86Med Genet 1998;35:836-840

Mosaic supernumerary ring chromosome 19identified by comparative genomic hybridisation

Saeed R Ghaffari, Elizabeth Boyd, J Michael Connor, Alison M Jones, J L Tolmie

AbstractWe report the use of comparative genomic

hybridisation (CGH) to define the originof a supernumerary ring chromosomewhich conventional cytogenetic banding

and fluorescence in situ hybridisation(FISH) methods had failed to identify.

Targeted FISH using whole chromosome19 library arm and site specific probesthen confirmed the CGH results. This

study shows the feasibility of using CGHfor the identification of supernumerary

marker chromosomes, even in fewer than

50% of cells, where no clinical or cyto- .i

genetic clues are present.

(JMed Genet 1998;35:836-840)

Keywords: chromosome 19; comparative genomichybridisation; ring chromosome; mosaicism

Institute ofMedicalGenetics, YorkhillHospitals Campus,Yorkhill, GlasgowG3 8SJ, UKS R GhaffariE BoydJ M ConnorJ L Tolmie

Merchiston Hospital,Brookfield, ByJohnstone PA5 8TY,UKAM Jones

Correspondence to:Dr Boyd.

Received 13 November 1997Revised version accepted forpublication 13 March 1998

The origin of supernumerary marker chromo-somes cannot always be resolved by means ofconventional banding techniques.' When thereare cytogenetic or clinical clues to the chromo-some abnormality, FISH using appropriateprobes may identify the origin of the markerchromosome, otherwise a stepwise approachusing different FISH probes or reverse chro-mosome painting,2-5 possibly combined withchromosome microdissection, is required."'However, when the marker chromosome isvery small, or has no significant band land-mark, or is present in a low percentage of cellsonly, even FISH may fail to identify the abnor-mality. Both FISH and microdissection alsodepend on good quality metaphases with highmitotic index, conditions that may be difficultto obtain. Therefore, a genome wide screeningtechnique is desirable.

Comparative genomic hybridisation (CGH)is a genome wide screening technique that wasfirst developed to study chromosome abnor-malities in tumours. 1-13 It has also been used asan adjunct to conventional cytogenetics for thediagnosis of chromosome imbalance in clinicalcytogenetics."'6 However, there are no reportsof CGH being used as a tool to diagnose theorigin of the extra DNA material in patientswho have low percentage mosaicism for asmall, supernumerary, marker chromosome.Here we report a 71 year old woman with

mental retardation and a supernumerarymarker chromosome, the genetic origin ofwhich was not detected despite attempts usingconventional cytogenetics and, latterly, FISHusing centromeric probes for a number ofchromosomes. In contrast, CGH readilyshowed its origin from chromosome 19 and

Figure 1 The proband.

complementary experimentsfirmed the CGH results.

by FISH con-

Materials and methodsCASE REPORTThe proband (fig 1), a 71 year old woman withmental handicap, had been in residential carefor 55 years. No details of her birth history orantenatal life were available. The familyrecalled her walking at the age of 2 years andthe development of speech was delayed untilthe age of 4 or 5 years. She was not clumsy butalways had some difficulty with fine motorskills. She attended a private convent school inearly childhood but upon her mother's death,when the proband was aged 16 years, she wasreceived into care and has resided in institu-tions ever since. The proband was the ninthchild born in a family of 10. In the family, nosib was similarly affected to the proband andthere was no history of mental or physicalillness. In 1976 the proband was admitted tohospital for treatment for hypertension andimpaired renal function of unknown origin. Atthat time, she also had septic skin lesions andosteoarthritis affecting the hips. At present, sheis able to feed, wash, and dress herself (withsome assistance), handle toys, obey simplecommands, and guard against physical dan-gers. Speech is mostly indistinct but she canutter short simple sentences that convey mean-ing. She is able to write a few words and canalso read several words, but is notably skilled atcompleting jigsaw puzzles. It is thought osteo-arthritis has led to loss of some fine motorskills. The proband has never been independ-ent or able to go out on her own but she enjoyssocial events. She has urinary and bowel conti-nence.

836 on July 21, 2021 by guest. P

rotected by copyright.http://jm

g.bmj.com

/J M

ed Genet: first published as 10.1136/jm

g.35.10.836 on 1 October 1998. D

ownloaded from

Mosaic supernumerary ring chromosome 19

Physical examination showed a height of 132cm, head circumference of 52.5 cm, bloodpressure of 180/110 mm Hg, poor peripheralcirculation, early hypertensive changes in bothoptic fundi, poor vision, upward slantingpalpebral fissures, large tongue, protrudinglower lip, normal palmar creases, normalfemale genitalia, and deep set toe nails. She hasreduced muscle power, a shuffling gait, bilat-eral pes planus, and radiological signs ofsubluxed, osteoarthritic hip joints. An x ray in1994 was reported to show considerabledegenerative changes in the cervical spine withgross osteophyte formation and partial, pre-sumed congenital, fusion of vertebral bodiesC3 and C4. Recently, renal function hasdeclined with recurrent urinary infections andher right kidney is non-functioning.

Detailed psychological examinations wereperformed and the highest estimate of intellectsuggested moderate to mild learning disability.However, the British Picture Vocabulary Scaleresult and the examiners' impressions indi-cated that moderate to severe learning disabil-ity was present. Notably, although there wassome evidence of behavioural and intellectualdeterioration, there were no psychological orpsychiatric difficulties in an emotionally stablewoman who has good interpersonal relation-ships.

In the proband's residence, it was assumedthat she was affected by Down's syndrome butthe true karyotypic abnormality was discoveredin 1976 when a blood sample was first submit-ted for chromosomal analysis.

CHROMOSOME PREPARATIONSMetaphase spreads were prepared from phyto-haemagglutinin (PHA) stimulated peripheralblood lymphocytes of the patient (for FISH) orhealthy males (for CGH studies) using stand-ard procedures of hypotonic treatment andmethanol/acetic acid fixation (3:1 v/v).

CGH AND DIGITAL IMAGE ANALYSISGenomic DNA probe preparation, labellingprocedures, hybridisation, and posthybridisa-tion washings were carried out as describedpreviously.'6 Briefly, test and control DNA wasextracted by proteinase K and RNase diges-tions. Control genomic DNA was preparedfrom blood of healthy females (46,XX). Test(patient) human genomic DNA was directlylabelled with FITC-12-dUTP (DuPont) andcontrol DNA was labelled with Texas Red-5-dUTP (DuPont) by standard nick translationreaction; DNase I concentration was adjustedto result in an average fragment size of500-1000 bp. One microgram of fluoresceinlabelled test DNA, 1 jg of Texas Red labelledreference, and 50 gg of unlabelled humanCot-i DNA (Gibco) were cohybridised to nor-mal female metaphase spreads. Hybridisationwas allowed to proceed for 48 hours at 37°C ina moist chamber. After posthybridisationwashings, images for CGH analysis wereacquired using an epifluorescence microscope(Axioplan, Zeiss, Germany) equipped with acooled CCD camera (Photometrics) control-led by an image analysis system (Digital Scien-

tific, Cambridge, UK). Green, red, and bluefluorescence images were captured from eachhigh intensity, uniformly hybridised metaphaseand were analysed as separate grey scaleimages. The image representing the blue DAPIcounterstain was inverted and used for chro-mosome identification based on its coarsebanding pattern. The mean of the individualratio profiles of 20 metaphase spreads was cal-culated. For normalisation of the ratio profiles,the model value of the green to red ratio for theentire metaphase was set to 1.0. Finally, theindividual ratio profiles were displayed next tothe chromosome diagrams.

FLUORESCENCE IN SITU HYBRIDISATIONIn an earlier attempt to identify the supernu-merary chromosome, FISH to metaphasechromosomes was carried out using a numberof commercially available centromeric probes(chromosomes 4, 12, 20, 22, 13/21, and1/5/19). After the CGH results were obtained,targeted FISH was carried out using wholechromosome 19 paint (Oncor), 19 centromericprobe, 19p and 19q arm paints, and telomerespecific probes for l9p and 19q(LI), accordingto the standard FISH procedures.

ResultsCONVENTIONAL CYTOGENETIC ANALYSIS (FIG 2A)Cytogenetic studies by G banding in 1976showed the presence of two cell lines. Eighteenout of 25 cells showed a 46,XX karyotype whileseven out of 25 (28%) showed an additionalmarker chromosome of unknown origin(46,XX/47,XX,+marker). Chromosomeanalysis by G banding in 1986 confirmed thefinding of two cell lines. Twenty-one out of 25cells examined showed an apparently normalfemale karyotype while the remaining four cells(16%) showed a count of 47 including oneadditional small chromosome. This markerappeared to be a ring chromosome, approxi-mately halfthe size of a chromosome 21, whichcontained at least one G band positive area (fig2A). C banding showed one centromere to bepresent, and Ag NOR staining was negativewith respect to the ring chromosome. No othermembers of the family underwent cytogeneticinvestigation.

FIRST SERIES OF FISH STUDIESFISH was carried out using centromericprobes for chromosomes 4, 12, 20, 22, 13/21,and 1/5/19 empirically. The studies failed todetect the origin of the extra chromosome andno further probes were investigated becausedifficulty was experienced in identifying themarker with certainty within the metaphases.

CGH ANALYSIS (FIG 2B)CGH was carried out using a DNA samplefrom the patient. Analysis using a very restric-tive threshold with a fluorescence ratio of 0.5-1.5 (which requires 100% abnormal cells todetect any gain or loss of the DNA material)failed to detect any abnormality. Progressivelyless restrictive thresholds were then used and,surprisingly, a threshold of 0.8-1.20 (whichtheoretically detects abnormal cells comprising

837 on July 21, 2021 by guest. P

rotected by copyright.http://jm

g.bmj.com

/J M

ed Genet: first published as 10.1136/jm

g.35.10.836 on 1 October 1998. D

ownloaded from

Ghaffari, Boyd, Connor, et al

Ie.f.*::I

v

i.P

.+t I:j * !P .i

:

.~~~~~~~~~~~~~~~.i^,4 4 is-

z4/

4 .: /k

I*3s

Figure 2 Conventional cytogenetic, CGH, and targeted FISH results. (A) G banded metaphase spread of the patient showing a supernumerary ringchromosome (arrowed). (B) Standard ratio profile ofchromosome 19 is shown above the chromosome idiogram. The horizontal lines represent the balancedstate of the chromosomal copy number (middle line, ratio 1. 0) and the lower and upper thresholds applied to detect chromosomal losses and gains,respectively. These thresholds correspond to ratios of 0. 8 and 1. 20, which are the theoretical values expectedfor a monosomy or trisomy present in 40% ofdiploid cells. (C, D) FISH to metaphase chromosome from the patient using chromosome 19p (C) and 1 9q arm specific painting probes (D).

40% of the total) showed a gain in DNA mate-rial corresponding to 1 9q (fig 2B). In CGH thecentromeric areas are strongly suppressed byunlabelled human Cot-i DNA, so gain or lossof this area could not be interpreted reliably.

FISH CONFIRMATION OF THE CGH RESULTS (FIG2C, D)Further FISH studies using a whole chromo-some 19 paint (WCP19, Oncor), 19p and 19qarm specific paints, and telomeric probes for19p and 19q (LI) and 19 centromeric probeconfirmed the CGH finding. The ring chromo-some hybridised to the whole chromosome 19and 19q arm specific paints and chromosome19 centromeric probe but not to l9p arm spe-cific paint or l9p or 19q telomeric probes. Itwas thus concluded that the proband wastrisomic for almost the entire long arm of chro-mosome 19 with only the telomeric regionbeing absent from the ring.

DiscussionTo date, only four cases with a supernumeraryring 19 have been reported.2 17 18 Althoughthere are 11 reported cases with partial 19qtrisomy,9 26 none is directly comparable to our

case since in each instance non-homologouschromosomes were involved and the patternsof clinical abnormalities which resulted werequite variable.Among the four cases with supernumerary

ring 19, one, a young infant with macrocephalyand developmental delay, is not directlycomparable as metaphase cells in that casecontained dicentric marker chromosome 19 aswell as another marker chromosome, not 19.2Among the remaining three cases, the firstcase2 was a newborn infant aged 3 weeks withfailure to thrive, hypotonia, and pneumoniawith a marker chromosome 19 in 50% of theperipheral lymphocytes examined. The thirdcase with supernumerary r(19)" had a verysmall supernumerary r(19) present in 46% ofumbilical cord blood cells and estimated to beabout 1/6 the size of chromosome 21. This wasidentified after amniocentesis was performedfor advanced maternal age. The normal maleinfant weighed 3700 g at birth and he had nor-mal growth and development when evaluatedaged 18 months. In another case, Quack et al17reported a non-mosaic, supernumerary r(19)present in a boy with somatic overgrowth andmental retardation. Unusually, this boy's

838

I -.0-

qgppr

on July 21, 2021 by guest. Protected by copyright.

http://jmg.bm

j.com/

J Med G

enet: first published as 10.1136/jmg.35.10.836 on 1 O

ctober 1998. Dow

nloaded from

Mosaic supernumerary ring chromosome 19

mother carried the same ring 19 but had a bal-anced karyotype since one other chromosome19 had an interstitial deletion of the long armsegment of chromosome 19 which formed thering. Thus, the ring chromosome was shown tobe derived from the deleted chromosome afterthe occurrence of two breaks, one in thecentromere region, the other in band q13.2 inthe long arm of chromosome 19. There are fewclinical resemblances between our elderlypatient with ring 19 mosaicism and thenon-mosaic child reported by Quack et all'7;neither patient has any major congenitalmalformation, both have moderate mentalhandicap, but the striking difference is that ourelderly patient has reduced stature and propor-tionately reduced head circumference, whereasthe subject of the report of Quack et all'7 haslength, weight, and head circumference morethan 3 SD above the mean. Quack et all'7estimated that the IQ of their proband was 69and reported particular difficulties with lan-guage, which is reminiscent of the findings inour patient whose psychological evaluationsshowed no evidence of a behavioural pheno-type and only some evidence of age relatedgradual decline in intellect. In summary, inthree cases non-specific moderate to mildmental retardation or developmental delay hasbeen caused by mosaicism for an abnormal,supernumerary chromosome 19.So far the approach to determining the

origin of cytogenetically unidentifiable chro-mosomes has involved stepwise FISH usingforward painting methods with alpha satellite,single copy, or whole chromosome paintingprobes, or reverse chromosome painting proce-dures with either flow sorting or microdissec-tion of the particular chromosome.'" Micro-dissection followed by DOP-PCRamplification of the dissected DNA is increas-ingly being used and its modified approachesare reported to be so efficient that just one dis-sected DNA fragment is sufficient for asuccessful analysis.9 However, not all diagnos-tic cytogenetics laboratories have easy access tomicrodissection technology. Here we presentan alternative approach which is not dependenton patient metaphase and is a genome widescreening technique enabling the result to beachieved in just one experiment. Many imageanalysis systems now in use in cytogeneticlaboratories include software for CGH, andresults of analysis are available in three workingdays, which is comparable to the time requiredfor microdissection. Where further confirma-tion by FISH is necessary one more day isrequired.An advantage of CGH over microdissection

is that, as shown here, it can provide anestimate of the proportion of cells containingthe ring chromosome in the tissue studied.This information is laborious to obtain bychromosome analysis or interphase FISH. Theminimum percentage of abnormal cells thatcan be detected by CGH has not yet beendefined, but based on this report and ourunpublished data, as little as 15% mosaicism incultured metaphase cells could be detected.Theoretically, using a medium restrictive fluo-

rescence ratio of 0.75-1.25, there should be atleast 50% abnormal cells present for the chro-mosome imbalance to be detected.'7-29 In thepresent case, fluorescence ratio 0.8-1.20 wasthe most restrictive threshold at which theabnormality was detected, which leads to anestimation of the percentage of abnormal cellsin uncultured blood of at least 40%. This pro-portion differs from the result of cytogeneticanalysis in cultured cells which is based on asmall sample of metaphases obtained fromPHA stimulated T lymphocytes, and whichshowed not more than 28% abnormal cells. Wecurrently use CGH to determine the origin ofcytogenetically unidentifiable chromosomesand so far have obtained successful results in10 cases with different sizes and percentages ofring and marker chromosomes (unpublisheddata). However, since in CGH the centromericregions are suppressed, the method would notbe applicable to ring or marker chromosomescomposed almost entirely of centromericheterochromatin.The classical explanation for formation of a

centric ring involves breaks in both arms of thechromosome with subsequent fusion of theproximal ends to form a ring and loss of thematerial distal to the breakpoints. Morerecently the possibility ofU type fusions or a Utype fusion in combination with transversemisdivision of the centromere has been putforward as a mechanism for the generation ofsmall ring chromosomes.3" In the case de-scribed here, no short arm material could beseen in the ring chromosome either by CGH orby FISH using short arm paint, suggesting abreakpoint at or immediately adjacent to thecentromere. No more detailed studies have yetbeen undertaken to define this breakpointmore precisely. CGH studies indicate that vir-tually the entire long arm of chromosome 19 ispresent in the ring, although by FISH the telo-meric region is shown not to be present. Wehave previously shown that partial trisomy orpartial monosomy of as little as 4-5 Mb can bedetected in telomeric regions by CGH.'6 Thefact that CGH has not shown any variation incopy number of any part of 19q suggests thatthe breakpoint is within 5 Mb of the telomere,and that the loss of long arm material in theformation of the ring chromosome is beyondthe resolution of CGH. However, although wehave precisely identified the identity of the ringchromosome, we are not able to distinguish themechanism by which it arose.The first series of FISH experiments using

empirically selected centromeric probes in-cluding one for chromosome 19 failed to indi-cate the origin of microdissection. This may beexplained in part by the low percentage of cellscontaining the ring and difficulties in identify-ing it with certainty in some mitoses. Also thealpha satellite probe hybridising to chromo-somes 1, 5, and 19 was used in the study, sinceat that time it was the only availablecentromeric probe for chromosome 19. It ispossible that a small change in stringency con-ditions caused some loss of signal for chromo-some 19.

839 on July 21, 2021 by guest. P

rotected by copyright.http://jm

g.bmj.com

/J M

ed Genet: first published as 10.1136/jm

g.35.10.836 on 1 October 1998. D

ownloaded from

Ghaffari, Boyd, Connor, et al

As discussed above, when no cytogenetic orclinical clues are available, a multicolour orstepwise FISH approach with or without chro-mosome microdissection may be used to iden-tify the origin of a marker chromosome. How-ever, in investigating small markerchromosomes, many service laboratories atpresent carry out sequential experiments usingFISH probes that are selected empirically. Ourexperience indicates that in FISH studies theempirical approach may not detect a markerthat is present in a lower percentage ofmetaphases, even though it is detected subse-quently in a targeted FISH study. Although thisexperience is unlikely to be a frequent occur-rence, it is one that is important for clinicalcytogenetic practice.

We thank colleagues in the cytogenetics laboratories at theDuncan Guthrie Institute for preparing and examining manyconventional cytogenetic preparations from the patient. We alsothank Jackie Squire and Keith Bonham for expert psychologicalevaluations. SRG received a scholarship from the Ministry ofHealth and Medical Education, Iran.

1 Buckton KE, Spowart G, Newton MS, Evans HJ. Forty-fourprobands with additional "marker" chromosome. HumGenet 1985;69:353-70.

2 Crolla JA, Dennis NR, Jacobs PA. A non-isotopic in situhybridisation study of the chromosomal origin of 15 super-numerary marker chromosomes in man. J Med Genet1992;29:699-703.

3 Blennow E, Telenius H, Larson C, et al. Completecharacterization of a large marker chromosome by reverseand forward chromosome painting. Hum Genet 1992;90:371-4.

4 Carter N, Ferguson-Smith MA, Perryman M, et al. Reversechromosome painting: a method for rapid analysis of aber-rant chromosomes in clinical cytogenetics. J Med Genet1992;29:299-307.

5 Magnani I, Sacchi N, Darfier M, Nisson P, Tornaghi R,Fuhrmann-Conti A. Identification of the chromosome 14origin of a C-negative marker associated with a 14q32deletion by chromosome painting. Clin Genet 1993;43: 180-5.

6 Lengauer C, Eckelt A, Weith A, et al. Painting of definedchromosomal regions by in situ supression hybridization oflibraries from laser-microdissection chromosomes. Cy-togenet Cell Genet 199 1;56:27-30.

7 Deng H, Yoshura K, Dirks R, et al. Chromosome-band-specific painting; chromosome in situ suppression hybridi-zation using PCR products from a microdissected chromo-some band as a probe pool. Hum Genet 1992;89:13-17.

8 Ohta T, Tohma T, Soejima H, et al. The origin of cytologi-cally unidentifiable chromosome abnormalities: six casesascertained by targeted chromosome-band painting. HumGenet 1993;92:1-5.

9 MullerNavia J, Nebel A, Schleiermacher E. Complete andprecise characterization of marker chromosomes byapplication of microdissection in prenatal diagnosis. HumGenet 1995;96:661-7.

10 De A, Coelho KEF, Egashira M, et al. Diagnosis of fourchromosome abnormalities of unknown origin by chromo-some microdissection and subsequent reverse and forwardpainting. Am J Med Genet 1996;63:468-7 1.

11 Kallioniemi A, Kallioniemi OP, Sudar D, et al. Comparativegenomic hybridization for molecular cytogenetic analysis ofsolid tumors. Science 1992;258:818-21.

12 Kallioniemi OP, Kallioniemi A, Piper J, et al. Optimizingcomparative genomic hybridization for analysis of DNAsequence copy number changes in solid tumors. GenesChrom Cancer 1994;10:231-43.

13 Du Manoir S, Speicher MR, Joos S, et al. Detection of com-plete and partial chromosome gains and losses bycomparative genomic in situ hybridization. Hum Genet1993;90:590-610.

14 Bryndorf T, Kirchhoff M, Rose H, et al. Comparativegenomic hybridization in clinical cytogenetics. Am J HumGenet 1995;57:1211-20.

15 Erdel M, Duba HC, Verdorfer I, et al. Comparative genomichybridization reveals a partial de novo trisomy 6q23-qter inan infant with congenital malformations: delineation of thephenotype. Hum Genet 1997;99:596-601.

16 Ghaffari SR, Boyd E, Tolmie JL, Crow YJ, Trainer AH,Connor JM. A new strategy for cryptic telomeric transloca-tion screening in patients with idiopathic mental retarda-tion. _Med Genet 1998;35:225-33.

17 Quack B, Van RN, Verschraegen Spae MR, Klein F.Interstitial deletion and ring chromosome derived from19q. Proximal 19q trisomy phenotype. Ann Genet 1992;35:241-4.

18 Michalski K, Rauer M, Williamson N, Perszyk A, Hoo JJ.Identification, counselling, and outcome of two cases ofprenatally diagnosed supernumerary small ring chromo-somes. Am J Med Genet 1993;46:88-94.

19 Lange M, Alfi OS. Trisomy 19q. Ann Genet 1976;19:17-21.20 Schmid W Trisomy for the distal part of the long arm of

chromosome 19 in brother and sister. Hum Genet 1979;46:263-70.

21 Pangalos A, Ghica M, Couturier J. Trisomy 1 9q associatedwith partial trisomy 22q in two members of the same fam-ily. Analysis using the thymidine and RBA bandingtechniques. Proceedings of the 6th international Congress ofHuman Genetics, Jerusalem, 1981, P171.

22 Zonana J, Brown MG, Magenis RE. Distal 1 9q duplication.Hum Genet 1982;60:267-70.

23 Rivas F, Garcia-Cruz D, Rivera H, Plascencia ML, GonzalezRM. 19q distal trisomy due to a de novo (19;22)(ql3.2;p1 1) translocation. Ann Genet 1985;28:113-15.

24 Madokoro H, Ohdo S, Sonodo T, Kawaguchi K, Ohba K.Partial trisomy for 19q due to paternal 17/19 reciprocaltranslocation. Am JHum Genet 1988;33:61-5.

25 Boyd E, Grass FS, Park JC, Knutson K, Stevenson RE.Duplication of distal 19q: clinical report and review. Am 7Med Genet 1992;42:326-30.

26 Valerio D, Lavorgna F, Scalona M, Conte A. A new case ofpartial trisomy 19q (ql3.2-*qter) owing to an unusualmaternal translocation. J Med Genet 1993;30:697-9.

27 Du Manoir S, Schrock E, Bentz M, et al. Quantitative analy-sis of comparative genomic hybridization. Cytometry 1995;19:27-41.

28 Piper J, Rutovitz D, Sudar D, et al. Computer image analysisof comparative genomic hybridization. Cytometry 1995;19:10-26.

29 Du Manoir S, Kallioniemi OP, Lichter P, et al. Hardwareand software requirements for quantitative analysis of com-parative genomic hybridization. Cytometry 1995;19:4-9.

30 Callen DF, Eyre HJ, Ringenbergs ML, Freemantle CJ,Woodroffe P, Haan EA. Chromosomal origin of small ringmarker chromosomes in man: characterization by molecu-lar genetics. Am J Hum Genet 199 1;48:769-82.

840 on July 21, 2021 by guest. P

rotected by copyright.http://jm

g.bmj.com

/J M

ed Genet: first published as 10.1136/jm

g.35.10.836 on 1 October 1998. D

ownloaded from