Published ahead of print on July 20, 2006, doi:10.1164/rccm.200603-370OC
© 2006 American Thoracic Society doi: 10.1164/rccm.200603-370OC
Mutations of DNAI1 in Primary Ciliary DyskinesiaEvidence of Founder Effect in a Common MutationUniversity of North Carolina at Chapel Hill, Chapel Hill, North Carolina; Institut de la Santé et de la Recherche Médicale, Créteil, France; Department of Pediatrics and Adolescent Medicine, University Hospital Freiburg, Freiburg, Germany; Royal Free and University College Medical School, London, United Kingdom; Concord Hospital, Sydney; Anatomy & Cell Biology, University of Melbourne, Melbourne, Australia; and Cystic Fibrosis Center, Verona, Italy Correspondence and requests for reprints should be addressed to Maimoona Zariwala, Ph.D., F.A.C.M.G., The University of North Carolina at Chapel Hill, Department of Pathology and Laboratory Medicine, CB# 7248, 7123 Thurston-Bowles Bldg., Chapel Hill, NC 27599-7248. E-mail: zariwala{at}med.unc.edu
Rationale: Primary ciliary dyskinesia (PCD) is a rare, usually autosomal recessive, genetic disorder characterized by ciliary dysfunction, sino-pulmonary disease, and situs inversus. Disease-causing mutations have been reported in DNAI1 and DNAH5 encoding outer dynein arm (ODA) proteins of cilia. Objectives: We analyzed DNAI1 to identify disease-causing mutations in PCD and to determine if the previously reported IVS1+2_3insT (219+3insT) mutation represents a "founder" or "hot spot" mutation. Methods: Patients with PCD from 179 unrelated families were studied. Exclusion mapping showed no linkage to DNAI1 for 13 families; the entire coding region was sequenced in a patient from the remaining 166 families. Reverse transcriptase–polymerase chain reaction (RT-PCR) was performed on nasal epithelial RNA in 14 families. Results: Mutations in DNAI1 including 12 novel mutations were identified in 16 of 179 (9%) families; 14 harbored biallelic mutations. Deep intronic splice mutations were not identified by reverse transcriptase–polymerase chain reaction. The prevalence of mutations in families with defined ODA defect was 13%; no mutations were found in patients without a defined ODA defect. The previously reported IVS1+2_3insT mutation accounted for 57% (17/30) of mutant alleles, and marker analysis indicates a common founder for this mutation. Seven mutations occurred in three exons (13, 16, and 17); taken together with previous reports, these three exons are emerging as mutation clusters harboring 29% (12/42) of mutant alleles. Conclusions: A total of 10% of patients with PCD are estimated to harbor mutations in DNAI1; most occur as a common founder IVS1+2_3insT or in exons 13, 16, and 17. This information is useful for establishing a clinical molecular genetic test for PCD.
Key Words: cilia • dynein • dextrocardia • Kartagener syndrome • mutation Primary ciliary dyskinesia (PCD; online Mendelian inheritance in man [OMIM] 242650) is a heterogeneous genetic disorder associated with impairment of flagellar and ciliary structure and function (1–3). PCD has been reported in all major ethnic groups worldwide (1) and is usually inherited as an autosomal recessive trait, although other modes of inheritance have been reported (4, 5). The incidence of PCD is estimated to be 1 in 16,000 individuals (6, 7). Defective ciliary function in the respiratory tract leads to impaired mucociliary clearance, resulting in chronic oto-sinu-pulmonary disease leading to bronchiectasis. Defective nodal cilial function during embryogenesis leads to situs inversus in 50% of patients with PCD (1, 8–10) and is termed "Kartagener syndrome" (OMIM 244400) (1). The diagnosis of PCD is important especially because of implications for the management of respiratory tract disease (11); however, establishing the diagnosis may be difficult. The present "gold standard" diagnostic test is ciliary biopsy with documentation of abnormal ciliary ultrastructure by electron microscopy (EM) (3, 11–13). This test is not easily accessible and requires specialized expertise.
The respiratory cilium is a highly conserved, complex structure consisting of a 9+2 arrangement of microtubules and approximately 250 proteins (14, 15). The force for ciliary beating is provided by outer dynein arms (ODA) and inner dynein arms (IDA), which are recognizable structures on transmission EM images of ciliary cross section (1, 16). The ODA is comprised of several heavy-, intermediate-, and light-chain dyneins (17, 18). The composition of IDA is highly complex and varies along the length of the axoneme (13, 19, 20). Mutations in any protein or structural subunit of cilia could lead to PCD; however, most patients ( No disease-causing gene has been defined from a genome-wide linkage analysis using multiple families (23). Mutations have been identified in a small number of patients, in DNAI1 (OMIM 604366) and DNAH5 (OMIM 603335), which encode an intermediate and heavy chain of ODA, respectively (24–28). DNAI1 and DNAH5 were identified by candidate gene approach (24) and "homozygosity mapping," respectively (27, 29). In rare instances, mutations in RPGR (involved in X-linked retinitis pigmentosa) have been identified in male patients with retinal degeneration cosegregating with PCD (5, 30–33). The purpose of the current study was to perform a large-scale genetic analysis for DNAI1 to identify mutations in well characterized, geographically diverse patients with PCD. This information will be useful for the development of a clinical molecular genetic test for PCD. In addition, we analyzed whether the recurring 219+3insT (IVS1+2_3insT) mutation (24–26) represents a mutational hot spot or a common founder. Parts of the current study have been previously reported as abstracts (34–36).
Subjects: Clinical Evaluation We studied 218 patients with PCD (179 unrelated families) from the University of North Carolina (UNC) and other specialized PCD centers. Over half of the patients with PCD (n = 112 from 89 families) were thoroughly evaluated at UNC; the evaluation included medical and family history, physical examination, sinus and/or chest radiographs, spirometry, and sputum microbiology (3). Ciliary ultrastructural analysis was carried out on nasal epithelial cells obtained by noninvasive nasal curettage (3, 37) and processed as described previously (3, 26, 37). Nasal nitric oxide (NO) was measured as described previously (3, 26). Blood or buccal cells were obtained from the proband (and relatives when possible) for DNA extraction (3, 26, 38). Genomic DNA was obtained from 106 patients with PCD (90 families) from diverse geographical locations, including Germany (n = 28), France (n = 23), United Kingdom (n = 18), Australia (n = 11), Italy (n = 6), Israel (n = 2), Pakistan (n = 1), and the United States (n = 1). Ciliary ultrastructural analysis and compatible clinical phenotype were used for the diagnosis of PCD. In the absence of ultrastructural analysis, adjunct tests such as immunofluorescent antibody staining of axonemal proteins, ciliary beat frequency, or nasal NO measurement were used. This study was approved by the committee for the protection of the rights of human subjects at the UNC-Chapel Hill and collaborating institutions, and written consent was obtained (see online supplement).
Mutation Profiling
Population Studies
cDNA Analysis
Genotyping of Microsatellite D9S1805 Close to DNAI1
Clinical Phenotype of Patients with PCD We identified 218 subjects with documented PCD from 179 unrelated families. The majority of the families were white (87%). There was a broad mixture of ethnicity in the remaining families (see online supplement). Parental consanguinity was noted in 21 (12%) families (as shown in Table E1 in the online supplement). The clinical phenotypes are representative for that previously reported in PCD (Table E1): the patients ranged from 0 to 73 yr of age, there was an equal distribution of male and female patients, and 52% had dextrocardia (Kartagener syndrome). Most of these patients also had neonatal respiratory distress (72%), recurrent otitis media (91%), sinusitis (97%), and bronchiectasis by radiographic imaging (68%; Table E1). Nasal NO measurements were available for 100 patients with PCD, and the mean ± SD was 21.7 ± 17.4 nl/min. Values of < 100 nl/min are consistent with PCD, whereas mean values for normal control subjects were 376 ± 124 nl/min (mean ± SD) (3).
Exclusion Mapping
Mutation Profiling
Frequency of DNAI1 Mutations The current data (n = 179 families) combined with previously published reports (n = 47 families) (24–26) were used to calculate mutation prevalence. From the total of 226 unrelated patients with PCD, 22 (10%) harbored mutations, of which 20 had two mutations (biallelic), giving 42 known mutant alleles. Of these 42 mutant alleles, 23 (55%) were IVS1+2_3insT mutation; 82% of all unrelated patients with mutations carried at least one IVS1+2_3insT mutant allele. The high prevalence of this IVS1+2_3insT mutation suggested a mutational hot spot or a founder effect (see specifics below). A cluster of mutations was noted in exons 13, 16, and 17, where 12 of 42 mutant alleles resided (Figure 1); 46% of the families carried at least one mutant allele in these three exons. Indeed, 21 (96%) of the 22 unrelated patients carrying mutations would have been detected by analysis restricted to exons 13, 16, and 17 and IVS1+2_3insT mutations. From these 21 patients, biallelic mutations would have been picked up in 15 (68%) unrelated patients.
Correlation of Genotype with Ultrastructure
Population Studies To test whether defined mutations are disease-causing for the PCD phenotype or represent benign polymorphisms, we determined the frequency of all missense and some of the splice and nonsense mutations in the general population by analyzing over 100 control alleles; none of the control chromosomes carried any of the tested mutations (Table 3). Although 100 chromosomes were analyzed earlier to define the frequency of IVS1+2_3insT mutation (24), we studied an additional 382 alleles and found that none carried this mutation.
Polymorphisms and Variants of Unknown Significance During the current study, we found several polymorphisms that had been noted previously, and their reference sequence numbers (www.ncbi.nlm.nih.gov/SNP/) are presented in Table E4. We detected two variants of unknown significance (1177G>A [V393M] and 526_528delGAA [176delE]) on one allele of two unrelated patients with PCD. Full gene sequencing did not reveal another mutation on the trans allele (see online supplement for details).
cDNA Analysis
Founder Effect Given the relatively high frequency and wide geographic distribution of the IVS1+2_3insT mutation among patients with PCD, we wondered whether the IVS1+2_3insT allele has a common origin or reflects the existence of a mutational hotspot in the DNAI1 gene. To address this issue, we analyzed a microsatellite marker (D9S1805) located 0.26 Mb upstream of DNAI1 in eight unrelated patients with PCD, two of whom were homozygous for the mutation and six of whom were heterozygous (i.e., compound heterozygotes). These unrelated patients originated from different geographic regions: four from the United States, two from France, and two from Germany. This analysis showed that these patients shared the same D9S1805 allele, characterized by 19-CA repeats. This allele was heterozygous in all patients except for two, who were homozygous for the IVS1+2_3insT mutation and for the 19-CA allele. In addition, a similar analysis performed in a parent carrying the IVS1+2_3insT mutation on one allele showed a direct association of this mutation with the 19-CA allele at the D9S1805 locus. Moreover, the genotype of the D9S1805 marker on the trans allele was found to be different among six compound heterozygous patients (data not shown). To estimate the frequency of the 19-CA D9S1805 allele in the general population, the D9S1805 marker was genotyped in 60 independent control subjects. Nine different alleles were identified among these 120 chromosomes (Figure E1); the 19-CA allele represented 17.5% of all tested chromosomes.
PCD is a genetically heterogeneous disease associated with abnormal ciliary ultrastructure and function. Before this study was done, disease-causing mutations had been described in two genes (DNAI1 and DNAH5) encoding ciliary outer dynein arms in a small number of patients with PCD (24–27). However, genotype–phenotype correlations and the prevalence of mutations in these genes have not been investigated. We undertook a mutation analysis of DNAI1 in a large number of well characterized patients with PCD to test for the correlation of mutations with ciliary ultrastructural defects (ODA versus other defects) and clinical phenotype and to define the prevalence of DNAI1 as a disease-causing gene in PCD. Determining the prevalence and distribution of mutations in DNAI1 is essential if a clinical molecular genetic screening test for PCD is to be developed. Because PCD is a rare genetic disorder, we obtained DNA samples from six geographically different cities to increase the number of patients tested. All specimens came from the specialized PCD centers that are experienced with PCD diagnosis using established methods. The majority of the patients (141 of 179) had specific ultrastructural defects in cilia (Table E1). In addition, the patients without the defined ultrastructural abnormalities were included if they had a strong clinical phenotype, including oto-sinu-pulmonary disease, situs abnormalities, bronchiectasis, neonatal respiratory distress in term neonates, history of infertility, low nasal NO, or abnormal ciliary beat frequency. A case report on a patient with PCD and normal dynein arms has been reported previously (40). Our current estimates of 218 subjects from 179 unrelated families indicate that approximately 9% of all patients with PCD have disease-causing mutations in DNAI1 and that approximately 13% of patients with PCD with ODA defects have DNAI1 mutations (Table 2). None of the 32 unrelated patients in the current study without defined ODA defects had any DNAI1 mutation. Thus, DNAI1 mutations are detected exclusively in association with the ODA defects. Similar observations have been made in Chlamydomonas, where IC78 (ortholog of DNAI1) mutations lead to ODA defects in flagella (17). We also observed a similar prevalence of mutations in DNAI1 when we combined our current study with three previous studies (24–26) involving 47 families (10% of all the 226 families had DNAI1 mutations, and of those 14% with a known ODA defect harbored DNAI1 mutations; Table 2). We believe that this observation is a realistic frequency and that there is no ascertainment bias because all 89 unrelated patients with PCD evaluated at UNC were sequenced without preselection, and mutations were seen in 9% (n = 8 patients). In addition, the current study shows that 14% of patients with PCD with defined ODA defects have mutations in the DNAI1, and because approximately 66% of patients with PCD present with ODA defects (1, 3, 21), approximately 9% of patients with PCD are likely to have mutations in DNAI1 gene. The use of polymorphisms(s) to "exclude" families is useful in a genetically heterogeneous disorder like PCD, especially when dealing with the large genes, because it reduces the need for full gene sequencing in a large number of subjects. During this study, we found nine polymorphisms that were useful for exclusion mapping because they permitted us to exclude 13 families from linkage to DNAI1. The pattern of inheritance of DNAI1 mutations in PCD is autosomal recessive (Figure 2). Of 22 patients with DNAI1 mutations from the current and previous studies (24–26), 20 had biallelic mutations (five patients were homozygous, and nine patients were compound heterozygous). Furthermore, the segregation pattern, if available (Table 1 and References 24–26), showed that these mutations were transallelic. In two patients from the current study, only monoallelic mutations were observed. In these two patients, it is likely that the mutation on the trans allele was missed because only genomic DNA sequence analysis was performed. There is no clinical evidence of dominant inheritance in these two patients because a parent from each of these patients carried the same mutation but was clinically unaffected. In addition, one patient carried the common mutation in the heterozygous state, and we know that this mutation segregates in a recessive manner. Although we have not observed evidence for a digenic pattern in PCD, it cannot be ruled out. There are other genetically heterogeneous disorders where this mode of inheritance has been described (41–43). The type of mutations in DNAI1 varied. Of the 42 mutant DNAI1 alleles from 22 unrelated patients from current and previous studies (24–26), 18 different mutations have been described, including frameshift/stop mutations, base substitutions, and splice mutation. Twelve of the mutations described in the current study are novel mutations (Figure 1). There were two missense variants (V408M and A538T), which we believe are likely to be disease-causing mutations because (1) one of the variants (A538T) was found to be present in two unrelated families, (2) segregation analysis showed that they were transallelic with respect to the second mutation, (3) both variants involved residues that are highly conserved across the species and belong to the conserved tryptophan and aspartic acid (WD) repeat domain of the gene, and (4) neither variant was detected in more than 100 control chromosomes from the general population (Table 3). Similarly, none of the splice mutations was found in any of the control chromosomes (Table 3). Although mutations were present throughout the DNAI1, most mutations occurred in intron 1 (IVS1+2_3insT was present in 55%) and in exons 13, 16, and 17 (29% of the mutant alleles). Moreover, nearly half of the families (46%) carried at least one mutant allele in these three exons. Having a high prevalence of mutations in discrete regions (intron 1 and exons 13, 16, and 17) of DNAI1 will aid in the development of the clinical screening panel for PCD. The status (mutation vs. polymorphism) of two heterozygous variants (526_528delGAA [176delE] and 1177G>A [V393M]) discovered during this study is uncertain. Both variants are evolutionarily conserved among some species, and they were not present in control chromosomes from the general population (108 chromosomes for 176delE and 394 chromosomes for V393M), suggesting that they may be true mutations. However, it seems more likely that these are polymorphisms for several reasons: (1) a patient who carried these variants did not have an ODA defect (based on EM data), and mutations in DNAI1 have been associated with ODA defects; (2) a patient who carried 176delE in a heterozygous state was born to consanguineous parents, and such a patient would likely depict "identity by descent" and hence homozygosity for the mutations or polymorphisms at the disease locus; and (3) in rats, methionine is used at the equivalent codon of V393M, suggesting that methionine is the normal variant and hence V393M represents a rare polymorphism (see online supplement). We also observed the consequence of three splice mutations on mRNA levels. We confirmed that the IVS1+2_3insT mutation causes aberrant splicing, as previously reported (24). Another mutation, 1490G>A, predicts an amino acid substitution; however, we demonstrated that it also affects splicing, leading to in-frame deletion of exons 15 and 16 (56 amino acids). Finally, a splice mutation in intron 19 resulted in an in-frame deletion of exon 19 (61 amino acids). Although approximately 15% of the known disease-causing mutations are splice mutations (44), we found no (deep intronic) mutations by the analysis of RT-PCR products (Table 4). The mutations in DNAI1 seem to be most common in white individuals with PCD. Combining data from the current study with previous studies (24–26), ethnicity information is available on 182 unrelated patients with PCD (364 alleles). Most of the 364 alleles were in patients with Caucasian origin (315 alleles), and 34 alleles carried DNAI1 mutations (11%). Of the alleles from the patients of non–Caucasian ethnicity (n = 47 alleles), only one allele carried a DNAI1 mutation (2%). The non–Caucasian dataset is small but suggests that DNAI1 mutations may be found more commonly in patients with a Caucasian ethnicity. One reason for DNAI1 mutations in white patients could be the fact that the IVS1+2_3insT mutation was present in 55% (23 of the total 42) of the mutant alleles; moreover, 82% (18 of 22) of unrelated patients had IVS1+2_3insT on at least one allele. Based on the high frequency of the IVS1+2_3insT mutation, founder effect analysis was carried out, and several lines of evidence strongly suggest the presence of a founder effect. First, a careful examination of the nucleotide environment did not reveal structural features that would favor the insertion of one base pair at this position. Second, all the patients who carried this mutation and who were genotyped at the closely linked D9S1805 marker were shown to carry the same polymorphic allele bearing 19-CA repeats. The association of the disease-causing mutation with the 19-CA D9S1805 allele was clearly demonstrated in two homozygous and one compound heterozygous patient. Although the parental DNA samples were not available for the other compound heterozygous patients, all these patients were heterozygous for the 19-CA D9S1805 allele. These data strongly indicate that the IVS1+2_3insT alleles originate from an ancestral mutational event. Taken together, the prevalence of the IVS1+2_3insT mutation and the cluster of mutations in exons 13, 16, and 17 would allow identification of most patients with DNAI1 mutations. Specifically, 21 of all 22 unrelated patients described in the present and past studies (24–26) that carry mutations of DNAI1 would be detected by screening for IVS1+2_3insT and mutations in exons 13, 16, and 17. This suggests that the development of clinical molecular genetic testing is a realistic and useful goal. Similar data are available for another gene encoding an ODA protein, DNAH5 (28), which has a frequency of mutation in all PCD of 28% (38 of 134 unrelated families with PCD). Moreover, in 47% of all mutant DNAH5 alleles (31 of 66), the mutations were located in clusters of five exons. Combining the data for both genes (DNAI1 or DNAH5), it seems that approximately 38% of all the patients with PCD are estimated to carry mutations in DNAI1 (22 families) or DNAH5 (38 families). For feasibility of genetic screening, it is important to note that over half (66 of 108) of the mutant alleles for both genes fall into the mutation cluster of four exons in DNAI1 and five exons in DNAH5. Thus, screening for only these nine exons would permit detection of 41 of 60 (68%) families that harbor mutations in at least one allele. A reasonable approach for developing a clinical test panel would be to test by "tiered" approach, for example, screening for mutation cluster region of the DNAI1 and DNAH5 in phase I, followed by full gene sequencing of the appropriate gene for the patients with monoallelic mutations in a second phase. Such genetic analysis would help expedite and improve the diagnosis of PCD and subsequent counseling for patients and families. In conclusion, we show that mutations in DNAI1 are estimated to occur in up to 14% of patients with PCD and reflect an autosomal recessive mode of inheritance. Most mutations in DNAI1 are truncating mutations and seem to occur more commonly in patients with Caucasian ethnicity. These data, coupled with the data for DNAH5 mutations (28), indicate that development of a clinical molecular genetic panel is feasible and would improve diagnosis and counseling for PCD.
The authors thank the patients with PCD and the family members from North America, Europe, and Australia for their participation. The authors thank Ms. S. Minnix, Mr. S. Lyons, Ms. R. Pace, Ms. E. Godwin, Dr. Wanda O'Neal, and the Genome Sequencing Center (Washington University, St. Louis) for their contributions. The authors also thank Ms. Michele Manion and the PCD Foundation for their continued support. The authors acknowledge Dr. Eitan Kerem from Hadassah University Hospital, Israel; Dr. H. Blau from Schneider Medical Center of Israel, Israel; Ms. Lindsay Stout, Texas General Hospital; Dr. Andrew Bush from the Royal Brompton Hospital, London, UK; Dr. Christopher O'Callaghan from the Leicester Royal Infirmary, UK; and Ms. Carie Drovdlic from Ohio State University for providing clinical information and DNA from a PCD patient/family.
Supported by grants GCRC 00046, MO1 RR00046-42, 1 RO1 HL071798, Deutsch Forschungsgemeinschaft (SFB592 and DFG Om6/2), GIS Maladies Rares A03091, the ANR Maladies Rares, the Assistance Publique-Hôpitaux de Paris (CRC 96125), Concord Hospital, the Medical Research Council (UK), the Milena Carvajal Pro-Kartagener Foundation, the Wellcome Trust (UK), and grant 5 U54 RR019480 from the National Center for Research Resources (NCRR), a component of the National Institutes of Health (NIH). Its contents are solely the responsibility of the authors and do not necessarily represent the official view of NCRR or NIH. This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Originally Published in Press as DOI: 10.1164/rccm.200603-370OC on August 3, 2006 Conflict of Interest Statement: M.A.Z. is a coinvestigator on a research grant from the North Carolina Technology Center to develop assays relating to DNAI1. The grant amount is $250,000 for the total of 2 yr (June 1, 2006 through May 31, 2008). No work in the current study was supported by this grant. M.W.L. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. F.C. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. M.P.K. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. P.G.N. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. J.L.C. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. M.J.H. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. A.L. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. J.H. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. H.O. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. N.T.L. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. A.-M.B. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. G.P. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. B.D. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. E.E. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. H.M.M. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. R.C. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. E.M.K.C. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. L.M.M. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. R.U.d.I. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. J.R. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. U.P. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. H.O. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. S.A. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. M.R.K. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form March 13, 2006; accepted in final form July 17, 2006
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
60–70%) have defective ODAs (
