Somatic mutations in transcription factor genes pertinent to cardiac tissue development have been put forward as a molecular rationale of CHD. Nkx2-5 is a homeodomain-containing transcription factor and is conserved in many organisms from flies to humans. It is an important transcriptional regulator of mammalian heart development. Absence of Nkx2-5 in animal models results in lethality due to impaired heart tube looping (for instance, see review [1]). In this regard, Draus et al. [2] reported that somatic mutations in NKX2-5 do not represent an important aetiologic pathway in pathologic cardiac development in patients with cardiac septal defects. This conclusion was based on the genetic analysis of frozen cardiac tissue samples of 28 patients with septal defects (ASD, n=13), ventricular septal defects (VSD, n=5), and atrioventricular canal defects (AVCD, n=10). Cardiac tissue samples were collected from diseased tissue located immediately adjacent to the defect and from anatomically normal tissue located at a site remote from the defect (right atrial appendage). Except for one nonsynonymous germline sequence variant in 1 patient, two synonymous germline sequence variants in 2 separate patients, and a common single nucleotide polymorphism (SNP) in 16 patients, no NKX2-5 somatic mutations were obtained.

The motivation behind Draus et al. study [2] was to replicate in frozen cardiac tissues our previous findings of somatic mutations in the NKX2-5 gene that we analyzed in the Leipzig collection of malformed hearts [3;4]. We investigated 68 malformed hearts from unrelated individuals collected from 1954 to 1982 and stored in formalin at the Institute of Anatomy, University of Leipzig, Germany. After detailed morpho- pathological characterization, the malformed hearts were classified by sequential segmental analysis [5] according to their septal defects, e.g. 29 VSDs, 16 ASDs and 23 AVSDs. But besides their septal defects, the patients had complex cardiac malformations, and many were affected by patent foramen ovale (PFO), patent ductus arteriosus (PDA) or both. Indeed, at least 42 different cardiac malformations were represented in this patients' cohort. Most of the patients died during early infancy from the disease and four patients within VSDs and 14 within AVSDs had Down syndrome. As control, we also investigated 10 formalin-fixed, normal hearts from the same collection, as well as five frozen normal hearts. To summarize briefly our previous findings on NKX2-5 [3;4], direct sequencing in 'affected' tissues (i.e. near septal defect) identified 35 different nonsynonymous mutations which were mainly absent in 'unaffected' tissues (i.e. away septal defect) in the same patient's heart. Certain mutations were frequent yet specific to a particular septal defect. Several of the frequent nonsynonymous mutations impaired protein function by reporter assays [6] or by bioinformatic prediction. Most important, detected mutations were absent in control hearts from the same collection. The NKX2-5 dbSNPs rs2277923 (c.63A>G, p.E21) and rs703752 (c.*61T>G) were also found indicating integrity of the archival DNA even after many years of formalin storage. Surprisingly, many patients had several NKX2-5 mutations, and after cloning of PCR fragments with several closely-spaced heterozygous mutations, more than the two expected haplotypes were found. We proposed that somatic NKX2-5 mutations as a possible mechanism of disease in complex congenital heart disease. We tried to explain the simultaneous occurrence of multiple NKX2-5 mutations as a reflection of the complexity of the observed cardiac malformations, with patients carrying multiple defects in the heart. We also suggested that a likely explanation for the multiple haplotypes would be a mixed population of cardiomyocytes carrying different mutations or de novo chromosomal rearrangements and gene duplications in the tissues of patients affected by various mutations.

We wish to comment on the Draus et al. study on the following points:

Draus et al. stated: "Six hundred and five somatic sequence variants were reported in the coding and flanking intron regions of NKX2-5. When we analysed the spectrum of these sequence variants found after amplification from DNA extracted from formalin fixed tissue, we were surprised to discover that A>G substitutions dominated these studies (58% of these sequence variants were A>G substitutions) and only 31% were G>A substitutions." Comment: We find such representation of our results inaccurate. For NKX2- 5, we reported 35 different nonsynonymous mutations, several of which are common, specific to septal defects, and affect protein function. Among these 35 nonsynonymous mutations, 8 are A>G substitutions and include the loss-of-function mutation p.K183E which is located in the third helix of homeodomain, and detected in 22 of 23 of AVSDs, but not in VSDs. Furthermore, 4 different G>A substitutions were found. We also found 14 synonymous mutations in the coding region of NKX2-5, 5 of these are A>G including dbSNPs rs2277923 and 3 of these A>G were frequently detected. Consequently, there needs to be an erratum in the publication of Draus et al. that it incorrectly quoted our findings.

Draus et al. stated: "Similarly, the same group found somatic sequence variants in the TBX5 gene from the same archival collection of hearts. TBX5 encodes a transcription factor that interacts with NKX2-5 and is important in vertebrate cardiac development. One hundred and thirty- five somatic sequence variations have been reported by the same authors in the coding and flanking intron regions of TBX5. Of the single nucleotide substitutions reported, 87 percent were adenine-to-guanine substitutions. Only 11 percent were guanine-to-adenine substitutions." Comment: Again, the representation of our results is incorrect. For TBX5, we obtained only 9 single base substitutions in the coding region of TBX5 resulting in nonsynonymous change that were absent in matched 'unaffected' tissues [7]. Six mutations would affect amino acids in the T-domain. No frequent mutation was obtained except c.236C>T (p.A79V) which was detected in five cases. Only 11 of 68 patients had nonsynonymous TBX5 mutations and none were found in VSDs.

Draus et al. stated: "In our study, an average of 1.3 microgram of genomic DNA (gDNA) was isolated from 5 mg of tissue obtained as surgical discards that were immediately frozen. In comparison, yields from the Leipzig collection were reportedly 0.5 to 1 microgram of genomic DNA from 25 mg DNA obtained from archival tissue. Therefore, the recovery of gDNA from our fresh frozen samples was more than five times that of the formalin-fixed samples. The relatively poor gDNA yield from the Leipzig collection likely reflects the poor quality of DNA taken from samples that had been fixed in formalin between 22 to 50 years." Comment: We wish to point out that the average size of genomic DNA isolated from the Leipzig collection was about 2 kb, and 20-50 ng was used to amplify three NKX2-5 fragments of sizes 489, 472, and 573 bp for sequence analysis. Although the quantity of DNA from fixed material was not comparable to that of fresh tissue, the quality of the genomic DNA as shown by the amplified NKX2-5 fragments was amenable for genetic analysis.

Draus et al. stated: "The Reamon-Buettner studies are significantly different from germline NKX2-5 mutations as well as significantly different from inherited disease associated mutations, tumour derived somatic mutations, and mitochondrial mutations. While these differences may reflect true differences between the patients at the mutational pathway level (germline versus somatic versus mitochondrial mutations), geographic and/or patient level (Germany and/or CHD patients versus an international collections of cancer databases), or the temporal level (over 22 years ago versus the current era), it is also possible that the previously reported somatic variants reflect postmortem artefacts of fixation or low quality DNA template. Recent studies also show that lower starting yield of genomic DNA from archival samples can result in higher misincorporation rates during PCR in a sequence dependent manner" Comment: Assuming the NKX2-5 mutations identified by us are postmortem artefacts of fixation, then why are the unaffected tissues or same formalin-fixed normal hearts did not contain the mutations? Why do patients carry common pathogenic mutations?

Draus et al. stated: " Recent studies also show that lower starting yield of genomic DNA from archival samples can result in higher misincorporation rates during PCR in a sequence dependent manner 8" Comment: As already mentioned earlier, the average size of genomic DNA isolated from the Leipzig collection was about 2 kb, and 20-50 ng was used to amplify three NKX2-5 fragments of sizes 489, 472, and 573 bp for sequence analysis. The paper by Akbari et al. [8] used paraffin-embedded tissues and the source for genomic DNA isolation was few microdissected cells.

Taken collectively, the study of Draus et al. provided erroneous information that requires an erratum and clarification. While the pathogenesis of CHD is complex, in which majority of CHD being sporadic, it is an acknowledged fact that patients have no family history of the disease and therefore does not follow a typical Mendelian trait. There is clinical or phenotype heterogeneity within affected individuals, as well as within and between families. Disease association is difficult owing to lack of clear genotype-phenotype correlation of mutations. Furthermore, at the molecular and cellular levels, heart development can practically go wrong in many directions, i.e. transcriptional regulators, signaling pathways and chromatin remodeling, leading to diverse cardiac malformations observed in CHD. Thus, somatic mutations in NKX2-5 may represent just one among the many causes of CHD, and their non-detection in another patients' cohort is, therefore, not totally surprising. Indeed, since NKX2-5 gene mutations have been implicated in human CHD [9], the coding region of NKX2-5 has been consistently analyzed for additional disease-associated sequence alterations. As of today, we documented 41 different NKX2-5 germline mutations most of which lead to amino acid change (nonsynonymous mutations). Most of these mutations would affect protein function, yet there is lack of genotype-phenotype correlation of NKX2-5 mutations in CHD (see review [10]). Notably, the detection frequency of NKX2-5 mutations in sporadic cases of CHD is about 2 % and in several studies none was found. Many identified NKX2-5 mutations are familial cases, and families have their private mutations. Thus, a single germline NKX2-5 mutation as the direct cause of disease or a simple genetic analysis does not reveal causation.

With this letter, we wish to highlight the erroneous quotation and interpretation of our studies.

References

1. Bartlett H, Veenstra GJ, Weeks DL. Examining the cardiac NK-2 genes in early heart development. Pediatr Cardiol 2010; 31:335-41.

2. Draus JM, Jr, Hauck MA, Goetsch M, Austin EH, III, Tomita-Mitchell A, Mitchell ME. Investigation of somatic NKX2-5 mutations in congenital heart disease. J Med Genet 2009; 46:115-22.

3. Reamon-Buettner SM, Hecker H, Spanel-Borowski K, Craatz S, Kuenzel E, Borlak J. Novel NKX2-5 mutations in diseased heart tissues of patients with cardiac malformations. Am J Pathol 2004; 164:2117-25.

4. Reamon-Buettner SM, Borlak J. Somatic NKX2-5 mutations as a novel mechanism of disease in complex congenital heart disease. J Med Genet 2004; 41:684-90.

5. Craatz S, Kunzel E, Spanel-Borowski K. Classification of a collection of malformed human hearts: practical experience in the use of sequential segmental analysis. Pediatr Cardiol 2002; 23:483-90.

6. Inga A, Reamon-Buettner SM, Borlak J, Resnick MA. Functional dissection of sequence-specific NKX2-5 DNA binding domain mutations associated with human heart septation defects using a yeast-based system. Hum Mol Genet 2005; 14:1965-75.

7. Reamon-Buettner SM, Borlak J. TBX5 mutations in Non-Holt-Oram Syndrome (HOS) malformed hearts. Hum Mutat 2004; 24:104.

8. Akbari M, Hansen MD, Halgunset J, Skorpen F, Krokan HE. Low copy number DNA template can render polymerase chain reaction error prone in a sequence-dependent manner. J Mol Diagn 2005; 7:36-9.

9. Schott JJ, Benson DW, Basson CT, Pease W, Silberbach GM, Moak JP, Maron BJ, Seidman CE, Seidman JG. Congenital heart disease caused by mutations in the transcription factor NKX2-5. Science 1998; 281:108-11.

10. Benson DW. Genetic Origins of Pediatric Heart Disease. Pediatr Cardiol 2010; 31:422-9.

Conflict of Interest:

None declared

New challenges for informed consent through whole-genome array testing

Christian Netzer1,2, Christine Klein3, Jürgen Kohlhase4, Christian Kubisch1,2

1Institute of Human Genetics, University of Cologne, Germany

2Center of Molecular Medicine Cologne, University of Cologne, Germany

3Department of Neurology, University of Lübeck, Germany

4Center for Human Genetics Freiburg, Freiburg, Germany

We read with great interest the article by Schwarzbraun et al. on predictive diagnosis of Li-Fraumeni syndrome established as an accidental finding in whole-genome array testing originally performed to identify the molecular cause of mental retardation (MR) in a seven-year-old child1. We would like to expand the spectrum of unexpected and unintended findings with this new technique: In a five-year-old Kurdish boy with mental retardation and normal karyotype, array CGH with an Agilent platform (using the Human Genome CGH 244A Microarray Kit with a medium spatial resolution of 12 kb and a threshold of 5 consecutive aberrant clones) identified a heterozygous intragenic deletion spanning exons 4 and 5 of the Parkin gene (PARK2) as the only significant change (minimal deletion boundaries for NCBI Build 36, chromosome 6: 162,339,755-162,582,128). Bi- allelic mutations of Parkin cause early-onset parkinsonism (OMIM 600116), a neurodegenerative disorder without phenotypic overlap to the MR syndrome of the five-year-old boy. Hence, array CGH accidentally revealed a carrier status for a more late-onset autosomal recessive disease.

We informed the patient´s parents about this result. They planned to have another child and asked us about the risk for their future offspring to be affected by this type of parkinsonism. The maximum frequency of putative recessive heterozygous Parkin mutations reported in the literature is 3.1% 2-5. Hence this risk would be up to 0.75%, if the Parkin deletion was inherited from either parent (offspring risk for carrying two mutations then can be calculated as 1 x 3/100 x 1/4, assuming no consanguinity between the parents). We complied with the parents´ request to clarify this situation through additional genetic tests and initiated MLPA analyses at the PARK2 locus (MRC Holland, SALSA Kit P051) on DNA samples of the parents and the MR-affected son, as well as direct sequencing of all coding exons and adjacent splice sites of Parkin on parental DNA. The Parkin exon 4-5 deletion turned out to be inherited from the patient´s mother, and neither the patient nor his parents carried another mutated Parkin allele.

Of note, testing the parents for mutations in Parkin by MLPA and direct sequencing had a low, but not negligible probability of being a predictive genetic test for early-onset parkinsonism: The disease is not fully penetrant by age 35 and 43, respectively, and since a de-novo Parkin exon 4-5 deletion in the index patient was in our view a priori rather unlikely, there was an up to 3% chance that the parent who inherited this deletion carried a mutation on the other allele, as well. In addition, testing the Parkin gene in the mentally retarded child by MLPA, which we considered to be necessary in order to validate the array CGH finding with an independent method, was associated with a (considerably lower) probability of identifying an exon-spanning deletion also on the other Parkin allele. This further complexity for genetic counselling, caused by the fact that the Parkin Type of Juvenile Parkinson Disease is not an early-childhood-onset autosomal recessive disorder, may be rather the exception than the rule. However, in our view it would have been a sufficient but not mandatory argument to decline the parents´ request for carrier testing in this case. Adding a final level of complexity, there is a growing body of evidence that heterozygous Parkin mutations may act as a risk factor for complex inherited (late-onset) Parkinson disease6, a potentially relevant issue for genetic counselling. Interestingly, the family history of the patient´s mother was positive for Parkinson disease, with her paternal grandfather being affected in the seventh decade of life.

This experience prompted us to rethink our procedure to obtain informed consent to array CGH. We reasoned that – as now reported by Schwarzbraun et al. – also mutations of genes underlying autosomal- dominantly inherited tumor predisposition syndromes or late-onset neurological disorders (e. g., Hereditary Neuropathy with Liability to Pressure Palsies [OMIM 162500], which is typically caused by a heterozygous 1.5-Mb deletion on chromosome 17p11.2) might be accidentally identified through array CGH analyses for mental retardation syndromes. These deletions or duplications do not necessarily have to be an integral part of the imbalance causing the phenotype in question, as it was the case with the seven-year-old mentally retarded boy, who carried a de-novo deletion of 47 genes including the Li-Fraumeni gene TP53. They can of course be inherited independently, and this knowledge could sometimes be of great importance for the entire family (e. g., if an HNPCC gene was affected, as life-saving surveillance programs exist for this cancer predisposition syndrome [OMIM 120435]). On the other hand, some of these conditions may be untreatable or fatal, and whether or not such predictive genetic information should be disclosed is a very personal decision. In addition to these considerations, we are aware of one individual with a non-hereditary hematologic disease (chronic lymphocytic leukemia, with somatic deletions involving chromosome 4 and 11q) accidentally “rediagnosed� in the course of an array CGH research project comparing the genomes of monozygotic twins on the basis of DNA extracted from peripheral leukocytes7.

As a consequence, we have now revised our procedure to obtain informed consent to array CGH prior to the analyses. Patients or their parents have to decide whether

- they wish to be informed about any additional genetic findings (without direct connection to the phenotype in question) with predictive value for the health of the proband and potentially her / his family;

- they only whish to be informed about such additional genetic findings if effective treatment options or surveillance programs are available;

- they whish to be informed about a carrier status for an autosomal recessive disease, i.e., about a condition which may have implications for reproductive decisions of the proband and/or family members.

Once the resolution of array CGH is high enough to accurately detect single-exon aberrations, the latter point may in fact be the one most frequently occurring in clinical practice, simply because there are hundreds if not thousands of autosomal recessive disease genes in the human genome. The resulting genetic risks can be substantial, especially if consanguinity is an issue.

To our knowledge, the implications and challenges for the informed consent procedure resulting from a more detailed look at the humane genome have not yet been broadly discussed with respect to routine clinical testing. However, they have to be addressed in order to cope with such “accidents� of array CGH analysis and to comply with patients´ autonomy.

Correspondence to: Dr. Christian Netzer, Institut für Humangenetik, Uniklinik Köln, Kerpener Str. 34, 50931 Köln, Germany; christian.netzer@uk-koeln.de

Competing interests: None.

REFERENCES

1. Schwarzbraun T, Obenauf A, Langmann A, et al. Predictive diagnosis of the cancer prone Li-Fraumeni syndrome by accident: new challenges through whole-genome array testing. J Med Genet. Published Online First: 5 March 2009. doi:10.1136/jmg.2008.064972

2. Clark LN, Afridi S, Karlins E, et al. Case-control study of the parkin gene in early-onset Parkinson disease. Arch Neurol 2006;63(4):548- 52.

3. Kann M, Jacobs H, Mohrmann K, et al. Role of parkin mutations in 111 community-based patients with early-onset parkinsonism. Ann Neurol 2002;51(5):621-5.

4. Kay DM, Moran D, Moses L, et al. Heterozygous parkin point mutations are as common in control subjects as in Parkinson's patients. Ann Neurol 2007;61(1):47-54.

5. Lincoln SJ, Maraganore DM, Lesnick TG, et al. Parkin variants in North American Parkinson's disease: cases and controls. Mov Disord 2003;18(11):1306-11.

6. Klein C, Lohmann-Hedrich K, Rogaeva E, et al. Deciphering the role of heterozygous mutations in genes associated with parkinsonism. Lancet Neurol 2007;6(7):652-62.

7. Bruder CE, Piotrowski A, Gijsbers AA, et al. Phenotypically concordant and discordant monozygotic twins display different DNA copy- number-variation profiles. Am J Hum Genet 2008;82(3):763-71.

With great interest we read the article of Malan et al., who reported on a novel clinically recognizable 19q13.11 microdeletion syndrome.1 Here we report on a fifth patient with an interstitial deletion overlapping the 19q13.11 region and compare our findings with those described by Malan et al. The proband was born after 37 weeks of gestation as one of dizygotic twins with a birth weight of 1620 g (-3.5 SD). His twin sister had a normal birth weight of 2290 g (p25). They were the first children of healthy non-consanguineous parents. Pregnancy was complicated by intrauterine growth retardation. At birth the proband had cutis aplasia over the posterior occiput, hypospadias, abnormal positioning of the feet, and a third nipple on the left side of the chest (the latter was also present in the father) whereas no dysmophisms where reported in his sister. During the first years of life, feeding difficulties, failure to thrive, and psychomotor delay were noted: he sat independently at two years of age, started crawling at the same age and started walking at the age of three years and ten months. Because of fatigue, he was able to walk short distances only, and walked with a stiff gait. He suffered from recurrent airway infections. At the age of four years and 10 months he could not speak, but was able to communicate with pictograms. He had hypermetropia (+3D/+4D) similar to his sister. At that age his height was 101.5 cm (-2.5 SD), weight was 15.4 kg (-0.5 SD), and head circumference was 45.5 cm (-3.5 SD). He had deep-set eyes, a broad nasal tip, a broad mouth with thin lips, broad gums, irregular placed teeth and retrognatia. Other features were long fingers with slender thumbs and long, dysplastic nails, a third nipple on the left side of the thorax, hypospadias, a sacral dimple and hyperlaxity of the joints (figure 1A-G).

Routine cytogenetic analysis with GTG-banding and subsequent Multiplex Ligation Dependent Probe Amplification (MLPA) analysis of the subtelomeric regions did not reveal any abnormalities. Array based comparative genomic hybridisation analysis (~ 32.000 BAC clones with an average resolution of 300 kb) was performed as previously described (de Vries et al. 2005 2) and revealed a 2.4 Mb loss in 19q13.11q13.12 (39.5- 41.9 Mb) encompassing ~ 50 genes (46,XY.arr cgh 19q13.11q13.12(RP11-615P - > RP11-679B20)x1). This deletion was subsequently confirmed by region specific Fluorescence In Situ Hybridisation (FISH) analysis using three BAC clones which map to the aberrant region (46,XY.ish del(19)(q13.11.q13.12)(RP11_615P5-,RP11_233i16-,RP11_679B20-). In addition, high resolution 250K SNP array analysis (Affymetrix, Inc., Santa Clara, CA, USA; average practical resolution of 200 kb) was performed to further characterize the extend of the deletion (figure 1H). Both SNP array and FISH analysis were performed in the parents showing that the deletion occurred de novo, that the maternal allele was deleted in the proband and that the mother was carrier of an intrachromosomal, submicroscopic insertion on chromosome 19 (46,XX.ish ins(19)(q13.?3q13.11q13.12)(q13.11)(RP11_615P5-, RP11_223i16- )(q13.12)(RP11_679B20-) (q13.?3)(RP11_615P5+,RP11_223i16+, RP11_679B20+). Therefore, recombination during meiosis is the underlying mechanism for the deletion to occur in her son (data not shown).

In addition to the report by Malan et al. and Kulharya et al., 3 this is the fifth patient with an interstitial deletion of the 19q13.11 region, overlapping the smallest region of overlap of the previously described patients by only 750 kb. Remarkably though, our patient shares several major features with the previously reported patients including intrauterine and postnatal growth retardation, primary microcephaly, speech disturbance, mental retardation, hypospadias and ectodermal dysplasia presented by scalp aplasia, dysplastic nails and a dry skin (table 1). The feeding problems in our patient did not seem to be as severe as in the patients described by Malan et al. No cardiac defects or deafness were noted, although the latter was only reported in the case with the large 11 Mb deletion.

The deletion in the current patient narrows down the critical region for this novel microdeletion syndrome to ~ 750 Kb (B36.1:CHR19:39 378 624 - 40 131 974) containing only 11 RefSeq genes (NCBI RNA reference sequence collection 4,5) (figure 1I). For three genes, PDCD2L, KIAA0355, SCGBL, the function is unknown. The glucose phosphate isomerase gene (GPI; OMIM #172400) encodes a housekeeping cytosolic enzyme playing a major role in the glycolysis and the gluconeogenesis pathway. As GPI mutations cause autosomal recessive chronic hemolytic anemia it is not likely that haploinsuffiency of this gene contributes to the phenotype in this novel microdeletion syndrome.6,7,8 Wilms tumor 1 interacting protein (WTIP) is shown to interact with WT1, which is a zinc finger transcription factor. ZNF302, ZNF181, ZNF599 and ZNF30 are clustered on chromosome 19q13.11 and belong to the KRAB-ZNF subfamily. KRAB-ZNFs have been described as transcription repressors. ZNF clusters have expanded in the primate evolution lineage and are suspected to contribute to higher cognitive functioning in primates. 9,10 As three zinc finger genes have previously been shown to be involved in X-linked mental retardation in males, ZNF41, ZNF81, ZNF674, 11,12,13 it might be that haploinsuffiency of the zinc finger cluster in 19q13.11 is involved in the pathogenesis in this novel syndrome. LSM14A is a Sm-like protein with sequence homology with the Sm protein family. Sm-like proteins are evolutionary conserved in eukaryotes and are thought to have a fundamental role in control of mRNA translation making LSM14A a interesting candidate gene.14,15 Ubiquitin-like modifier activation enzyme 2 (UBA2) is playing a role in the ubiquitin pathway. Ubiquitin and ubiquitin-like modifiers regulate multiple important cellular processes: cell differentiation, apoptosis, posttranslational modification, transcriptional regulation and DNA repair and are known to be involved in mental retardation syndromes such as Angelman syndrome (OMIM#105830), Opitz syndrome (OMIM #300000), Bardet-Biedl syndrome (OMIM#209900) and limb-girdle muscular dystrophy type 2H (OMIM #254110).16,17

As our patient has a similar phenotype as the patients described by Malan et al. and Kulharya et al. we narrowed down the critical region of this novel microdeletion syndrome from 2,87 Mb to 750 Kb encompassing 11 genes. It is very likely that haploinsufficiency of one or more genes in this region give rise to this microdeletion syndrome for which LSM14A and UBA2 seem the most promising candidates.

J.H.M. Schuurs-Hoeijmakers, S. Vermeer, B.W.M. van Bon, R. Pfundt, C. Marcelis, A.P.M. de Brouwer, N. de Leeuw, B.B.A. de Vries Department of Human Genetics, Nijmegen Centre for Molecular Life Sciences, Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands

Correspondence to: Bert B.A. de Vries, Department of Human Genetics 849, Radboud University Nijmegen Medical Centre, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands. Tel: +31 (0) 243613946; fax: +31 (0) 243668753; e-mail: b.devries@antrg.umcn.nl

Acknowledgements: We are sincerely appreciative to the patient and his parents for their support and cooperation. This work was supported by grants from the Netherlands Organisation for Health Research and Development (ZON-MW) (BBAdV), Hersenstichting Nederland (BBAdV), and grants from the AnEUploidy project (LSHG-CT-2006-037627) supported by the European Commission under FP6 (BBAdV, BvB). The authors wish to acknowledge the absence of any competing interest.

Competing interest: None.

Patient consent: Consent was obtained from the patient's family for publishing patient details and images in figure 1.

References 1. Malan V, Raoul O, Firth HV, Royer G, Turleau C, Bernheim A et al. 19q13.11 deletion syndrome: a novel clinically recognizable genetic condition identified by array-CGH. J Med Genet. Epub 2009 Jan 6. 2. de Vries BB, Pfundt R, Leisink M, Koolen DA, Vissers LE, Janssen IM et al. Diagnostic genome profiling in mental retardation. Am J Hum Genet. 2005;77(4):606-16. 3. Kulharya AS, Michaelis RC, Norris KS, Taylor HA, Garcia-Heras J. Constitutional del(19)(q12q13.1) in a three-year-old girl with severe phenotypic abnormalities affecting multiple organ systems. Am J Med Genet. 1998;77(5):391-4. 4. Kent WJ. BLAT - the BLAST-like alignment tool. Genome Res. 2002;12(4):656-64. 5. Pruitt KD, Tatusova T, Maglott DR. NCBI Reference Sequence (RefSeq): a curated non-redundant sequence database of genomes, transcripts and proteins. Nucleic Acids Res. 2005;33(Database issue):D501-4 6. Arnold H. Inherited glucosephosphate isomerase deficiency. A review of known variants and some aspects of the pathomechanism of the deficiency. Blut. 1979;39(6):405-17. Review. 7. Kanno H, Fujii H, Hirono A, Ishida Y, Ohga S, Fukumoto Y et al. Molecular analysis of glucose phosphate isomerase deficiency associated with hereditary hemolytic anemia. Blood. 1996;88(6):2321-5 8. Kugler W, Breme K, Laspe P, Muirhead H, Davies C, Winkler H et al. Molecular basis of neurological dysfunction coupled with haemolytic anaemia in human glucose-6-phosphate isomerase (GPI) deficiency. Hum Genet. 1998;103(4):450-4 9. Urrutia R. KRAB-containing zinc-finger repressor proteins. Genome Biol. 2003;4(10):231. Review. 10. Emerson RO, Thomas JH. Adaptive evolution in zinc finger transcription factors. PLoS Genet. 2009;5(1):e1000325. Epub 2009 Jan 2. 11. Lugtenberg D, Yntema HG, Banning MJ, Oudakker AR, Firth HV, Willatt L, et al. ZNF674: a new kruppel-associated box-containing zinc-finger gene involved in nonsyndromic X-linked mental retardation. Am J Hum Genet. 2006;78(2):265-78. 12. Kleefstra T, Yntema HG, Oudakker AR, Banning MJ, Kalscheuer VM, Chelly Jet al. Zinc finger 81 (ZNF81) mutations associated with X-linked mental retardation. J Med Genet 2004;41:394¡V399 13. Shoichet SA, Hoffmann K, Menzel C, Trautmann U, Moser B, Hoeltzenbein M et al. Mutations in the ZNF41 gene are associated with cognitive deficits: identification of a new candidate for X-linked mental retardation. Am J Hum Genet 2003;73:1341¡V1354. 14. Marnef A, Sommerville J, Ladomery . RAP55: Insights into an evolutionarily conserved protein family. Int J Biochem Cell Biol. E pub 2008 Aug 3. 15. Yang WH, Yu JH, Gulick T, Bloch KD, Bloch DB. RNA-associated protein 55 (RAP55) localizes to mRNA processing bodies and stress granules. RNA. 2006;12(4):547-54. 16. Lois LM, Lima CD. Structures of the SUMO E1 provide mechanistic insights into SUMO activation and E2 recruitment to E1. EMBO J. 2005;24(3):439-51. 17. Li T, Santockyte R, Shen RF, Tekle E, Wang G, Yang DC, Chock PB. A general approach for investigating enzymatic pathways and substrates for ubiquitin-like modifiers. Arch Biochem Biophys. 2006;453(1):70-4.

The statement that "an important predictor of dyslexia, phonological awareness, can be understood as poor auditory structuring ability applied to language" raises some questions for me.

/1/ If dyslexia correlates with poor auditory structuring ability, it seems very strange that some alleles leading to dyslexia would also lead to musical ability. One would generally assume that musical ability would involve making good (rather than poor) use of auditory information. This leads me to ask whether your research plans to find out what particular aspects of human auditory ability (in people with the music/dyslexia alleles) might enhance musical learning/performance while simultaneously worsening literacy learning/performance.

/2/ Does the possession of phonological awareness indeed predict the possession of poor reading ability? From other research, I had always assumed that possessing phonological awareness went along with good (rather than poor) reading ability.

/3/ Similarly, it seems strange to see phonological awareness (rather than its absence) described as "poor auditory structuring ability." Can you please clarify how the ability to isolate and sequence the successive sounds within a spoken word would constitute "poor" (rather than good) "auditory structuring ability"?