The study of variation found in DNA is fundamental in human genetic studies. Single nucleotide polymorphisms (SNPs) are simple to document because they can be captured in single DNA sequence reads. Larger structural variation including duplications, insertions, deletions, termed as copy number variation (CNV), inversions and translocations are more challenging to discover. Recent studies using microarray and sequencing technologies have demonstrated the prevalence of structural variation in humans. They can disrupt genic and regulatory sequences, be associated with disease, and fuel evolution. Therefore, it is important to identify and characterize both SNPs and structural variants to fully understand their impact. This thesis presents the analysis of structural variation in the human genome. The primary DNA sample used for my experiments is the DNA of J. Craig Venter, also termed HuRef. It was the first personal human genome sequenced. I combined computational re-analysis of sequence data with microarray-based analysis, and detected 12,178 structural variants covering 40.6 Mb that were not reported in the initial sequencing study. The results indicated that the genomes of two individuals differed 1.3% by CNV, 0.3% by inversion and 0.1% by SNP. Structural variation discovery is dependent on the strategy used. No single approach can readily capture all types of variation, and a combination of strategies is required. I analyzed the formation mechanisms of all HuRef structural variants. The results showed that the relative proportion of mutational processes changed across size range: the majority of small variants (<1kb) were associated with nonhomologous processes and microsatellite events; median size variants (<10kb) were commonly related to minisatellites and retrotransposons; and large variants were associated with nonallelic homologous recombination. Eight new breakpoint-resolved HuRef inversions were genotyped in populations to elucidate these understudied variants. I discovered that the structures of inversion could be complex, could create conjoined genes, and their frequencies could exhibit population differentiation. The data here contributes to our understanding of structural variation in humans. It shows the need to use multiple strategies to identify variants, and it emphasizes the importance to examine the full complement of variation in all biomedical studies.

Copy-number variation (CNV) has been implicated in many complex diseases. It is of great interest to detect and locate such regions through genetic association testings. However, the association testings are complicated by the fact that CNVs usually span multiple markers and thus such markers are correlated to each other. To overcome the difficulty, it is desirable to pool information across the markers. In this thesis, we propose a kernel-based method for aggregation of marker-level tests, in which first we obtain a bunch of p-values through association tests for every marker and then the association test involving CNV is based on the statistic of p-values combinations. In addition, we explore several aspects of its implementation. Since p-values among markers are correlated, it is complicated to obtain the null distribution of test statistics for kernel-base aggregation of marker-level tests. To solve the problem, we develop two proper methods that are both demonstrated to preserve the family-wise error rate of the test procedure. They are permutation based and correlation base approaches. Many implementation aspects of kernel-based method are compared through the empirical power studies in a number of simulations constructed from real data involving a pharmacogenomic study of gemcitabine. In addition, more performance comparisons are shown between permutation-based and correlation-based approach. We also apply those two approaches to the real data. The main contribution of the dissertation is the development of marker-level association testing, a comparable and powerful approach to detect phenotype-associated CNVs. Furthermore, the approach is extended to high dimension setting with high efficiency.

L’hypertension artérielle essentielle (HTA) est une pathologie complexe, multifactorielle et à forte composante génétique. L’impact de la variabilité dans le nombre de copies sur l’HTA est encore peu connu. Nous envisagions que des variants dans le nombre de copies (CNVs) communs pourraient augmenter ou diminuer le risque pour l’HTA. Nous avons exploré cette hypothèse en réalisant des associations pangénomiques de CNVs avec l’HTA et avec l’HTA et le diabète de type 2 (DT2), chez 21 familles du Saguenay-Lac-St-Jean (SLSJ) caractérisées par un développement précoce de l’HTA et de la dyslipidémie. Pour la réplication, nous disposions, d’une part, de 3349 sujets diabétiques de la cohorte ADVANCE sélectionnés pour des complications vasculaires. D’autre part, de 187 sujets de la cohorte Tchèque Post-MONICA (CTPM), choisis selon la présence/absence d’albuminurie et/ou de syndrome métabolique. Finalement, 134 sujets de la cohorte CARTaGENE ont été analysés pour la validation fonctionnelle. Nous avons détecté deux nouveaux loci, régions de CNVs (CNVRs) à effets quantitatifs sur 17q21.31, associés à l’hypertension et au DT2 chez les sujets SLSJ et associés à l’hypertension chez les diabétiques ADVANCE. Un modèle statistique incluant les deux variants a permis de souligner le rôle essentiel du locus CNVR1 sur l’insulino-résistance, la précocité et la durée du diabète, ainsi que sur le risque cardiovasculaire. CNVR1 régule l’expression du pseudogène LOC644172 dont le dosage est associé à la prévalence de l’HTA, du DT2 et plus particulièrement au risque cardiovasculaire et à l’âge vasculaire (P<2×10-16). Nos résultats suggèrent que les porteurs de la duplication au locus CNVR1 développent précocement une anomalie de la fonction bêta pancréatique et de l’insulino-résistance, dues à un dosage élevé de LOC644172 qui perturberait, en retour, la régulation du gène paralogue fonctionnel, MAPK8IP1. Nous avons également avons identifié six CNVRs hautement hérités et associés à l'HTA chez les sujets SLSJ. Le score des effets combinés de ces CNVRs est apparu positivement et étroitement relié à la prévalence de l’HTA (P=2×10-10) et à l’âge de diagnostic de l’HTA. Dans la population SLSJ, le score des effets combinés présente une statistique C, pour l’HTA, de 0.71 et apparaît aussi performant que le score de risque Framingham pour la prédiction de l’HTA chez les moins de 25 ans. Un seul nouveau locus de CNVR sur 19q13.12, où la délétion est associée à un risque pour l’HTA, a été confirmé chez les Caucasiens CTPM. Ce CNVR englobe le gène FFAR3. Chez la souris, il a été démontré que l’action hypotensive du propionate est en partie médiée par Ffar3, à travers une interférence entre la flore intestinale et les systèmes cardiovasculaire et rénal. Les CNVRs identifiées dans cette étude, affectent des gènes ou sont localisées dans des QTLs reliés majoritairement aux réponses inflammatoires et immunitaires, au système rénal ainsi qu’aux lésions/réparations rénales ou à la spéciation. Cette étude suggère que l’étiologie de l’HTA ou de l’HTA associée au DT2 est affectée par des effets additifs ou interactifs de CNVRs. / Essential hypertension (HT) is a multifactorial complex disease with a strong genetic component. However, little is known about the effects of copy number variance on HT. We hypothesized common Copy Number Variants (CNVs) could increase or decrease the risk for HT. We performed GWAS of CNVs with HT and, with HT and Type 2 Diabetes (T2D), in 21 families of the Saguenay-Lac-St-Jean region of Quebec (FC) affected by early-onset hypertension and dyslipidemia. Replication was tested in a cohort of 3349 unrelated diabetic subjects of Caucasian origin from the ADVANCE trial. Replication was also tested in 187 individuals from the Czech Post-Monica (CPM) cross-sectional survey, ascertained by the presence/absence of albuminuria and/or metabolic syndrome. We performed locus-specific transcriptional analyses in 134 subjects from the CARTaGENE population cohort. We identified two CNV Regions (CNVRs), at 17q21.31, associated with HT and T2D in FC and associated with hypertension in ADVANCE diabetics. A statistical model of association including both CNVRs underlined the main effect size of CNVR1 on insulin resistance, T2D early onset and duration, and risk for cardiovascular diseases (CVD). CNVR1 appeared to influence LOC644172 expression, whose transcript abundance was associated with the prevalence of HT and T2D, and strongly with the risk of CVD and vascular age (P<2×10-16). Our results suggest carriers of copy-number gain at these 17q21.31 loci, principally at the CNVR1 locus, undergo premature β-cell functional deregulation and insulin resistance, due to increase dosage of the LOC644172 pseudogene, which might in turn affect the regulation of expression of its functional paralog, MAPK8IP1. We also report six different CNVR loci, highly heritable and contributing to the risk of hypertension, in French Canadians. The combined CNV risk score appeared robustly related to prevalence of hypertension (p=2×10-10) and age at diagnosis of hypertension. In FC, this combined CNV risk score model showed a C-statistic of 0.71 for HT and appeared as powerful as Framingham HT risk score in predicting hypertension in individuals aged less than 25. We validated the association of a new locus, 19q13.12 deletion-CNVR, with hypertension, in CPM. FFAR3 surrounds this 19q13.12 deletion-CNVR. It has been demonstrated that in mice, a portion of propionate hypotensive effect is mediated by Ffar3, and involves a cross-talk between the gut microbiota and the renal-cardiovascular system. The identified CNVRs appear to influence genes and QTLs mainly related to immune and inflammatory responses and renal damaged and repair. Some CNVRs are exclusive to primates. This study suggests that additive and interactive actions of multiple copy-number variants are involved in the etiology of hypertension or of hypertension associated with T2D.

The study of variation found in DNA is fundamental in human genetic studies. Single nucleotide polymorphisms (SNPs) are simple to document because they can be captured in single DNA sequence reads. Larger structural variation including duplications, insertions, deletions, termed as copy number variation (CNV), inversions and translocations are more challenging to discover. Recent studies using microarray and sequencing technologies have demonstrated the prevalence of structural variation in humans. They can disrupt genic and regulatory sequences, be associated with disease, and fuel evolution. Therefore, it is important to identify and characterize both SNPs and structural variants to fully understand their impact. This thesis presents the analysis of structural variation in the human genome. The primary DNA sample used for my experiments is the DNA of J. Craig Venter, also termed HuRef. It was the first personal human genome sequenced. I combined computational re-analysis of sequence data with microarray-based analysis, and detected 12,178 structural variants covering 40.6 Mb that were not reported in the initial sequencing study. The results indicated that the genomes of two individuals differed 1.3% by CNV, 0.3% by inversion and 0.1% by SNP. Structural variation discovery is dependent on the strategy used. No single approach can readily capture all types of variation, and a combination of strategies is required. I analyzed the formation mechanisms of all HuRef structural variants. The results showed that the relative proportion of mutational processes changed across size range: the majority of small variants (<1kb) were associated with nonhomologous processes and microsatellite events; median size variants (<10kb) were commonly related to minisatellites and retrotransposons; and large variants were associated with nonallelic homologous recombination. Eight new breakpoint-resolved HuRef inversions were genotyped in populations to elucidate these understudied variants. I discovered that the structures of inversion could be complex, could create conjoined genes, and their frequencies could exhibit population differentiation. The data here contributes to our understanding of structural variation in humans. It shows the need to use multiple strategies to identify variants, and it emphasizes the importance to examine the full complement of variation in all biomedical studies.

Mantle cell lymphoma (MCL) and chronic lymphocytic leukemia (CLL) both belong to the group of mature B-cell malignancies. However, MCL is typically clinically aggressive while the clinical course of CLL varies. CLL can be divided into prognostic subgroups based on IGHV mutational status and into multiple subsets based on closely homologous (stereotyped) B-cell receptors. In paper I we investigated 31 MCL cases using high-density 250K single-nucleotide polymorphism arrays and gene expression arrays. Although most copy-number aberrations (CNAs) were previously reported in MCL, a novel deletion was identified at 20q (16%) containing the candidate tumor suppressor gene ZFP64. A high proliferation gene expression signature was associated with poor prognosis, large CNAs, 7p gains and 9q losses. Losses at 1p/8p/13q/17p were associated with increased genomic complexity. In paper II we sequenced exons 4 to 8 of the TP53 gene in 119 MCL cases. 17p copy-number status was known from previous studies or determined by real-time quantitative polymerase chain reaction. TP53 mutations were detected in 14% of cases and were strongly associated with poor survival while 17p deletions were more common (32%) but did not predict survival. In papers III and IV we applied high-resolution genomic 27K methylation arrays to 20 MCL and 39 CLL samples. In paper III MCL displayed a homogenous methylation profile without correlation with the proliferation signature whereas MCL was clearly separated from CLL. Gene ontology analysis revealed enrichment of developmental genes, in particular homeobox transcription factor genes, among targets methylated in MCL. In paper IV we compared three different stereotyped CLL subsets: #1 (IGHV unmutated), #2 (IGHV3-21) and #4 (IGHV mutated). Many genes were differentially methylated between each two subsets and immune response genes (e.g. CD80 and CD86) were enriched among genes methylated in subset #1 but not in subsets #2/#4. In summary, CNAs were frequent and not random in MCL. Specific CNAs correlated with a high proliferation gene expression signature or genomic complexity. TP53 mutations predicted short survival whereas 17p deletions did not. A high proliferation signature was not associated with differential DNA methylation in MCL, which demonstrated a homogeneous methylation pattern. In contrast, genomic methylation patterns differed between MCL and CLL and between stereotyped CLL subsets.

The availability of DNA samples that are of adequate quality and quantity is essential for any genetic analysis. The fields of forensic biology and clinical diagnostic pathology testing often suffer from limited samples that yield insufficient DNA material to allow extensive analysis. This study examined the utility of a recently introduced whole genome amplification method termed Multiple Displacement Amplification (MDA) for amplifying a variety of limited sample types that are commonly encountered in the fields of forensic biology and clinical diagnostics. The MDA reaction, which employs the highly processive bacteriophage φ29 DNA polymerase, was found to generate high molecular weight template DNA suitable for a variety of downstream applications from low copy number DNA samples down to the single genome level. MDA of single cells yielded sufficient DNA for up to 20,000,000 PCR assays, allowing further confirmatory testing on samples of limited quantities or the archiving of precious DNA material for future work. The amplification of degraded DNA material using MDA identified a requirement for samples of sufficient quality to allow successful synthesis of product DNA templates. Furthermore, the utility of MDA products in comparative genomic hybridisation (CGH) assays identified the presence of amplification bias. However, this bias was overcome by introducing a novel modification to the MDA protocol. Future directions for this work include investigations into the utility of MDA products in short tandem repeat (STR) assays for human identifications and application of the modified MDA protocol for testing of single cell samples for genetic abnormalities.

L'autisme est un trouble du développement du système nerveux central défini par des altérations des interactions sociales et de la communication, et par des comportements restreints et répétitifs. Sa prévalence est actuellement évaluée jusqu'à 1% dans la population générale. L'autisme est caractérisé par une grande hétérogénéité sur les plans phénotypiques et génétiques. À ce jour, plus de 300 gènes candidats ont été caractérisés soit par des variations du nombre de copies (CNV) et/ou des variations nucléotidiques (SNV). Leur identification a permis de mettre en évidence une contribution significative de mutation de novo, ainsi que l'implication de voies physiopathologiques cibles, en particulier la densité post-synaptique (PSD). / Autism is a developmental disorder of the central nervous system defined by impairments in social interaction and communication, and by restricted and repetitive behavior. Its prevalence is currently estimated at around 1% in the general population. Autism is characterized by a wide heterogeneity at both phenotypic and genetic level. To date, more than 300 candidate genes were characterized either by copy number variations (CNV) and/or nucleotide variations (SNV). Their identification has highlighted a significant contribution of de novo mutations, as well as the involvement of targeted pathophysiological pathways, particularly post-synaptic density (PSD).

Abstract Breast cancer is the most common cancer among women, and inherited predisposition is one of the major recognized causes of increased breast cancer risk. Only about half of the hereditary cases are explained by mutations in the known susceptibility genes, including the DNA damage response genes BRCA1, BRCA2 and PALB2, leaving the majority still uncovered. Identification of the missing genetic predisposing factors is important for more effective diagnostics and counseling of the risk families, and also for better understanding of the etiology and cellular characteristics of breast cancer. The first aim of this study was to investigate the cancer associations of six rare germline copy number variant (CNV) deletions, which were previously identified in breast cancer patients by a genome-wide microarray approach. The second aim was to identify novel susceptibility alleles, both protein-truncating variants and missense mutations, by next-generation sequencing (NGS) of nearly 800 DNA damage response genes in 189 hereditary breast cancer patients. The cancer-associations of all selected candidate alleles (6 CNVs, 39 protein-truncating variants and 35 missense mutations) were studied by case-control approach using DNA samples from several hundred breast cancer patients and healthy controls. The prevalence of the studied CNVs did not significantly differ between the cases and controls, but when studying the associations with specific clinical parameters, deletion in the CYP2C19 gene showed enrichment in the breast cancer patients with hormonally triple-negative tumors (p=0.021). As CYP2C19 functions in estrogen metabolism, the results indicate that disturbance of hormonal balance due to enzyme defects may predispose specifically to the estrogen receptor-negative subtype of breast cancer. Two protein truncating-variants, TEX15 c.7253dupT and FANCD2 c.2715+1G&gt;A showed significant breast cancer association in the Northern Finnish cohort (p=0.018 and p=0.036, respectively). Similarly, two of the studied missense variants, RECQL p.Ile156Met (p=0.043) and POLG p.Leu392Val (p=0.010), were enriched in the breast cancer cases. Thus, this study provided novel connections between increased breast cancer risk and inherited mutations in TEX15, FANCD2 and POLG genes, and further supported the recently established role of RECQL as a breast cancer susceptibility gene. / Tiivistelmä Rintasyöpä on naisten yleisin syöpä, ja perinnöllinen alttius on yksi merkittävimmistä sairastumisriskiin vaikuttavista tekijöistä. Tunnetuimpia alttiustekijöitä ovat mutaatiot BRCA1-, BRCA2- ja PALB2-DNA-vauriovastegeeneissä, mutta ne yhdessä muiden altistavien geenimutaatioiden kanssa selittävät kuitenkin vain noin puolet perinnöllisistä rintasyöpätapauksista. Uusien alttiusgeenien löytäminen mahdollistaa tehokkaamman diagnostiikan ja korkeassa syöpäriskissä olevien sukujen perinnöllisyysneuvonnan, sekä auttaa ymmärtämään syvemmin rintasyövän etiologiaa ja syntymekanismeja solutasolla. Tämän väitöskirjan ensimmäisenä päämääränä oli tutkia tarkemmin aiemmin genominlaajuisella mikrosirumenetelmällä rintasyöpäpotilailta tunnistettujen harvinaisten perinnöllisten DNA-kopiolukuvariaatioiden (CNV) yhteyttä rintasyöpäriskiin. Toisena tavoitteena oli tunnistaa uusia rintasyöpäalttiusalleeleja, sekä proteiinitrunkaatioita että missense-mutaatioita, hyödyntämällä uuden sukupolven sekvensointitekniikkaa, jonka avulla tutkittiin mutaatioita lähes 800 DNA-vauriovastegeenistä 189 pohjoissuomalaiselta rintasyöpäpotilaalta. Valittujen kandidaattialleelien (6 deleetion aiheuttavaa CNV:tä, 39 proteiinitrunkaatiota ja 35 missense-mutaatiota) yhteyttä rintasyöpään tutkittiin tapaus-verrokkimenetelmällä käyttäen DNA-näytteitä usealta sadalta rintasyöpäpotilaalta ja terveeltä kontrollihenkilöltä. Tutkittujen CNV:iden esiintyvyydessä ei ollut merkitseviä eroja potilaiden ja kontrollien välillä, mutta tarkasteltaessa yhteyttä potilaiden kasvaimista saatuihin kliinisiin parametreihin, deleetio CYP2C19-geenissä oli yleisempi hormonaalisesti kolmoisnegatiivisissa rintatuumoreissa kuin muissa tuumorityypeissä (p=0.021). Koska CYP2C19 on estrogeenimetaboliaan osallistuva entsyymi, sen viallinen toiminta voi mahdollisesti altistaa erityisesti estrogeenireseptorinegatiiviselle rintasyövälle. Kaksi tutkituista proteiinitrunkaatioista, TEX15 c.7253dupT ja FANCD2 c.2715+1G&gt;A, olivat rikastuneet perinnöllisessä rintasyöpäpotilasaineistossa verrattuna kontrolleihin (p=0.018 ja p=0.036). Myös kaksi missense-alleelia, RECQL p.Ile156Met (p=0.043) ja POLG p.Leu392Val (p=0.010), olivat yleisempiä rintasyöpäpotilailla. Tulokset osoittivat uuden yhteyden kohonneen rintasyöpäriskin ja perinnöllisten muutosten TEX15-, FANCD2- ja POLG-geenien välillä, sekä tukivat aiempia tutkimustuloksia, joiden mukaan RECQL on kohtalaisen riskin rintasyöpäalttiusgeeni.

Peu d’hommes sont touchés par les maladies auto-immunes (MAI), maladies où la réponse immune est très forte et attaque l’hôte. La polyarthrite rhumatoïde (PR), une maladie inflammatoire chronique, suit cette règle avec 3 femmes pour 1 homme atteint. Dans cette thèse, nous analysons les différences en fonction du sexe et les raisons d’un tel biais. D’après des observations chez l’animal, nous nous sommes demandés si les rares hommes atteints de PR ont une augmentation du nombre de copies d’un gène impliqué dans la réponse immune et porté par le chromosome (Chr) X. Contrairement aux femmes, les hommes n’ont qu’un Chr X et de ce fait qu’une copie de ce gène. Cependant, nous avons montré par différentes techniques, que ces patients avaient 10% de cellules portant 2 copies de ce gène, et que cette augmentation venait de cellules ayant 2 Chr X. Nos recherches soulignent l’importance du Chr X dans l’auto-immunité et ouvrent un nouveau champ d'investigation pour les hommes atteints de MAI. / Few men are affected by autoimmune diseases (AID), diseases where the immune response is very strong and attacks the host. Rheumatoid arthritis (RA), a chronic inflammatory disease, follows this rule with 3 women affected for 1 man. In this thesis, we analyse gender differences and the reasons for such bias. Based on observations in animals, we wondered whether the rare men with RA have an increased copy number of a gene involved in the immune response and carried by the X chromosome (Chr). Unlike women, men have only one X Chr and one copy of this gene. However, we showed by different techniques that these patients had 10% of cells carrying 2 copies of this gene, and that this increase came from cells with 2 X Chr. Our research emphasizes the importance of the X Chr in autoimmunity and opens up a new field of investigation for men with AID.

A perda auditiva é o defeito mais comum ao nascimento e cerca de 70 milhões de pessoas no mundo apresentam algum grau de perda auditiva. Além da alta incidência, as implicações da perda auditiva na linguagem, na cognição e no desenvolvimento emocional e social reforçam sua importância. No entanto, em grande parte dos pacientes, a causa da deficiência auditiva não é esclarecida. Nós usamos hibridação comparativa do genoma baseada em arrays (Array Comparative Genomic Hybridization aCGH) para investigar alterações no número de cópias de segmentos de DNA (Copy Number Variation CNV) em 31 indivíduos que apresentavam deficiência auditiva e sinais clínicos adicionais, mas que não puderam ser classificados em síndrome conhecida. A escolha de indivíduos sindrômicos se baseou no pressuposto de que, em média, apresentam alterações genômicas maiores e, portanto, mais provavelmente detectáveis com o uso de aCGH de 1 Mb, que era a plataforma disponível no início do projeto. CNVs não descrita em bancos de dados de indivíduos normais foram identificadas em oito pacientes, quatro delas ocorreram de novo enquanto as outras quatro foram herdadas de um genitor fenotipicamente normal. As alterações de novo definem segmentos cromossômicos que provavelmente contém genes relacionados à deficiência auditiva e sensíveis a dose, especificamente: 1q23.3-q25.2, 2q22q23, 6p25.3 e 11q13.2-q13.4. As alterações raras identificadas tanto nos pacientes quanto em um genitor normal poderiam ser um evento ao acaso, sem papel na deficiência auditiva; no entanto, a possibilidade de que essas alterações possam funcionar como fatores de predisposição não podem ser descartadas. Se considerarmos apenas as CNVs de novo como causativas dos fenótipos investigados, detectamos quatro pacientes portadores entre os 31 investigados (13%). Se considerarmos também as CNVs herdadas como possivelmente causativas, a taxa de desequilíbrios cromossômicos associados à surdez será de 26%. Esses resultados são provavelmente uma substimativa e esses números seriam possivelmente maiores com o uso de uma das plataformas de alta resolução disponíveis atualmente. Esses resultados, embora limitados, indicam que investigação por aCGH em pacientes com surdez sindrômica idiopática está entre os testes mais eficientes para detectar etiologia dos fenótipos, devendo ser incorporado à rotina no diagnóstico e aconselhamento genético. / Hearing loss is the most common congenital deficiency and about 70 million people worldwide present some degree of hearing impairment. In addition to its high incidence, hearing loss impacts language, cognition and social and emotional development. However, in a large proportion of patients, the cause of the hearing deficiency cannot be elucidated. We screened copy number changes by 1 Mb-array Comparative Genomic Hybridization (aCGH) in 31 individuals with syndromic hearing impairment whose clinical features were untypical for known disorders. The choice of evaluating syndromic rather than non-syndromic individuals was based on the assumption that they are more likely to carry larger genomic alterations which could be more easily detected by the comparatively low resolution 1 Mb aCCG, which was the available platform when this project started. Copy number changes (CNV) not documented in the database of normal individuals were detected in eight patients, four de novo imbalances and four inherited from a normal parent. The de novo alterations define candidate chromosome segments likely to harbor dosage sensitive genes related to hearing impairment, namely 1q23.3-q25.2, 2q22q23, 6p25.3 and 11q13.2- q13.4. The rare imbalances also present in normal parents might be casually associated with hearing impairment, but also have a possible role as a predisposition factor. When only the de novo CNVs were considered causative for the disease phenotypes, our study revealed relevant copy number changes in 4 patients (13%). If we also count the rare CNVs that had been inherited as possibly causative, the frequency of chromosome imbalances associated with syndromic deafness in our sample becomes 26%. These figures are probably underestimates and will probably become larger when high resolution oligoarray platforms are applied. These results indicate that aCGH is an efficient tool for defining the etiology of syndromic deafness and its use in routine diagnosis of hearing impairment and for genetic counseling is highly recommended.