J7Med Genet 1997;34:395-399

Trisomy 15 rescue with jumping translocation ofdistal 1 5q in Prader-Willi syndrome

Koenraad Devriendt, Paul Petit, Gert Matthijs, Joris R Vermeesch, Maureen Holvoet,Alain De Muelenaere, Peter Marynen, Jean-Jacques Cassiman, Jean-Pierre Fryns

AbstractWe report a patient with Prader-Willisyndrome (PWS) and mosaicism for a denovo jumping translocation of distal chro-mosome 15q, resulting in partial trisomyfor 15q24-qter. A maternal uniparentalheterodisomy for chromosome 15 waspresent in all cells, defining the molecularbasis for the PWS in this patient. Thetranslocated distal 15q fragment was ofpaternal origin and was present as ajumping translocation, involving threedifferent translocation partners, chromo-somes 14q, 4q, and 16p. The recipientchromosomes appeared cytogeneticallyintact and interstitial telomere DNA se-quences were present at the breakpointjunctions. This strongly suggests that theinitial event leading to the translocation ofdistal 15q was a non-reciprocal transloca-tion, with fusion between the 15q24 break-point and the telomeres of the recipientchromosomes. These observations arebest explained by a partial zygotic trisomyrescue and comprise a previously unde-scribed mechanism leading to partialtrisomy.(JMed Genet 1997;34:395-399)

Keywords: Prader-Willi syndrome; jumping transloca-tion; uniparental disomy; partial trisomy

Centre for HumanGenetics, UniversityHospital Leuven,Herestraat 49, B-3000Leuven, BelgiumK DevriendtP PetitG MatthijsJ R VermeeschM HolvoetP MarynenJ-J CassimanJ-P Fryns

MPIOLV ter Engelen,Klerken, BelgiumA De Muelenaere

Correspondence to:Dr Devriendt.

Received 5 August 1996Revised version accepted forpublication 13 December1996

Prader-Willi syndrome (PWS) is caused by theabsence of a paternal genetic contribution tochromosomal region 15ql 1-13, either throughdeletion on the paternal chromosome orthrough the presence of maternal uniparentaldisomy for chromosome 15 (UPD 15 mat).' 2Other chromosomal aberrations have beenreported in PWS, such as unbalanced translo-cations or a marker chromosome derived fromchromosome 15, resulting in either a deletionin paternal chromosome 15ql 1-13 or inUPD15 mat.9 Jumping translocations areextremely rare and describe the translocationof the same chromosomal fragment to differenttranslocation partners in different cell lines of asingle person. Interestingly, the initial descrip-tion of a jumping translocation was also in apatient with PWS."' The majority of jumpingtranslocations reported so far involve hetero-chromatic chromosomal regions, such as telo-meres, centromeres, or satellites." Rarely,jumping translocations with an interstitialbreakpoint in one of the chromosomes havebeen reported and, interestingly, most of thesecases were patients with the PWS, with a

breakpoint in chromosomal region15ql 1-13.4We report here a patient with Prader-Willi

syndrome carrying a jumping translocationresulting in a partial trisomy.

Subjects and methodsCLINICAL DATAThe patient is a girl, the third child of healthy,non-consanguineous parents. Family history isnegative with regard to mental retardation orcongenital malformations. At the time of birth,the mother was 25 and the father 29 years old.During pregnancy, fetal movements werereduced and there was polyhydramnios. Shewas born at term with a birth weight of 2200 g(3rd centile=2500 g). The neonatal period andinfancy were characterised by severe hypotoniaand major feeding difficulties, necessitatingfrequent hospital admission and nasogastrictube feeding. Weight gain was poor. Aroundthe age of 2 years, she developed a markedchange in feeding behaviour, with an increasein appetite. This led to the gradual develop-ment of obesity during childhood. At the age of22 months, weight was 10 kg (3rd-25thcentile), height 85 cm (50th centile), and headcircumference 46 cm (3rd-25th centile). At theage of 10 years, weight was 42.5 kg (90th-97thcentile), height 135 cm (25th-50th centile),and head circumference 52 cm (25th-50thcentile). She suffered from recurrent urinaryinfections and at the age of 2 years unilateralvesicoureteral reflux was surgically corrected.Psychomotor development was severely de-layed: she could sit at 2.5 years and walk at 3years. First words appeared at 8 years. Now, atthe age of 22 years, there is truncal obesity,with weight 73.5 kg (90th-97th centile), height153 cm (3rd-25th centile), and head circum-ference 55 cm (75th centile). She is dysmor-phic, with short hands (hand length 15.5cmand finger length 6.5 cm, both below the 3rdcentile) and short feet, low set ears, andalmond shaped eyes (fig 1). The saliva is stickyand the teeth are carious. There is no hypopig-mentation. The corners of the mouth aredownturned and the palate is high arched. Sheis myopic. There is hypotonia, with genuvalgum and dorsal kyphosis. She is severelymentally retarded and exhibits obsessive be-haviour, temper tantrums, and skin picking.Sleep is disturbed, with frequent awakening.There is primary amenorrhoea. She foragesfood. Pain sensitivity is diminished. She fulfilsall major and seven minor criteria of the

395

on June 14, 2021 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.34.5.395 on 1 M

ay 1997. Dow

nloaded from

http://jmg.bmj.com/

Devriendt et al

Wl.

Figure 2 G banlded parti'al karyotype showinlg,from leftto right, chromosomes 4, 14, 15, and 16. Note the 4q+,14q+, and 16p+ (arrowheads). This fragmeint wastentatively identified as 15q24-qter.

I ...i;..... $.. i l

fl :;3.f

.. : . ::. . tc;. ........ ,

$;-1 N S.. .... .. ......... ......... .........

tg

Figure I Clinicalfeatures of the patient at 22 years. Notethe truncal obesity, short stature, small hands andfeet, andhypotonia, with downturned corners of the mouth.

recently proposed diagnostic criteria forPrader-Willi syndrome.'2

METHODS

Cytogenetic analysisChromosome studies on peripheral lym-phocytes and skin fibroblasts were performedaccording to standard cytogenetic techniques,and karyotyping was by Giemsa banding.Fluorescent in situ hybridisation (FISH) anddetection were carried out using coatasome 15chromosome paint and SNRPN/PML probes,obtained from ONCOR (Gaithersburg, MD).The PML probe supplied together with theSNRPN probe is present as a control probeand recognises sequences in 15q22. A biotinlabelled FES cosmid probe (15q26.1) (a giftfrom E Schoenmakers, Leuven) and oligonu-cleotide (TTAGGG)7 probes were applied asdescribed by Pinkel et al'3 and Vermeesch et al"respectively. Pictures were taken by digitalimaging microscopy using a cooled chargecoupled device camera system (Photometrics).Merging and pseudocolouring were performedusing the SmartCapture software (Vysis, Stutt-gart, Germany). At least 30 metaphases wereexamined with each probe.

Molecular analysisGenomic DNA was extracted from peripheralwhite blood cells and cultured fibroblasts andanalysed by Southern blotting using probes

Table I Distribution of the different karyotypes

Karvotype 46,XX 46,XX,4q+ 46,XX,14q+ 46,XX,16p+ Total

Lymphocytes (No of cells) 8 11 85 3 107Fibroblasts (No of cells) 51 14 35 0 100

PW7 1'5 (a gift from Dr B Horsthemke) andDN34E as previously described.'6

Analysis ofpolymorphic microsatellite mark-ers was done using PCR amplification on thepatient (white blood cells and fibroblasts) andher mother (white blood cells). DNA from thefather was not available. The following loci onchromosome 15 were examined: D15S122,D15S165, APW, D15S123, FES, D15S107,and D115S120. The primer sequences wereobtained from GDB.

ResultsCYTOGENETIC ANALYSISCytogenetic analysis on both lymphocyte andfibroblast tissues showed a normal 46,XX cellline and, in addition, unbalanced karyotypeswith partial trisomy 15q24-qter (table 1, fig 2).Mosaicism consistent with a jumping translo-cation was observed with a predominant46,XX,14q+ cell line as well as 46,XX,4q+and 46,XX,16p+ cell lines (table 1, fig 2).Chromosome painting with a chromosome 15specific probe confirmed that the extra chro-mosomal material originated from chromo-some 15 (fig 3A). FISH using a cosmid probefor the FES gene, located on chromosome15q26. 1, showed three signals in all unbal-anced metaphases, one on the distal region ofeach chromosome 15 and one on the extramaterial translocated onto chromosomes 14q,4q, and 16p (fig 3B). FISH using the SNRPN/PML cocktail probe showed normal signals at15ql 1 and 15q22 on both chromosomes,excluding a deletion involving the SNRPNgene (fig 3C). This also confirmed that thetranslocated region in the unbalanced cellsoriginated from the distal 15q region, with abreakpoint located between 15q22 (PMLcon-trol probe) and 15q26.1 (FES locus).The karyotypes of the parents and sibs were

normal after G banding.

PARENTAL ORIGIN OF THE CHROMOSOMES 15 ANDTRANSLOCATED FRAGMENT DISTAL 1 5qThe results obtained with probes DN34 andPW71 showed hypermethylation of the PWSregion, indicating the absence of a paternalallele (not shown).'556 This is compatible withUPD 15 mat. A submicroscopic deletion on thepaternal chromosome 15ql 1 gives the samemethylation pattern, but this was excluded byFISH using a cosmid probe from the SNRPNgene (fig 3C). A more detailed analysis of theparental origin of the chromosomes 15 wasdone using polymorphic markers distributedalong this chromosome. Four of the sixmarkers analysed were informative and showedthat the patient had inherited two different

396

on June 14, 2021 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.34.5.395 on 1 M

ay 1997. Dow

nloaded from

http://jmg.bmj.com/

J7umping translocation in Prader-Willi syndrome

Figure 3 (A) FISH analysis using a chromosome 15 painting library. Note that thefragment translocated to chromosome l4qter originatedfromchromosome 15. (B) FISH analysis with cosmid probe FES (located on chromosome 15q26). Three distinct signals were present, two on the normalchromosomes 15 and one on the distal long arm ofchromosome 4, at the site of the translocatedfragment. (C) FISH analysis using the SNRPN/PMLcocktail probe showed the presence of normal signals at 15q1 1 and 15q22 on both chromosomes, excluding a deletion involving the SNRPN gene. Moreover,the breakpoint on chromosome 15q must be located between PML (15q22) and FES (15q26).

PatientWBC

Patientskin f

APW FES

Mother '-FWBC

Figure 4 Analysis of microsatellite markers APWand FES. DNA from the patient'speripheral white blood cells (top row) andfibroblasts (middle row) andfrom maternalwhite blood cells (bottom row) was usedfor PCR amplification of the polymorphic repeatsAPW (dinucleotide repeat) and FES (tetranucleotide repeat). For FES, three alleles areobservedfor the patient, two of which are inheritedfrom the mother, while the third must beofpaternal origin (arrowhead). In fibroblasts, the amount of the paternal allele is reduced,reflecting the presence of mosaicism, with a normal 46,XX cellfibroblast line.

maternal chromosomes, which is fully compat-ible with the presence of maternal heterodis-omy (fig 4, other results not shown). However,in the absence of paternal DNA, this could notbe proven with absolute certainty.The parental origin of the translocated chro-

mosome 1 5q fragment was determined bymeans of polymorphic microsatellite markersin 1 5q26 to 1 5qter. Analysis of the polymor-phic marker FES, located in 15q26, detectedthree different alleles, both in skin fibroblastsand in peripheral white blood cells (fig 4). Oneof the alleles was not present in the mother andtherefore almost certainly represents the pater-nal allele. Similarly, for marker D15S107, alsoon distal 1 5q, an allele not present in themother was found in the patient. This is incontrast to the markers proximal to 15q24,where no alleles absent in the mother could befound (results not shown). In conjunction withthe cytogenetic and methylation studies, thesefindings are fully consistent with a paternal ori-

gin of the translocated fragment of distal chro-mosome 15q.Dosage analysis of the microsatellite markers

FES and D15S107 showed that the paternalallele was present in the majority ofwhite bloodcells, whereas in skin fibroblasts a lower dose ofthe paternal allele was found compared to thematernal alleles (fig 4). This is in agreementwith the cytogenetic findings, showing a partialtrisomy for distal 1 5q in the majority oflymphocytes, but only in approximately 50% offibroblasts (table 1).

INTERSTITIAL TELOMERE SEQUENCES AT THEBREAKPOINT JUNCTION SITESThe distal part of chromosome 1 5q was trans-located onto three different chromosomes,14q, 4q, and 16p. FISH, using a telomereprobe, showed the presence of two signals onall normal chromosomes, including the twochromosomes 15. On the chromosomes 14 and4, carrying the translocated 1 5q fragment,three signals were detected, two at thetelomeres and one interstitial signal (fig 5).These interstitial signals coincided with thejunction sites between the translocated distalchromosome 15q fragment and the transloca-tion partners. No metaphases with a 1 6p+could be analysed.

DiscussionThe patient reported here fulfils the diagnosticcriteria of PWS according to Holm et al,'2 withthe presence of all main clinical features andseven minor criteria. By conventional cytoge-netics, mosaicism consistent with a jumpingtranslocation of distal 15q and resulting in apartial trisomy for distal 15q was detected inboth lymphocytes and in fibroblasts. This rear-rangement must have occurred de novo as thekaryotypes of both parents were normal.Maternal uniparental heterodisomy for chro-mosome 15 was present, explaining the PWSphenotype. This was shown by methylationanalysis of the imprinted region on chromo-some 1 5ql 1 and further supported by theanalysis ofpolymorphic microsatellite markers,distributed along chromosome 15. On top of

397

on June 14, 2021 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.34.5.395 on 1 M

ay 1997. Dow

nloaded from

http://jmg.bmj.com/

Devriendt et al

Figure 5 FISH analysis with a telonmeric probe. Note that on7 the derivative chromosome14, with the extra distal 15q fragment, three signals are present, two at the ends of thechromosomes and one interstitial (indicated by arrow) at the translocation junction region.

2

47,XX, 1 5+ 46,XX,UPD1 5mat,der(1 4q+) 46,XX, U PD1 5mat,der(4q+)

3~~~~4\A -3 ^

46,XX,UPD1 5mat

Maternal chromosome 15

Paternal chromosome 15

Figure 6 Proposed nmechanism leading to maternal uniparental heterodisonmy and partialtrisomy of distal 15q. (1) Trisomy rescue event, with partial loss of chromosome 15 andtelonmeric translocation of distal chron2osonme 15 to chromosonie 14q. (2) Subsequenttranslocationi of the distal 15q fragnment from 14qter to the telomeres of chronlosomle 4q anld16p. (3) Loss of the distal 15q fragment fronm the telonmeres might lead to a euploid cell line.(4) A second trisomy rescue event with loss of the entire chromosome 15 is an alternativemechanism resulting in a euploid cell line in this patient.

this, the patient's karyotype showed partial tri-somy for distal 15q, and this could possiblyexplain why her mental retardation was moresevere than usually observed in PWS. Paternalinheritance for alleles on chromosome 15could only be shown for distal chromosome15q, and this, together with a maternal origin

of both intact chromosomes 15, is fullycompatible with a paternal origin of this trans-located chromosomal fragment. The mostlikely mechanism explaining these findings isshown in fig 6. Fertilisation by a normal malegamete of a disomic oocyte results in a zygotewith trisomy 15. These embryos are unviable,unless a postzygotic correction occurs. The lossof the paternal chromosome 15 results in aUPD 15 mat. Trisomy rescue is a recognisedmechanism leading to UPD.'7 lx Interestingly,whereas the correction usually involves the lossof the entire chromosome 15, a partial loss ofchromosome 15 occurred in this patient. Thedistal chromosome 1 5q fragment was retainedand translocated to other chromosomes, result-ing in partial trisomy for distal 15q. To ourknowledge, this is the first report showing thatin trisomy rescue both UPD and partialtrisomy can occur simultaneously.The distal 15q fragment in this patient was

present as a jumping translocation involvingthree different recipient partners, chromo-somes 14q, 4q, and 16p. In addition, a normal46,XX cell line without the partial trisomy wasfound, an unprecedented finding in patientswith a jumping translocation.3 " This cell linealso has a UPD 15 mat, since in skin fibroblasts,where approximately 50% of the cells have a46,XX karyotype, an exclusively maternalmethylation pattern was detected. The resultsof the microsatellite analysis are also fully con-sistent with this.There are several different possible explana-

tions for these observations. In a first possiblemechanism, translocation of the distal 1 5qfragment to another chromosome coincideswith the process of trisomy rescue (fig 6). Dur-ing subsequent cell divisions, this fragment isthen translocated to other chromosomes, thatis, a real jumping process (fig 6). Thismechanism has been proposed before.3 Theeuploid 46,XX cell line could be the result ofthe loss of the distal 1 5q chromosomalfragment during the jumping process (fig 6,step 3). Alternatively, an independent trisomyrescue event might have occurred in a differentcell line with the loss of an entire paternalchromosome 15 (fig 6, step 4) or after trisomyrescue, the distal chromosome 1 5q couldinitially remain as a free acentric chromosomefragment. During subsequent cell divisions andin different cells, the fragment could either belost or translocated to different chromosomes.

In this patient, the telomeric regions of thetranslocation partners appeared cytogeneti-cally intact. In addition, by means of FISH,interstitial telomeric sequences were shown atthe breakpoint junctions. This would imply aninitial translocation of the distal 15q fragmentto the telomere of one chromosome. The pres-ence of interstitial telomeres might renderthese derivative chromosomes unstable andprone to breakage at this site and recombina-tion with other telomeric sequences, as wassuggested before.4 The jumping translocationprocess would then not represent a recurrentreciprocal translocation but rather a transposi-tion of a chromosomal fragment from one telo-mere to another. Loss of the chromosome

398

on June 14, 2021 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.34.5.395 on 1 M

ay 1997. Dow

nloaded from

http://jmg.bmj.com/

J7umping translocation in Prader-Willi syndrome

15qter fragment, leading to a euploid cell line,would occur upon chromosome breakage atthe junction site without a simultaneousrecombination with another telomere. Furtherexperimental evidence to support this mech-anism could be gained from long term culturesof clonal cell lines carrying a jumping translo-cation, and the finding of either a loss of thefragment or transposition to the telomeres ofanother chromosome.Another proposed mechanism would require

that the distal chromosome 1 5q acentric chro-mosome fragment remains stable during sev-eral sequential cell divisions before being lostor translocated to the telomeres of differentchromosomes. It is difficult to envisage how achromosomal fragment without a centromereand missing a telomere at one end could bestably retained during several cell divisions.Therefore, we favour the first mechanism.

Interestingly, besides one patient with abreakpoint in chromosome 17q23, all sevenother reported patients with a telomeric trans-location have a breakpoint in chromosome15ql1-13 and have the PWS phenotype."6Involvement of the same chromosomal regionhas led to the suggestion that in this chromo-somal region specific DNA sequences must bepresent, with an affinity for recombination withtelomeres.4 6 On the other hand, in the twoother patients with a constitutional jumpingtranslocation investigated so far, interstitialtelomeric sequences were also found, as in thepresent case (case 1 of Park et al,5 case 3 ofRossi et at). The detection of interstitialtelomeres in the present patient with abreakpoint in 15q24 suggests that the jumpingprocess could be related to the presence ofinterstitial telomeric sequences and not merelyto the chromosomal region involved. Addi-tional studies are needed and, more specifi-cally, molecular cloning of the breakpointregion on 1 5q24 will be of particular interest inaddressing this question.

In conclusion, the present observation is afurther illustration of trisomy rescue leading touniparental disomy and shows the presence ofUPD in association with a partial trisomy. Asimilar observation has been made before in apatient with UPD16 mat and mosaic trisomyfor distal 16p.'9 For the first time, interstitialtelomere sequences were also found in aconstitutional jumping translocation involvinga chromosomal region outside 15ql 1-13.Therefore, we suggest that the jumping processis related to the presence of interstitialtelomeres.

We thank Reinhilde Thoelen for expert technical help and EricSchoenmakers for the FES probe. This work is supported by akrediet aan navorser, 1994, from the Nationaal Fonds voorWetenschappelijk Onderzoek of Belgium. Peter Marynen is aonderzoeksdirecteur and Gert Matthijs and Joris Vermeesch areaangesteld navorsers of the Nationaal Fonds voor Wetenschap-pelijk Onderzoek, Belgium.

1 Ledbetter DH, Riccardi VM, Airhart SD, Strobel RJ,Keenan BS, Crawford JD. Deletions of chromosome 15 asa cause of the Prader-Willi syndrome. N Engl J Med1981;304:325-9.

2 Nicholls RD, Knoll JHM, Butler MG, Karam S, Lalande M.Genetic imprinting suggested by maternal heterodisomy innon-deletion Prader-Willi syndrome. Nature 1989;342:281-5.

3 Rivera H, Zuffardi 0, Gargantini L. Non-reciprocal andjumping translocations of 15ql-qter in Prader-Willisyndrome. Am J Med Genet 1990;37:31 1-17.

4 Rossi E, Floridia G, Casali M, et al. Types, stability and phe-notypic consequences of chromosome rearrangementsleading to interstitial telomeric sequences. 7 Med Genet1993;30:926-31.

5 Park VM, Gustashaw KM, Wathen TM. The presence ofinterstitial telomeric sequences in constitutional chromo-some abnormalities. Am J Hum Genet 1992;50:914-23.

6 Reeve A, Norman A, Sinclair P, et al. True telomeric trans-location in a baby with the Prader-Willi phenotype. Am J7Med Genet 1993;47:1-6.

7 Jauch A, Robson L, Smith A. Investigations with fluores-cence in situ hybridisation (FISH) demonstrate loss of thetelomeres on the reciprocal chromosome in three unbal-anced translocations involving chromosome 15 in thePrader-Willi and Angelman syndromes. Hum Genet 1995;96:345-9.

8 Robinson WP, Wagstaff J, Bernasconi F, et al. Uniparentaldisomy explains the occurrence of the Angelman orPrader-Willi syndrome in patients with an additional smallinv dup(1 5) chromosome. J Med Genet 1993;30:756-60.

9 Toth-Fejel S, Olson S, Gunter K, et al. The impact ofimprinting: Prader-Willi syndrome resulting from chromo-some translocation, recombination and nondisjunction.AmJrHum Genet 1996;58:1008-76.

10 Lejeune J, Maunoury C, Prieur M, Van den Akker J. Trans-location sauteuse (5p;15q), (8q;15q), (12q;15q). AnnGenet (Paris) 1979;22:210-13.

11 Farrell SA, Winsor EJT, Markovic VD. Moving satellites andunstable chromosome translocations: clinical and cytoge-netic implications. AmJMed Genet 1993;46:715-20.

12 Holm VA, Cassidy SB, Butler MG, et al. Prader-Willisyndrome: consensus diagnostic criteria. Pediatrics 1993;91:398-402.

13 Pinkel D, Straume T, Gray JW. Cytogenetic analysis usingquantitative, high-sensitivity, fluoresence hybridisation.Proc NatlAcad Sci USA 1986;83:2934-8.

14 Vermeesch JR, De Meurichy W, Van Den Berghe H,Marynen P, Petit P. Differences in the distribution andnature of the interstitial telomeric (TTAGGG)n sequencesin the chromosomes of the Giraffidae, okapi (OkapiJohnstoni), and giraffe (Giraffa camelopardalis): evidencefor ancestral telomeres at the okapi polymorphic rob (4;26)fusion site. Cytogenet Cell Genet 1996;72:310-15.

15 Dittrich B, Robinson WP, Knoblauch H, et al. Moleculardiagnosis of the Prader-Willi and Angelman syndromes bydetection of parent-of-origin specific DNA methylation in15qI 1-13. Hum Genet 1992;90:313-15.

16 Driscoll DJ, Waters MF, Williams CA, et al. A DNA meth-ylation imprint, determined by the sex of the parent, distin-guishes the Angeiman and Prader-Willi syndromes. Ge-nomics 1992;3:917-24.

17 Cassidy SB, Lai LW, Erickson RP, etal. Trisomy 15 with lossof the paternal 15 as a cause of Prader-Willi syndrome dueto maternal disomy. Am J Hum Genet 1992;51:701-8.

18 Engel E. La disomie uniparentale: revue des causes et con-sequences en clinique humaine. Ann Genet (Paris)1995;38:113-16.

19 Schinzel A, Kotzot D, Brecevic L. An unusual consequence ofmaternal uniparental disomy 16. Genet Counsel 1996;7:87.

399

on June 14, 2021 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.34.5.395 on 1 M

ay 1997. Dow

nloaded from

http://jmg.bmj.com/