwo flanking polymorphic DNA markers

(linkage/recombination fraction/restriction fragment length polymorphism/genetic counseling)

I. OBERLt*, R. HEILIG*, J. P. MOISAN*, C. KLOEPFER*, M. G. MATTtIt, J. F. MATTtIt, J. Bou0t,U. FROSTER-ISKENIUS§, P. A. JACOBS¶, G. M. LATHROP11, J. M. LALOUEL11, AND J. L. MANDEL* ***Laboratoire de Gdndtique Moldculaire des Eucaryotes du Centre National de la Recherche Scientifique, U.184 Biologie Moldculaire et Genie Gdndtique, 11rue Humann, 67085 Strasbourg, Cedex, France; tU.242 H6pital d'Enfants de la Timone, 13385 Marseille, France; tU.73 Chfteau de Longchamp, 75016 Paris,France; §Institut fur Humangenetik, Medizinische Hochschule-D-2400 Lubeck, Federal Republic of Germany; ¶J. A. Burns School of Medicine,Honolulu, HI 96822; and IlLaboratoire Anthropologie Biologique, Universitd Paris VII, 75221 Paris, France

Communicated by Jean Dausset, August 26, 1985

ABSTRACT The fragile-X mental retardation syndrome,one of the most prevalent chromosome X-linked diseases (,1 of2000 newborn males), is characterized by the presence inaffected males and in a portion of carrier females of a fragilesite at chromosome band Xq27. We have performed a linkageanalysis in 16 families between the locus for the fragile-Xsyndrome, FRAXQ27, and two polymorphic DNA markers thatcorrespond to the anonymous probe Stl4 and to the coagulationfactor IX gene F9. Our results indicate that the order of loci iscentromere-F9-FRAXQ27-Stl4-Xqter. The estimate of therecombination fraction for the linkage F9-FRAXQ27 is 0.12(90% confidence limits: 0.044-0.225) and 0.10 for FRAX-Q27-Stl4 (90% confidence limits: 0.040-0.185). Recombina-tion between Stl4 and F9 does not appear to be significantlydifferent in normal and fragile-X families. The two flankingprobes were used for diagnosis of the carrier state and fordetection of transmission of the disease through phenotypicallynormal males. They should also allow first-trimester diagnosiswith a reliability of about 98% in 40% of the families. Used inconjunction with the cytogenetic analysis, the segregationstudies with both probes should improve the genetic counselingfor the fragile-X syndrome and should be useful for the formalgenetic analysis of this unique disease.

The fragile-X mental retardation syndrome (fragile-X syn-drome) accounts for one-quarter to one-third of families withchromosome X-linked mental retardation, and it is present inapproximately 1 of 2000 newborn males. It may also accountfor 3-4% of all mental retardation in otherwise normalfemales. The fragile-X syndrome is a pleiotropic trait con-sisting of (i) the presence of a fragile site on the X chromo-some at the q27-q28 interface, induced in vitro by thymidinedeprivation, (ii) a variable degree of mental retardation inhemizygous males, usually accompanied by characteristicphysical features, and (iii) a 35% risk of mental impairmentin heterozygous females (see refs. 1-3).The genetics of the fragile-X syndrome departs from classic

chromosome X-linked inheritance in several respects. Thegene for the fragile-X syndrome, FRAXQ27, does not appearto be fully penetrant in males. Apparently normal males(cytogenetically and/or clinically) can transmit the disease,as first suggested from retrospective analysis of large pedi-grees (reviewed in refs. 1 and 2). On the other hand, thepercentage of clinically expressing females is much higherthan in other sex-linked diseases.The diagnosis of carrier females is difficult because only

about one-half of the females who carry the fragile-X muta-tion can be detected by their phenotype (mental retardationand/or fragile-site expression) (4). Prenatal diagnosis can be

performed by assaying for the fragile site in fetal cellsobtained by amniocentesis (5), fetal blood sampling (6), andmore recently by microvilli biopsy (7). The first two tech-niques can be performed in only a few centers because of thedifficulty of detection of the fragile site in amniocytes or thegreater obstetrical complexity of fetal blood sampling com-pared to amniocentesis. The reliability of fragile-site detec-tion in trophoblast cells is not yet documented. Thus,because of the prevalence of the disease and of the problemsencountered for genetic counseling, it is important to improvethe methods for carrier detection and prenatal diagnosis.Other aspects of the formal genetics of the fragile-X syn-drome appear intriguing. A segregation analysis suggestedthat the mutation rate at this locus is very high (7.2 x 10-4),but that mutations occur only in sperm (4). The gene seemsto be more penetrant in the offspring of daughters of trans-mitting males than in offspring of mothers of transmittingmales (20).

Segregation analysis with restriction fragment length poly-morphism (RFLP) linked to the disease locus could be veryuseful in investigating such problems, since these markersallow the assessment of the genotype of individuals. In thispaper, we report a linkage analysis between the fragile-Xsyndrome locus FRAXQ27 and two probes located in theq27-qter region of the human X chromosome. We havereported that RFLP within the coagulation factor IX gene F9was closely linked to FRAXQ27 and could detect transmis-sion through normal males (8). Here we show that F9 and avery polymorphic locus defined by the anonymous probeStl4 (9) flank the disease locus. We illustrate the use of thetwo probes for the detection of carrier women or normal maletransmitters. The two probes should be useful for prenataldiagnosis at present in about 40% of the families.

MATERIALS AND METHODSPreparation ofDNA and Blot Hybridization. Total genomic

DNA was extracted from human leukocytes or cultured cells,digested to completion with restriction endonucleases, frac-tionated by electrophoresis on 0.9% agarose gels as described(10), and blotted onto diazobenzyloxymethyl paper (11).

Probes. The probe for the coagulation factor IX gene F9was a 5.5-kilobase (kb) EcoRI genomic fragment called FIXP1, cloned in X1149, and screened with a coagulation factorIX cDNA probe (12). It contains exons b, c, and d andneighboring introns and detects a Taq I polymorphism (13).For the StJ4 locus, we first used a 9.3-kb probe (Stl4.9) thatdetects at least 10 allelic fragments in DNAs digested withTaq I and three independent RFLPs with the enzyme Msp I

Abbreviations: RFLP, restriction fragment length polymorphism;kb, kilobase(s); fragile-X syndrome, fragile-X mental retardationsyndrome.**To whom reprint requests should be addressed.

1016

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

4, 2

020

Proc. Natl. Acad. Sci. USA 83 (1986) 1017

(9). We recently have cloned a different genomic fragmentthat detects a simpler pattern of constant bands than theoriginal probe and gives a better signal for the polymorphicTaq I bands (probe St14.1). This probe detects a supplemen-tary Taq I RFLP with two alleles (R.H. and I.O., unpublisheddata).

Fragile-X Families. Pedigrees and cytogenetic data for thefamilies 3, 4, and 13 have been published (8, 14, 15). Thecytogenetic analysis in 13 other families was carried out asdescribed (14, 16). To detect possible false paternity, wehybridized the blotted DNAs to several chromosome X-specific probes detecting Taq I RFLPs: the high number ofalleles at the StJ4 locus should in many cases reveal falsepaternity.

Linkage Analysis. Detection of linkage, estimation of re-combination fraction, and inference of gene order wereperformed with the LINKAGE programs (17). We assumeda gene frequency of 0.0006 and a mutation rate of 0.0002,equal in males and females. We verified that our resultsremained unaffected when these two parameters were variedover a wide range. Values assigned to other parameters, suchas penetrances, are discussed in the text.

RESULTSLinkage Analysis. We analyzed DNAs from 169 individuals

belonging to 16 families with fragile-X syndrome. Seven ofthese families are two-generation nuclear families, while theothers are three-generation or have more complex pedigrees(see the legend to Fig. 2). The proportion of families infor-mative for the Taq I polymorphism detected with the FIX P1probe is higher than in the general population because suchfamilies were initially selected for study. Eleven familieswere informative for both FIX P1 and St14 probes. Severallarge families exhibited no recombination between FRAXQ27and either the F9 or the St14 loci (see Figs. 1 and 2). Incontrast, one family showed several recombination eventswith the two probes (family 5, Figs. 1 and 2).Two-point linkage analysis was performed by using the

LINKAGE computer program (17). Penetrance, as estimatedfrom the clinical or cytogenetic status, is not complete anddiffers in males and in females; this was taken into accountin the calculations. Because only rough estimates of pene-trances are available, we considered the influence of varyingpenetrance values for males and females (Table 1). Twoalternative coding rules were considered for the definition ofaffection status. First, the phenotype of all family memberswas defined as affected or normal according to cytogeneticand/or clinical data. Calculations were then performed withpenetrance values varying from 0.8 to 0.95 for males and

fixed at 0.56 for females. The first and third figures arederived from the segregation analysis of Sherman et al. (4).Second, unambiguous male transmitters were entered asaffected, the status of cytogenetically normal adult womenwas considered as unknown, and a high penetrance value(0.9) was tested for younger females (under 18), since it islikely that most of the carriers are detected when cytogeneticanalysis is carried out in this age category. (In our pedigreeset, out of 25 young females at 50% risk of being carrier, 13showed the fragile site; see also refs. 18 and 19.)The lod scores (logarithm of odds of linkage) obtained for

each family, as calculated with model A (see Table 1) arepresented in Fig. 2. Only families 5 and 16 showed negativelod scores at a recombination fraction of 0.1 for both StJ4 andF9 linkage. When the data from all 16 families were pooled,the recombination fraction (6) for which the maximum lodscore is obtained was stable under the various modes ofcalculation (6 varying from 0.114 to 0.125 for the F9-FRAXQ27 linkage and from 0.097 to 0.101 for Stl4-FRAXQ27). The maximum lod score varied from 4.95 to 6.18for F9-FRAXQ27 and from 7.90 to 9.50 for StJ4-FRAXQ27(Table 1). In both cases, the highest lod score value wasobtained when the least ambiguity (model G) was attributedto the genotype of normal individuals at the fragile-X syn-drome locus-i.e., high penetrance for young females and, assuggested (20), for brothers of affected males. Unless other-wise specified, these parameter values were used in subse-quent calculations. The 90% confidence limits for the recom-bination fractions are 0.04 to 0.22 for F9-FRAXQ27 and 0.04to 0.18 for Stl4-FRAXQ27 (these are the recombinationfractions for which the relative probability of linkage is1/10th of the maximum; see ref. 21).Order of Loci. Linkage analysis in normal families has

shown that the recombination fraction between the Stl4 andF9 loci was approximately 0.30 (11 recombinations out of 37meioses) (9). The families with fragile-X syndrome yielded amaximum lod score of 3.83 for a recombination fraction of0.186 (Table 2). Thus, the genetic distance between Stl4 andF9 appears to be close to the sum of the distance of eachmarker to FRAXQ27, suggesting that the fragile-X syndromelocus is located between the two markers. This was support-ed by inspection of families that are doubly informative: forinstance, in family 13, out of 16 meioses there was norecombinant with F9 and there were two recombinants withStJ4 (in individuals III8 and II19). The reverse was found infamily 9 in which, out of eight meioses, two show recombi-nation between FRAXQ27 and F9 (individuals II3 and II8)and none between Stl4 and FRAXQ27 (Fig. 1). The jointsegregation of the three loci was analyzed with the LINK-

a AA

III

A Au4 2'

a A a a AA

7 2 2 7 24

2 3 4 5

F A M I L Y

FA MI LY 1 3 (4) (58)

t .7 a (A) Aa a7 4 8 (3) 5 8 a482 4 5 a 7 9

A a A1

9 Iv OD AMILY 5

FIG. 1. Inheritance of theTaq I polymorphisms detectedby the probe Stl4 (alleles 1-8)and by the probe FIX P1 [allelesA (1.8 kb) and -a (1.3 kb)] forcoagulation factor IX gene F9,in families with the fragile-Xsyndrome. Genotypes in paren-theses were deduced from thegenealogy. a, o, Phenotypicallynormal but not examined; z, ®,deceased. In the following sym-bols, the left half denotes theclinical status and the right halfdenotes the cytogenetic analy-sis: mJ, (D, normal with no fragilesite in cells [0% fra(X)(q27)cells]; *, a, dull with 1-2% offra(X)(q27) cells; *, *, mentallyretarded with more than 2% offra(X)(q27) cells.

Genetics: Oberle' et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

4, 2

020

Proc. Natl. Acad. Sci. USA 83 (1986)

St 14- FRA X

U-

FIX - FRA X13

.E\IU

-4

-3

2

0

CO)0

0

1

16 061I -2

I:~~~~18

9

1 10 20 1 10 20

RECOMBINATION FRACTION (%)

FIG. 2. Estimation of the recombination fraction between thefragile-X syndrome and two polymorphic DNA markers: Stl4 andthe coagulation factor IX gene F9. Lod scores for the linkagesStJ4-FRAXQ27 and F9-FRAXQ27 were calculated in each familyunder model A (see Table 1). Families are numbered on the left as inthe text and in the other figures. Families 7-12 and 16 are two-generation nuclear families, families 13-15 are three-generationfamilies, and families 1-6 are complex multigeneration families.

AGE program. There was no evidence for interference (XI =

0.1), which was therefore neglected for tests of gene order.This analysis favored the order F9-FRAXQ27-Stl4 againstthe two alternatives F9-StJ4-FRAXQ27 and Stl4-F9-FRAXQ27 with relative odds of 2333:43:1, respectively.

Carrier Diagnosis with DNA Markers. The probability of adouble recombination between StJ4 and F9 can be estimatedto be 1-2%. Thus, in families informative for both markers,segregation analysis can be used for carrier and/or prenataldiagnosis, provided that no recombination has occurredbetween the two marker loci in the individual at risk. Anexample is shown in Fig. 3, where the sister (IV10) of an

affected male asked for genetic counseling on the occasion ofa pregnancy. She was clinically and cytogenetically normal,

Table 1. Influence of varying penetrance values in males andfemales on the estimation of the recombination fractionbetween FRAXQ27 and Stl4 or F9

Penetrance Maximum lod score

Cod- values Stl4-FRAXQ27 F9-FRAXQ27ing Fe- (0.097 - 0- (0.114 -< 0rules Model males Males 0.101) 0.125)

1 A 0.56 0.8 8.17 4.95B 0.56 0.9 8.43 5.28C 0.56 0.95 8.50 5.43

2 D 0.56 0.8 7.90 5.10E 0.56 0.9 8.22 5.48F 0.56 1.0 8.48 5.89G 0.9 1.0 9.50 6.18

Under coding rules 1, individual phenotype at the FRAXQ27 locuswas defined as affected or normal according to cytogenetic and/orclinical data. Under coding rules 2, unambiguous male transmitterswere entered as affected, cytogenetically normal adult women asunknown2 and cytogenetically normal young women (under 18) asnormal. 6 is the recombination fraction for which the lod score ismaximum.

with thus a 30% chance only of being a carrier (assuming apenetrance of 0.56). She has received from her doublyheterozygous mother F9 and StJ4 alleles that are differentfrom those of her affected brother or cousins. Risk calcula-tion shows that her final risk of being a carrier is only 1.3%.In the same family, the demonstration that transmissionoccurred through a male (114, see below) implies that II18 isa carrier. Her cytogenetically normal daughter has a proba-bility of about 76% of being a carrier since she inherited theF9 allele from the grandfather (the Stl4 locus is not infor-mative in this case because 1118 is homozygous for thismarker).

Detection of Probable Male Transmitters. In several fami-lies the segregation analysis strongly suggested transmissionthrough a male. In family 1 (Fig. 3), where the right part ofthe family was analyzed first, all three affected males haveinherited the allele combination of the maternal grandfather.Further enquiry revealed that the half-sister II11 of this manwas the grandmother of two affected cases, IV3 and IV5.Thus, male transmission was demonstrated unequivocally. Inother cases, such pedigree evidence was not available. Infamily 2, individuals of generation III with fragile-X haveinherited the A-5 combination while unaffected males havethe a-1 combination (Fig. 4). It is unlikely that the disease wastransmitted through the grandmother, for it would imply atleast three recombination events between F9 and FRAXQ27.On the contrary, no recombination would have occurred ifthe grandfather was the carrier of the mutation. In family 14,the mother of affected children (112) was heterozygous for the

Table 2. Test of heterogeneity in recombination among fragile-X families or between fragile-X andnormal families

Estimation of recombination

90% Hetero-confidence Z geneity test

Linkage 6 limits (a)Among fragile-X F9-FRAXQ27 0.125 0.047-0.25 4.95 12 = 15.7

families Stl4-FRAXQ27 0.097 0.037-0.19 8.17 4= 15.6

Between fragile-X F9-StJ4 0.186 0.09-0.32 3.83 - 1and normal families F9-Stl4 0.284 0.14-0.47 1.24 x2 - .06

Recombination fractions were estimated under model A (see Table 1). Tests of heterogeneity were

performed as described by Morton (22): XI-, = (21n 10) (E P - 2) where Zi is the maximum lod score

for the jth family and 2 is the maximum lod score for the whole set of families.

1018 Genetics: Oberld et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

4, 2

020

Proc. Natl. Acad. Sci. USA 83 (1986) 1019

III

'V

v

FA M I LY 1

F9 RFLP but not for the Taq I RFLP detected by the St14probe (Fig. 4). In the three sibs of generation III, theFRAXQ27 mutation appears to segregate with the a allele,which comes from the grandfather. It is very unlikely (about1 chance in 400) that the grandmother was a carrier, whichwould require three recombination events in three meioses.This suggests that the grandfather was a silent carrier or thata new mutation occurred in sperm.

DISCUSSIONWe have performed an analysis of linkage between a locusassociated with a prevalent chromosome X-linked disease,the fragile-X syndrome, and two polymorphic loci detectedby DNA probes previously mapped in the q27-qter region ofthe X chromosome. In our study, most families were infor-mative for both polymorphic markers; therefore, we ana-lyzed the three loci jointly, for it is more efficient thanpairwise linkage to infer gene order (23). It was important totake into account different, incomplete penetrances in malesand females, which are particular features of this syndrome.For instance, in three-generation families, the knowledge ofthe genotype of the maternal grandfather at the marker locidoes not always yield the phase as with other chromosomeX-linked diseases, since he might be a "normal" carrier (seefamilies 1 and 13). The FRAXQ27 penetrance in males wasestimated as 80% by Sherman et al., but it might be higher insibships with affected males (20). The determination of thecarrier status in females is also difficult: only about 56% ofcarrier females show clinical or cytogenetic signs of theFRAXQ27 mutation. This value is age dependent and, youngfemales apparently exhibit the fragile site in a much greaterproportion of the cases (18, 19).We have analyzed the data under various assumptions.

Penetrance was varied from 0.8 to 0.95 in males. We used apenetrance value of 0.56 for females of all ages, or, alterna-tively, we used a penetrance value of0.9 for girls under 18 andscored adult women as phenotypically unknown. Althoughthe odds in favor of linkage vary with the various modes of

11

III

(A) (Aa)

| A A a A a Aa|5 1 6 5 1 5 1

6< 8Z

A a A a AS 6 5 1 5 1

1 2 3 4 5

F A MILY 2

FIG. 3. Carrier detection and transmis-sion by a phenotypically normal male in afamily with the fragile-X syndrome. Symbolsare as in Fig. 1.

calculation (Table 1), the recombination fraction estimateremains stable. This is also true for gene order: with higherpenetrance values in males and/or females, the most likelyorder is preserved, but odds in its favor are increased.Our estimate of recombination between F9 and FRAXQ27

is 0.12 (90% confidence limits of 0.04 to 0.22). Although ourfirst study (8) had suggested that the two loci might be closer,we had pointed out that the 90% confidence limits were0-12% recombination. Meanwhile, Choo et al. (24) publishedan analysis of five families showing approximately 15%recombination (5 of29 or 4 of 23, depending on whether or notnormal females were taken into account), similar to the valuefound in our study. The Stl4 locus also displays about 10%recombination with FRAXQ27 with a maximum lod score of9.5 (90% confidence limits of 0.04-0.18). The three-pointlinkage analysis as well as simple examination of largenuclear pedigree (Fig. 1) places FRAXQ27 between the twomarker loci. Since a similar conclusion has been reached byin situ hybridization of the FIX and Stl4 probes to chromo-somes displaying the fragile site at Xq27, this supports furtherthe notion that the mutation is located in the same region asthe cytogenetically demonstrable fragile site (25).The multilocus linkage data can be used to investigate two

important questions. First, is there genetic heterogeneity infragile-X families-i.e., can the data be explained by theexistence of a single locus? Linkage analysis does not provideevidence for heterogeneity among families as tested by themethod of Morton (22) (Table 2; see, however, Note Addedin Proof). Second, does the presence of the FRAXQ27mutation affect recombination in the telomeric region of theX chromosome? It has been proposed that the fragile siteregion exhibits a high frequency of recombination in normalfamilies, since the two probes 52A and FIX P1 mapping in q27by in situ hybridization are at 0.30 recombination unit fromthe cluster of loci in q28 that includes those defined by probesDX13 and Stl4 and the G6PD (glucose-6-phosphate dehy-drogenase) and HEMA (hemophilia A) loci (9, 26). Szabo etal. (27) have suggested that the recombination fraction

(a) AA(4) 4 4

A Aa4 44

2 3

A a A A a4 4 4 4

aFIG. 4. Possible transmission by a phe-

<b CD) g notypically normal male in two families with2 3 the fragile-X syndrome. Symbols are as in

F A M I L Y 1 4 Fig. 1.

Genetics: Oberle' et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

4, 2

020

Proc. Natl. Acad. Sci. USA 83 (1986)

between F9 and the G6PD cluster is much smaller in fragile-Xfamilies. The present data do not support the hypothesis ofSzabo et al. because the recombination fraction between Stl4and F9 in fragile-X families, as estimated from three-locusanalysis, is about 0.19 (or 0.22 when estimated from two-point linkage data), whereas there was 28% recombination innormal families. A x2 test showed that these differences arenot significant (Table 2). Part of the discrepancy with theconclusions presented by Szabo et al. (27) derives from theiruse of our initial linkage data (8), which suggested closerlinkage between the F9 and FRAXQ27 loci.Our results establish the validity of the StJ4 and F9

markers as tools in the genetic analysis in fragile-X families.They flank the disease locus and can be used in conjunctionwith cytogenetic tests for prenatal diagnosis and carrierdetection in families informative at both loci. Given therecombination fraction between each test locus and theFRAXQ27 locus, double recombinants should occur in only1-2% of the meioses in the Stl4-F9 interval. The heterozy-gosity for the combined Taq I and Dde I RFLPs detected bythe FIX P1 probe is 60% in Caucasians (28), while that of theTaq I and Msp I RFLPs at the Stl4 locus is about 90% (9).Thus, 54% of the families would be doubly informative.However, in 25-30o of the cases, a single recombinationevent will occur between the two marker loci, which willprevent diagnosis. Therefore, we estimate that about 40% ofCaucasian cases could benefit from a segregation analysiswith the two probes.The marker study can also help to detect new mutations (as

shown for hemophilia A; unpublished data) and/or thosefamilies in which the disease was transmitted through normalmales. Out of the 16 families analyzed, families 3 and 4 wereknown to have normal male transmitters, based on thepedigree alone (a conclusion only possible for large pedi-grees). In family 13, the F9 segregation data were sufficientto establish male transmission beyond doubt (8), and the Stl4segregation data confirm this conclusion. In family 1 the F9segregation data also suggested male transmission, whichwas confirmed by further pedigree analysis (Fig. 3). Sugges-tive evidence for male transmission (or for a neomutationarising in a male) has been obtained in two other families(families 2 and 14, Fig. 4). These results would be inagreement with the high percentage of phenotypically normalmale carriers estimated by Sherman et al. (20), although it ispossible that an ascertainment bias exists in favor of suchfamilies-i.e., they will tend to have a larger number ofaffected individuals in the third generation because all of thedaughters in generation II are carriers. It must be emphasizedthat detection of male carriers is of great importance forgenetic counseling. Finally, the segregation analysis withlinked markers could help in testing the hypothesis thatpenetrance might be different among sibships of normal malecarriers and sibships of affected males (20) by allowing us toinfer the genotype of clinically and cytogenetically normalmales.Although the present probes represent useful tools, it is

desirable to find markers closer to the fragile-X syndromelocus FRAXQ27 and to increase the number of familiesinformative for probes proximal to that locus. The familiesalready investigated are very useful to quickly map any newpolymorphic probe with respect to FRAXQ27, Stl4, and F9,since it is necessary to analyze only those meioses showingrecombination between FRAXQ27 and one of the test loci. Inthis way, we have recently mapped several probes betweenthe centromere and F9, close to this latter locus on theproximal side of FRAXQ27 (unpublished data). However, ifthe FRAXQ27 region is a hot spot for recombination (9, 27),it might be difficult to find probes genetically closer unlessthey are physically in the immediate vicinity of the mutation.

Note Added in Proof. Since submission of this manuscript, Brown etal. (29) have reported finding significant heterogeneity among fam-ilies for the recombination fraction between F9 and FRAXQ27 lociwhen using Morton's test on families with more than three informa-tive males. With the same criteria we still found no significantheterogeneity for the Stl4-FRAXQ27 linkage (XI = 14.0), but wefound heterogeneity for the F9-FRAXQ27 linkage (XI = 11.7; 0.05 6P - 0.025). When family 13 was taken out of our sample, we nolonger found significant heterogeneity (X2= 6.3, for6 = 0.18). Similarconclusions were obtained when comparing families with demon-strated normal male carriers to all other families-i.e., no heteroge-neity for the Stl4-FRAXQ27 linkage and significant heterogeneityfor F9-FRAXQ27 only when family 13 was included.

We thank Dr. De Grouchy (Paris), Dr. J. Fraisse (St Etienne), andDr. R. Walbaum (Roubaix) for referring families to us, and C. Werle,C. Aron, and J. Kaibach for help in the preparation ofthe manuscript.This work was supported by grants from the Ministbre de l'Industrieet de la Recherche (C0671) and from Agir (to J.L.M.).

1. Mattei, J. F., Mattei, M. G., Auger, M. & Giraud, F. (1984) J. Ge'net.Hum. 32, 167-192.

2. Turner, G. & Jacobs, P. A. (1983) in Advances in Human Genetics, eds.Harris, H. & Hirschorn, K. (Plenum, New York), Vol. 13, pp. 83-112.

3. Fishburn, J., Turner, G., Daniel, A. & Brookwell, R. (1983) Am. J. Med.Genet. 14, 713-724.

4. Sherman, S. L., Morton, N. E., Jacobs, P. A. & Turner, G. (1984) Ann.Hum. Genet. 48, 21-37.

5. Jenkins, E. C., Brown, W. T., Duncan, C. J., Brooks, J., Ben-Yishay,M., Giordano, F. M. & Nitowsky, H. M. (1981) Lancet ii, 1292.

6. Webb, T., Butler, D., Insley, J., Weaver, J. B., Green, S. & Rocek, C.(1981) Lancet i, 1423.

7. Tommerup, N., Sondergaard, F., Tonnesen, T., Kristensen, M.,Arveiler, B. & Schinzel, A. (1985) Lancet i, 870.

8. Camerino, G., Mattdi, M. G., Mattdi, J. F., Jaye, M. & Mandel, J. L.(1983) Nature (London) 306, 701-704.

9. Oberld, I., Drayna, D., Camerino, G., White, R. & Mandel, J. L. (1985)Proc. Nati. Acad. Sci. USA 82, 2824-2828.

10. Grunebaum, L., Cazenave, J. P., Camerino, G., Kloepfer, C., Mandel,J. L., Tolstoshev, P., Jaye, M., De La Salle, H. & Lecocq, J. P. (1984)J. Clin. Invest. 73, 1491-1495.

11. Alwine, J. C., Kemp, D. J., Parker, B. A., Reiser, J., Renart, J., Stark,G. R. & Wahl, G. M. (1979) Methods Enzymol. 68, 220-242.

12. Camerino, G., Grzeschik, K. H., Jaye, M., De La Salle, H., Tolstoshev,P., Lecocq, J. P., Heilig, R. & Mandel, J. L. (1984) Proc. Natl. Acad.Sci. USA 81, 498-502.

13. Tonnesen, T., Sondergaard, F., Guttler, F., Oberl6, I., Moisan, J. P.,Mandel, J. L., Hauge, M. & Damsgard, E. M. (1984) Lancet iU, 932.

14. Rhoads, F. A., Oglesby, A. C., Mayer, M. & Jacobs, P. A. (1982) Am.J. Med. Genet. 12, 205-217.

15. Froster-Iskenius, U., Schulze, A. & Schwinger, E. (1984) Hum. Genet.67, 419-427.

16. Mattdi, M. G., Mattdi, J. F., Vidal, I. & Giraud, F. (1981) Hum. Genet.59, 166-169.

17. Lathrop, G. M. & Lalouel, J. M. (1984) Am. J. Hum. Genet. 36,460-465.

18. Jacobs, P. A., Glover, T. W., Mayer, M., Fox, P., Gerrard, J. W.,Dunn, H. G. & Herbst, D. S. (1980) Am. J. Hum. Genet. 7, 471-489.

19. Fryns, J. P. (1984) Clin. Genet. 26, 497-528.20. Sherman, S. L., Jacobs, P. A., Morton, N. E., Froster-Iskenius, U.,

Howard-Peebles, P. N., Nielsen, K. B., Partington, M. W., Sutherland,G. R., Turner, G. & Watson, M. (1985) Hum. Genet. 69, 289-299.

21. Gusella, J. F., Wexler, N. S., Conneally, P. M., Naylor, S. L.,Anderson, M. A., Tanzi, R. E., Watkins, P. C., Ottina, K., Wallace,M. R., Sakaguchi, A. Y., Young, A. B., Shoulson, I., Bonilla, E. &Martin, J. B. (1983) Nature (London) 306, 234-238.

22. Morton, N. E. (1978) Cytogenet. Cell. Genet. 22, 15-36.23. Lathrop, G. M., Lalouel, J. M., Julier, C. & Ott, J. (1984) Proc. Nati.

Acad. Sci. USA 81, 3443-3446.24. Choo, K. H., George, D., Filby, G., Halliday, J. L., Leversha, M.,

Webb, G. & Danks, D. M. (1984) Lancet Ii, 349.25. Mattdi, M. G., Baeteman, M. A., Heilig, R., Oberl6, I., Davies, K.,

Mandel, J. L. & Mattdi, J. F. (1985) Hum. Genet. 69, 327-331.26. Drayna, D., Davies, K., Hartley, D., Mandel, J. L., Camerino, G.,

Williamson, R. & White, R. (1984) Proc. Nati. Acad. Sci. USA 81,2836-2839.

27. Szabo, P., Purrello, M., Rocchi, M., Archidiacono, N., Alhadeff, B.,Filippi, G., Toniolo, D., Martini, G., Luzzatto, L. & Siniscalco, M.(1984) Proc. Nati. Acad. Sci. USA 81, 7855-7859.

28. Winship, P. R., Anson, D. S., Rizza, C . & Brownlee, G. G. (1984)Nucleic Acids Res. 12, 8861-8872.

29. Brown, W. T., Gross, A. C., Chan, C. B. & Jenkins, E. C. (1985) Hum.Genet. 71, 11-18.

1020 Genetics: Oberle' et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

4, 2

020