Hum Genet (1986) 74: 270-274

© Springer-Verlag 1986

Development of additional RFLP probes near the locus for Duchenne muscular dystrophy by cosmid cloning of the DXS84 (754) locus M. H. Hofker 1, G.J.B. van Ommen 1, E. Bakker ~, M. Burmeister 2, and P. L. Pearson 1 1Department of Human Genetics, Sylvius Laboratories, State University of Leiden, P.O. Box 9503, NL-2300 RA Leiden, The Netherlands 2European Molecular Biology Laboratory, Meyerhofstrasse 1, D-6900 Heidelberg, Federal Republic of Germany

Summary. We have isolated 70 kb of sequences surrounding probe 754 (DXS84), linked with Duchenne muscular dystrophy. In addition to the original PstI RFLP detected by 754, BglII and EcoRI RFLPs were detected with the single copy subclone 754.11 and a HindIII RFLP with the subclone 754.6. The BglII and HindIII and HindIII RFLPs both have minor allele frequencies of 40%, as in PstI polymorphism. The EcoRI polymorphism has a minor allele frequency of 23%. Since a linkage disequilibrium is observed between these RFLPs (P < 0.001), the BglII and the HindIII RFLPs do not contribute to the heterozygosity. However, the minor allele of the EcoRI RFLP segregates exclusively with the major haplotype of the PstI-BglII-HindIII complex, and consequently 47% of the homozygotes for the haplotype become heterozygous. As a result, the overal heterozygote frequency of the DXS84 locus increases from 50% to 65%.

Introduction Probe 754, isolated in our laboratory (HoNer et al. 1985) is one of the closest proximal probes to the Duchenne muscular dystrophy (DMD) locus. It detects a PstI polymorphism with a minor allele frequency of 40%. Our own current estimate of the genetic distance between DMD and 754 is 5 cM with a Lod of +12.44 (95% confidence limits 0.01-0.12) (Bakker et al. 1986) and the pooled data for several collaborators gives a peak Lod score at 11 cM (Goodfellow et al. 1985). Physically it maps proximal to those DMD-related Xp21 translocation breakpoints (HoNer et al. 1985; Francke et al. 1985; Davies et al. 1985), on which it has been tested, and into an interstitial deletion of 2-5 million basepairs in Xp21, present in a male patient with DMD, retinitis pigmentosa, chronic granulomatous disease and McLeod syndrome (Francke et al. 1985). Subsequently, 754 was found to map into several other independent Xp21 deletions, associated with glycerol kinase deficiency, adrenal hypoplasia and DMD (Wieringa et al. 1985; Patil et al. 1985). The value of 754 for family studies and the key role it has played in opening up the DMD region for high resolution deletion mapping has prompted us to clone the sequences adja-

cent to 754 and to search for additional RFLP probes linked to it. In the present study we describe the isolation of three cosmid clones with 754, together covering 70kb, the detection of three new RFLPs, using subcloned probes, and the study of their contribution to the increase of the heterozygosity of the 754 (DXS84) locus in the Dutch population.

Materials and methods Cosmid libraries

One library was constructed by the ligation of size-fractionated partial Mbo I-digested female D N A into the vector c2RB (Bates and Swift 1983). In vitro packaging in phage lambda particles and transfection into Escherichia coli 1046 was carried out essentially as described by van Ommen et al. (1983). An efficiency of 2 x 106 clones per gg D N A was obtained. The primary library was grown at a density of 40,000 colonies per 90 mm dish. Duplicate filters were prepared according to the method of Hanahan and Meselson (1980) and hybridized with the labelled insert of 754, using standard conditions (see below). The second library was prepared by Dr. A. M. Frischauf (EMBL, Heidelberg). Its construction in pCOS2 EMBL, allowed screening by homologous recombination in E. coli as described (Poustka et al. 1984). Restriction maps

Two approaches have been employed. The cosmids were first linearized with SalI, followed by partial digestion with different enzymes. The samples were run on 0.35% agarose gels, transferred to duplicate Genescreen filters by sandwich blotting. Each filter was hybridized, either with the labelled SalI/ PstI or with the SalI/BamHI fragment of pAT153, thus detecting sets of partials, from both ends inwards, in a complementary fashion; this is shown schematically in Fig. 1. The EcoRI and PstI maps of the pCOS2 EMBL cosmid were obtained by linearizing at its cohesive sites by lambda terminase cleavage, partially digestion with restriction enzymes and hybridization with end-labelled CosL or CosR sequences (Rackwitz et al. 1985). Genomic D N A

Offprint requests to: M. H. HoNer, Department of Human Genetics, Sylvius Laboratories, State University of Leiden, P.O. Box 9503, NL-2300 RA Leiden, The Netherlands

The test panel for polymorphism screening consists of D N A prepared from 50 randomly obtained Dutch placentas. For

271

jQ-

p A T 153

S

)~2 = (P11P22 - PI2P21)2N/(Ptl 4- P12)(P21 + P22) (Pll 4- P21)(P12 + xP22)

°B

The value of Z 2 is related to the correlation coefficient "r" of the association between a pair of polymorphic sites and whose value is intuitively easier to use than either delta or Z2:

E

E

E

E

i

I

I

I

S ,

r = ( Z 2 / N 4- X2) 1/2

Results insert

C 2 RB

Fig. 1. Schematic representation of the restriction site mapping procedure. SalI/BamHI and SalI/PstI fragments of pAT153 hybridize to either side of the C2RB cosmid clone, linearized with SalI (S). The cosmid DNA is partially cleaved with a particular restriction enzyme (E). One set of partials, showing the order of the restriction sites from the SalI site towards the middle of the cosmid clone, can be revealed as shown in the diagram and in Fig. 2

family studies, blood was sampled from patients and their relatives and used for D N A isolation. Clone 2D (C12D) is a human/hamster hybrid cell line containing a single human Xchromosome and was kindly provided by Dr. S. Goss, Dunn School of Pathology, Oxford. Southern blot analysis

Genomic D N A isolations were performed as described ( H o N e r et al. 1985). The D N A was digested with various restriction enzymes obtained from Promega Biotec or Boehringer Mannheim. B g l l I was prepared in our own laboratory. Incubations were carried out with 3 U/~tg in medium salt buffer (Maniatis et al. 1982) except for BglI1 (10 m M Tris-HC1, 25 m M MgCI2, 60 mM NaC1, 1 m M mercaptoethanol, p H 9.5) and TaqI (High Salt, 1). A n amount of 7 gg of the D N A was electrophoresed through 0.7% agarose in T E A buffer (Manintis et al. 1982) overnight at 1V/cm. Gels were denatured for 2 × 15 min in 0.4 M NaOH/1.5 M NaC1 and transferred in the same solution to Genescreen plus (New England Nuclear). Probes were labelled with cd2pdCTP to a specific activity of 5 x 10s dpm/Iag, using a Nick Translation kit (Amersham). The hybridization and washing conditions were as described by van Ommen et al. (1983).

Isolation and characterization o f the cosmids

From the 360,000 clones of the human female cosmid library in vector c2RB, we obtained two almost completely overlapping cosmids, CX754 and CX754-8. HX754 was found in the male cosmid library in pCOS2 EMBL using homologous recombination screening (Poustka et al. 1984) with 754 in pUC12. First, restriction maps were made via a protocol generating ladder patterns of partials from both ends inwards, which are interpretable like D N A sequence gels (Fig.2). This "restriction site sequencing" is an extremely rapid mapping procedure, essentially similar to the cosmid mapping method developed by Rackwitz et al. (1985). With this procedure several cosmids can be mapped in one experiment. The restriction site sequencing protocol is particularly suitable for mapping cosmids with genomic inserts. Other methods, such as crossblotting or hybridization with gel-purified D N A fragments, are sometimes difficult to interpret, due to the hybridization of repetitive sequences to the different parts of the clones. In this approach, some positional uncertainty remains with sites in the middle of each cosmid. Further refinement of the maps was obtained by hybridization of

Calculations

The measure of disparity between the observed and expected haplotype frequencies is given by the linkage disequilibrium parameter Delta (Vogel and Motulsky 1979): Delta = (P22/N) 1/2 - [(P12 + P22)(P21 4- P22)/N2] 1/2 where Pn, Plz, P21, and P22 are the observed numbers of the + +, 4- - , - +, - - haplotypes respectively and where N is the total number of chromosomes observed. A " + " has been used for the most frequent or major allele and a " - " has been used for the less frequent or minor allele. The value of Delta is zero at complete equilibrium between the alleles. To test wether Delta deviates significantly from zero in case of a linkage disequilibrium, a 2 x 2 table ~2 test has been applied as follows (Vogel and Motulsky 1979):

Fig. 2. Autoradiograph of cosmid HX754 DNA, partially digested with BglII (B), PstI (P), EcoRI (E) and HindIII (H) and hybridized with the labelled pAT153 SalI/BamHI fragment. The restriction sites can be read from the bottom to the top of the gel and show the restriction map of HX754 from the left side onwards

272

~75441

754 12a

RFLP

RFLPs

RFLP

Y

YY

v

754.12b

754

754.11

7546 CX 754

HX 754 CX 754-8

i

5kbj

Fig.3. Map of the cloned DXS84 region for the restriction enzymes EcoRI (E), BglII (B), HindIII (H) and PstI. The isolated cosmid clones HX754, CX754 and CX754.8, the positions of the single copy clones HX754.41, HX754.12a, HX754.12b, 754.11 and 754.6 and polymorphic sites (RFLP) are indicated SauIIIA subclones of cosmids in pUC12. Single copy subclones were identified by hybridizing them to filters containing C12D D N A and human D N A digested with various enzymes. C12D was included to confirm the X-chromosomal nature of the clones. Selected subclones were subsequently hybridized to the cosmid D N A , digested with one or more restriction enzymes. Cosmid CX754 contains 42kb of human insert, cosmid HX754 33 kb (Fig. 3). They overlap about 12 kb and cover in total approximately 70 kb of sequences surrounding 754. Cosmid CX754-8 does not extend beyond this region, but it has been useful in providing the other allele of one of the newly detected RFLPs.

RFLP investigations Five subcloned probes (Fig. 3) have been used for the detection of additional RFLPs. The D N A of seven females and three males was employed in a test panel, using minimally six different enzymes (TaqI, MspI, HindIII, EcoRI, BglII and PstI). In addition to the PstI polymorphism, three polymorphic sites were discovered (Table 1). Clone 754.11 detects a BglII RFLP (minor allele frequency 0.40) (Fig. 4a) and an EcoRI RFLP (minor allele frequency 0.23). Clone 754.6 detects a HindIII RFLP (minor allele frequency 0.40) (Fig. 4b). A Mendelian segregation pattern of the BglII, H i n d I I I and EcoRI RFLPs is observed in Dutch DMD families, as is shown for the EcoRI RFLP (Fig. 5). The polymorphic restriction sites have been mapped in the cosmids to determine their physical distance (Fig. 3). The polyrnorphic PstI site could be readily deduced from the heterozygosity of the two cosmids CX754 and HX754 and maps 3 kb "left" of the actual 754 sequence. The 5.3 kb allele of the BglII RFLP is found in CX754. CX754-8 has the 6.5 kb allele, and shows that the right-hand site of the 5.3 kb fragment is polymorphic. The EcoRI RFLP, also detected by

Fig.4. a This is an autoradiograph of genomic DNA digested with BglII and hybridized with clone 754.11. The lower band is the 5.3 kb major allele, the 6.5kb band is the minor allele, b This is an autoradiograph of DNA of 3 females digested with HindIII and hybridized with 754.6. Two are homozygous for the major 2.1kb allele and one is homozygous for the minor, 5,1 kb allele

754.11, has a minor allele of 2.2kb. This allele implies the presence of a new EcoRI site at the right-hand side of 754.11, 2.2 kb from the left-ward EcoRI site. The EcoRI RFLP thus maps within a few hundred bp of the polymorphic BglII site. Clone 754.6 detects a HindIII constant fragment of 1.1 kb and a variable fragment of 2.1 kb or 5.1 kb. The cosmid contains the 2.1 kb allele. Absence of its right-hand site would indeed result in a 5.1 kb allele. According to the restriction map, the PstI and BglII RFLPs are 16kb appart and the H i n d I I I RFLP maps 11kb further to the right. The BglII and EcoRI RFLP, detected by 754.11 are not caused by the same mutation because of their different allele frequencies.

Linkage disequilibrium Table 1. Polymorphisms at the DXS84 locus Probe name

Enzyme

Alleles

Designation

Frequency

754

Pst I

12 kb 9 kb

P p

0.60 0.40

754.11

BglII

5.3 kb 6.5 kb 4.2 kb 2.2 kb

B b E e

0.60 0.40 0.77 0.23

2.1 kb 5.1 kb

H h

0.60 0.40

EcoRI 754.6

Hind III

To test the possibility of linkage disequilibrium, D N A from 46 unrelated females from Dutch D M D families, homozygous for the PstI RFLP, was digested with HindIII, EcoRI and BglII. Southern blots of this D N A were hybridized with the probes 754.6 and 754. f 1, and the haplotypes were determined (Table 2). In total, all but one were also homozygous for BglII, and all but two for HindIII. In 61 chromosomes the major PstI allele is coupled to the major BglII allele and the major H i n d I I I allele, giving the major haplotype P-B-H. In 29 chromosomes the three minor alleles are coupled, giving the p-b-h haplotype. Only one p-B-H and one p-b-H haplotype have been found. In contrast, the minor EcoRI allele occurs exclusively in the major P-B-H haplotype, giving three pre-

273 dominant haplotypes. The observed haplotype frequencies differ from the expected values (Table 2). The PstI, BglII, EcoRI and HindIH RFLPs were tested pair wise for non random association by calculating the values for Delta and "r" (Table 3). The observed linkage disequilibrium was found to be significant at a 0.t% level, when the 2 x 2 table ~2 test was used. As a consequence, 47% of the homozygotes for the major P-B-H haplotype will be heterozygote for the EcoRI RFLP and the overall heterozygosity of the DXS84 locus (754) increases from about 50% to 65%.

10,i

Discussion

Linkage disequilibrium

Fig.5. a This is a pedigree of family DL 85 segregating the DMD locus. ©, Female; ®, female carrier by genetic proof; II, affected male; n , healthy male;ll, deceased male. b This is an autoradiograph of the EcoRI polymorphism of 754.11. The mother I is heterozygous. Individual 10 is a DMD patient carrying the minor 2.2 kb allele. Individual 3 and 11 are healthy boys with the major 4.2 kb allele. The DMD couples to the 2.2kb EcoRI fragment, thus individuals 4 and 8 have a low risk, and individuals 6 and 7 have a high risk of being carriers

2. Haplotype frequencies of the Pst I, Bgi II, EcoRI and Hind III RFLPs in 92 chromosomes. Major alleles are represented by "+" and minor alleles by " - " Table

PstI 754

BglII EcoRI HindIII No. of 754-11 754-6 observed haplotypes

Observed Expected haplotype haplotype frequency frequency

+ + +

+ + +

+ + +

+ + -

-

+

+

+

0.41 0.31 0.25 0.01 0.01 0.00

38 29 23 1 1

Other haplotypes

0

0.16 0.05 0.05 0.11 0.11 0.52

3. Correlation coefficient (r) of the pairwise associations are shown in the upper right part of the table; Delta values are given in the lower left part. In all cases Delta deviates significantly from zero at a 0.1% level Table

Delta value PstI BglII HindIII EcoRI

r-Value PstI x 0.24 0.25 - 0.29

BglII 0.98 x 0.24 - 0.28

HindIII 0.95 0.98 x - 0.29

EcoRI 0.36 0.35 0.33 x

RFLPs are relatively rare on the X-chromosome, 17 polymorphic sites were found in 6200 basepairs screened per haploid genome (Hofl~er et al. 1986b). The 754 region appears to be an exception; three additional RFLPs were found in 252 basepairs tested. However, haplotype analysis revealed very tight association between all 4 RFLPs found in a region spanning 27 kb. Several other cases of linkage disequilibrium between RFLPs have already been documented, for instance for ~-globin (Chakravarti et al. 1984) and the anonymous locus D l l S l 2 (Barker et al. 1985). In contrast with the DXS84 region, the ~-globin region shows two clusters of polymorphic sites which are in linkage disequilibrium, separated by only 9.1 kb. Moreover, at the D l l S 1 2 locus two strongly correlated polymorphisms show a smaller degree of linkage disequilibrium with two RFLPs that lie in between them. These two studies show, that a small physical distance does not always imply a high degree of non-random association. However, if a linkage disequilibrium is observed between two pairs of RFLPs, they can be expected to map close to each other. Because of the very small distance between the EcoRI and the BglII RFLPs (approximately 200bp, see Fig. 3), recombination, when it does occur, will probably take place in the interval between these two RFLPs on one hand, and the PstI or the HindIII RFLP on the other. Such recombination events are the most likely cause of the haplotypes pBEH in the first case and PBEh in the second, shown in Table 2, and can also generate PbEh, pbEH, and PBeh haplotypes. The fact that these latter haplotypes have not been observed in our study implies that genetic drift has either eliminated them or maintained them at low frequency. The pbeh haplotype would not reasonably be expected by meiotic exchange on the basis of present day haplotype frequencies and it could be caused by a new mutation in the EcoRI site in a pbh haplotype. The EcoRI polymorphism of DXS84 has a lower allele frequency than the other RFLPs. Because the minor allele occurs exclusively in a portion of the major PBH-haplotype, we suggest that the EcoRI polymorphism evolved later in this haplotype. To investigate the origin of the EcoRI RFLP further, it may be interesting to test other populations for its association with the PBH haplotype as well, because in principle it appears to postdate the establishment of linkage disequilibrium of the PBH and pbh haplotypes. Due to the EcoRI polymorphism, 47% of the homozygotes for the major PBH allele will be heterozygous for the EcoRI RFLP, increasing the heterozygosity of the DXS84 (754) locus to 65%. The fraction of cases in which combinations of probes

274 will give a 99% predictive reliability in diagnosis of D M D is raised to approximately 90%. The linkage disequilibrium is in contrast with the observed overall increase in recombination in the Xp21 region. The genetic distance of 754 to D M D ranges in various studies from 5 c M upto 19cM, with a peak Lod score at l l c M , obtained over all known data (Goodfellow et al. 1985); 754 maps into the overlap of the BB-deletion (Francke et al. 1985), estim a t e d to be 2000-5000 kb with other D M D associated deletions (Wieringa et al. 1985; Patil et al. 1985), extending from within the BB deletion, towards the telomere ( H o N e r et al. 1986a). O n this basis, the distance between D M D and 754 is likely to be 1000 to 2000 kb, which results in a recombination frequency of 1 cM per 100-200kb. This is tenfold higher in comparison with the estimated average recombination frequency of 1 cM/1000 kb for the whole X-chromosome. It is therefore very likely that we are at a point in the molecular resolution of this specific chromosomal segment, in which one begins to detect discrepancies from an average recombinational behaviour, dissecting regions of frequent and rare recombination.

Future prospects Because the genetic defect in D M D is unknown, cloning of the D M D gene has to be achieved without any knowledge of its structure or function. A l t h o u g h on a chromosomal level, the distance between 754 and D M D is very small, it may still be too large to clone the gene by employing a chromosomal walk with cosmids. A possible cloning strategy could be based on the recent discovery of small deletions associated with D M D , to which P E R T 87 maps (Monaco et al. 1985). A t present, we are studying these chromosomal disorders with probes of the DXS84 locus, using pulsed field gradient electrophoresis (Carle and Olson 1984; Schwartz and Cantor 1984). This technique allows the examination of the chromosome over a distance of 100-1000 kb in both directions from 754. In this way, both the molecular distance between D M D and 754 and the orientation of the cloned region along the c h r o m o s o m e can be estimated.

Acknowledgements. We would like to thank Dr. J. d'Amaro for the helpful discussions concerning the computing of the data, Drs. E.C. Klasen and M.H. Breuning for their critical reading of the manuscript. We are also indebted to Dr. H. Lehrach, in whose laboratory some of the work was carried out, for helpful discussions and invaluable support. The project was supported by the Foundation for Medical Research FUNGO, grant no. 13-23-47, which is subsidized by the Netherlands Organization for the Advancement of Pure Scientific Research (ZWO), the Netherlands Prevention Fund, grant no. 28-878 and the British Muscular Dystrophy Group.

References Bakker E, Bonten EJ, deLange LF, Veenema H, Majoor-Krakauer D, HoNer MH, van Ommen GJB, Pearson PL (1986) DNA probe analysis for carrier detection and prenatal diagnosis of DMD: a standard diagnostic procedure. J Med Genet (in press) Barker D, Holm T, White R (1985) A locus on chromosome 11p with multiple restriction site polymorphisms. Am J Hum Genet 36: 1159-1171

Bates PF, Swift RA (1983) Double cos site vectors: simplified cosmid cloning. Gene 26 : 137-146 Carle CF, Olson MV (1984) Separation of chromosomal DNA molecules from yeast by orthogonal field alternation gel electroforesis. Nucleic Acids Res 12 : 5647-5664 Chakravarti A, Buetow KH, Antonarakis SE, Waber PG, Boehm CD, Kazazian HH (1984) Nonuniform recombination within the human 13globin gene cluster. Am J Hum Genet 36 : 1239-1258 Davies KE, Speer A, Herrmann F, Spiegler AWJ, McGlade S, HoNer MH, Briand P, Hanke S, Schwartz M, Steinbicker V, Szibor R, Korner H, Sommer D, Pearson PL, Coutelle Ch (1985) Human X-chromosome markers and Duchenne muscular dystrophy. Nucleic Acids Res 13 : 319-326 Francke U, Ochs HD, de Martinville B, Giacolone J, Lindgren V, Disteche C, Pagon RA, Hofker MH, van Ommen GJB, Pearson PL, Wedgewood R (1985) Minor Xp21 chromosome deletion associated with expression of Duchenne muscular dystrophy, chronic granulomatous disease, retinitis pigmentosa and McLeod syndrome. Am J Hum Genet 37 : 250-267 Goodfellow PN, Davies KE, Ropers HH (1985) Report of the committee on the genetic constitution of the X and Y chromosomes. Cytogenet Cell Genet 40 : 296-352 Hanahan D, Meselson M (1980) Plasmid screening at high colony density. Gene 10: 63-67 HoNer MH, Wapenaar MC, Goor N, Bakker E, van Ommen GJB, Pearson PL (1985) Isolation of probes detecting restriction fragment length polymorphisms from X-chromosome specific libraries. Hum Genet 70 : 148-156 HoNer MH, Bergen AAB, Skraastad MI, Bakker E, Francke U, Wieringa B, Bartley J, van Ommen GJB, Pearson PL (1986a) Isolation of a random cosmid clone, cX5, which defines a new polymorphic locus DXS148 near the locus for Duchenne muscular dystrophy. Hum Genet 74: 275-279 HoNer MH, Skraastad MI, Bergen AAB, Wapenaar MC, Bakker E, Millington-Ward A, van Ommen GJB, Pearson PL (1986b) The X-chromosome shows less variation at restriction sites than the autosomes. Am J Hum Genet (in press) Maniatis T, Fritsch EF, Sambrook J (1982) Molecular cloning. A laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY Monaco AP, Bertelson CJ, Middlesworth W, Colletti C-A, Aldridge J, Fischbeck KH, Bartlett R, Pericak-Vance MA, Roses AD, Kunkel LM (1985) Detection of deletions spanning the Duchenne muscular dystrophy locus using a tightly linked DNA segment. Nature 316 : 842-845 Patil SR, Bartley JA, Murray JC, Ionasescu VV, Pearson PL (1985) X-linked glycerol kinase, adrenal hypoplasia and myopathy maps at Xp21. Cytogenet Cell Genet 40 : 720-721 Poustka A, Rackwitz HR, Frischauf AM, Hohn B, Lehrach H (1984) Selective isolation of cosmid clones by homologous recombination on Escherichia coli. Proc Natl Acad Sci USA 81 : 4129-4133 Rackwitz HR, Zehetner G, Murialdo H, Delius H, Chai JH, Poustka A, Frischauf AM, Lehrach H (1985) Analysis of cosmids using linearization by phage lambda terminase. Gene 40 : 259-266 Schwartz DC, Cantor CR (1984) Separation of yeast chromosome sized DNAs by pulsed field gradient electroforesis. Cell 37 : 67-75 van Ommen GJB, Arnberg AC, Baas F, Brocas H, Sterk A, Tegelaers WHH, Vassart G, Vijlder JJM (1983) The human thyroglobulin gene contains two 15-17kb introns near its 3' end. Nucleic Acids Res 11 : 2273-2285 Vogel F, Motulsky AG (1979) Human Genetics. Springer, Berlin Heidelberg New York Wieringa B, Hustinx Th, Scheres J, Renier W, ter Haar B (1985) Complex glycerol kinase syndrome explained as X-chromosomal deletion. Clin Genet 27 : 522-523

Received April 16, 1986 / Revised June 26, 1986