Dissertations / Theses on the topic 'Genome-wide copy number analysis'

DNA copy number change is an important form of structural variation in human genome. Somatic copy number alterations (CNAs) can cause over expression of oncogenes and loss of tumor suppressor genes in tumorigenesis. Recent development of SNP array technology has facilitated studies on copy number changes at a genome-wide scale, with high resolution. Quantitative analysis of somatic CNAs on genes has found broad applications in cancer research. Most tumors exhibit genomic instability at chromosome scale as a result of dynamically accumulated genomic mutations during the course of tumor progression. Such higher level cancer genomic characteristics cannot be effectively captured by the analysis of individual genes. We introduced two definitions of chromosome instability (CIN) index to mathematically and quantitatively characterize genome-wide genomic instability. The proposed CIN indices are derived from detected CNAs using circular binary segmentation and wavelet transform, which calculates a score based on both the amplitude and frequency of the copy number changes. We generated CIN indices on ovarian cancer subtypesâ copy number data and used them as features to train a SVM classifier. The experimental results show promising and high classification accuracy estimated through cross-validations. Additional survival analysis is constructed on the extracted CIN scores from TCGA ovarian cancer dataset and showed considerable correlation between CIN scores and various events and severity in ovarian cancer development. Currently our methods have been integrated into G-DOC. We expect these newly defined CINs to be predictors in tumors subtype diagnosis and to be a useful tool in cancer research.
Master of Science

The genetic basis of the different cancer phenotypes has been a continuous and accelerating subject of investigation. Data accumulated thanks to recently introduced genome-wide scanning technologies have revealed that human diversity and diseases susceptibility is also greatly influenced by structural alterations in the human genome, such as DNA copy number variants (CNVs) and copy number alterations (CNAs), which influence gene expression in both healthy and pathological cells. Our research aims to investigate the influence of structural alterations on gene expression in cancer cells using SNP microarray data. Specifically, we focus on analyzing DNA copy number alternations (CNAs), which can significantly influence gene expression in cancer cells. Several cancer-predisposing mutations affect genes that are responsible for maintaining the integrity of the chromosomes during cell division, which can result in translocations, gains or losses of large parts of chromosome. To our knowledge, there have been no publications that link whole-genome copy number alterations in cancer to gene expression variations using the full range of possibilities offered by SNP arrays. The accurate use of SNP arrays in the analysis of cancer has been difficult due to tumour purity, tumour heterogeneity, aneuploidy/polyploidy and complex patterns of CNA and loss-of-heterozygosity (LOH). In our work, we use and further extend a recently developed novel tool for tumour genome profiling called OncoSNP (Yau, Mouradov et al. 2010), in order to resolve some of those problems and accurately estimate copy number alterations (CNA) and loss-of-heterozygosity (LOH) from SNP array data in cancer cell samples. The methods developed in this thesis tackle the problem of cancer genomic investigation by developing and validating an extension (DPS smoothing) of a new method (OncoSNP). This approach is used in the analysis of global expression versus CNA patterns in experimental systems and large clinical datasets. We analyse various cancer SNP and gene expression arrays of increasing complexity and heterogeneity, starting with a dataset of head and neck squamous cell carcinoma (HNSCC) cell lines, followed by leukaemia samples and finally a large breast cancer dataset. The central findings of our research are multifold. In the HNSCC dataset we find that the level of genetic instability is not indicative of the pathological state; i.e. there are premalignant lesions displaying extensive mutations. However some genetic features are typical of certain lesion type; e.g. we consistently observe copy loss in the short arm of chromosome 3 in carcinoma. The pattern of homozygous deletion in the dataset reveals common deletion of cancer related genes, especially CDK4 (pI6). Furthermore we notice a significant positive correlation between the copy number and the expression on a systematic level. In Leukaemia, we do not observe extended uniparental disomy as previously published (Akagi, Shih et al. 2009) and expected. However large alterations (whole arm amplification) are observed in individual patients: copy loss in chromosome 7 (2 patients), copy gain in chromosome 8 (3 patients) as well as common alterations around the centromeres and telomeres. In the breast cancer dataset significantly different level of mutations were observed in the different subtypes in the cohort. Furthermore 499 genes were identified with significant correlation between their gene expression (GE) and underlying genomic alterations (either copy number (CN) or loss-of-heterozygosity (LOH)). Performing hierarchical clustering on the cohort using the 499 correlated genes enabled us to recover the subtypes' separation previously based on gene expression alone.

From 2007, there has been a huge proliferation in the discovery of genetic variants affecting human traits and diseases, achieved largely by the integration of multiple genome-wide association studies (GWAS) via meta-analysis. The principal objective of this thesis is to develop robust approaches for meta-analysis GWAS in order to reduce false positive findings and optimise statistical power. I consider both Single Nucleotide Polymorphism (SNP) and Copy Number Variant (CNV) GWAS. First, to gain background knowledge in GWAS and meta-analysis, I was involved in a large-scale meta-analysis GWAS to identify genetic variants associated with alcohol consumption, as the main statistical analyst. This study provided me with the opportunity to investigate ways of reducing the probability of false positive findings, via quality control procedures. The main discovery from the study was the identification of the Autism susceptibility candidate 2 gene (AUTS2) as associated with alcohol consumption at genome-wide significance. In the alcohol study, different phenotype transformations were applied to the data according to the inclusion or exclusion of non-drinkers, which led to questioning which transformation of skewed continuous phenotypes optimises statistical power in GWAS in general, forming the second major investigation in my thesis. It was shown that while the inverse normal transformation (INT) may not be the preferable choice of transformation in many epidemiological studies where effect sizes are large, its application to non-normal phenotypes in GWAS, where effect sizes are small and the priority is discovery over interpretability, may lead to an increase in the discovery of genetic variants affecting continuous traits. Finally, as knowledge about CNVs has accumulated in recent years, the meta-analysis of GWAS on CNVs has become a natural next step forward in the field. Therefore, I investigated and developed an approach to enable CNV meta-analysis to proceed in a similar way as SNP meta-analyses. This approach was developed into a software package, cnvPipe, which was applied to investigate CNVs associated with height and weight in the meta-analysis setting.

Rare copy number variants (CNV), defined as deletions and duplications of genetic material over 1,000 base pairs in length, have become the focus of considerable interest in psychiatric disorders, where a proportion of individuals harbour rare and de novo events not usually seen in controls. We have performed a genome wide association study of copy number variation in 3,106 cases of recurrent depressive disorder, 1,731 controls screened for a lifetime absence of psychiatric disorder, and 5,619 population controls from phase 2 of the Wellcome Trust Case Control Consortium. Analysing our data with the PennCNV method, we found an enrichment of rare deletion CNVs in our case cohort, especially when compared to our screened control cohort. This finding was supported by further analysis with the iPattern method, but not by the QuantiSNP method. We followed up a selection of cases and controls with a comparative genomic hybridisation (CGH) array focussed on the region 22qll.2, which is a neuro-gene rich region of the human genome under current active evolutionary selection and resident to a deletion syndrome which commonly manifests with psychiatric disorders. We found no significant differences in CNV burden between our case and control cohorts. Finally we ran association analyses with our CNV call sets, including a high quality intersected call set derived from all three methods, against various phenotypes obtained from a combined database of all studies that contributed samples to this GWAS. We found no associations that survived Bonferroni correction for multiple testing.

Genome wide studies have uncovered the existence of large-scale copy number variation (CNV) in the human genome. The human genome of different individuals was initially estimated to be 99.9% similar, but population studies on CNV have revealed that it is 12-16% copy number variable. Abnormal genomic CNVs are frequently found in subtelomeres of patients with mental retardation (MR) and other neurological disorders. Rearrangements of chromosome subtelomeric regions represent a high proportion of cytogenetic abnormalities and account for approximately 30% of pathogenic CNVs. Although DNA double strand breaks (DSBs) are implicated as a major factor in chromosomal rearrangements, the causes of chromosome breakage in subtelomeric regions have not been elucidated. But due to the presence of repetitive sequences in subtelomeres, we hypothesized that chromosomal rearrangements in these regions are not stochastic but driven by specific sequence motifs. In a collaborative effort with Dr. Rudd (Department of human genetics at Emory University), we characterized subtelomeric breakpoints on different chromosome ends in search of common motifs that cause double-strand breaks. Using a yeast-based gross chromosomal rearrangement (GCR) system, we have identified a subtelomeric breakage motif from chromosome 2 (2q SBM) with a GCR rate that is 340 fold higher than background levels. To determine if the fragility of 2q SBM was driven by the formation of secondary structures, the helicase activities of Sgs1 and Pif1 were disrupted. These helicases have been shown to destabilize DNA secondary structures such as G-quadruplex structures. Disruption of these helicases augmented chromosomal rearrangements induced by 2q SBM, indicating that these helicases are required for maintenance of this sequence. We also donwregulated replication fork components to determine if 2q SBM was imposing any problems to the replication fork machinery. Downregulation of replication fork components increased chromosomal rearrangements, indicating that intact replication fork was a critical determinant of 2q SBM fragility. Using a yeast-based functional assay, these experiments have linked human subtelomeric repetitive sequences to chromosomal breakage that could give rise to human CNV in subtelomeric regions.

ABSTRACT In the first part of our work we focused our attention on the biological question about the differences between two pathologies: T-cell lymphoblastic leukemia and T-cell lymphoblastic lymphoma. These two diseases share many features such as immunophenotypic features, lymphoblast morphology and clinical characteristics and are differentially diagnosed only on the base of bone marrow involvement. We tried to understand whether T-cell leukemia and lymphoma are a unique pathology with a different manifestation or whether they are two different diseases. The results obtained by gene expression profiling revealed an intrinsic difference in the expression of 78 genes between T-ALL and T-LBL. In particular since these genes belong to the angiogenesis and the chemotactic response we supposed that the two malignancies have different ability to respond to several cyto- and chemokines and that T-LBL need to modulate transcription to promote angiogenesis as well as to deal with hypoxic conditions. Also by analysis of copy number we were able to identify some abnormalities that seemed to be specific for each group regardless the limited data set of patients. Although this work provides additional elements in the characterization of these two pathologies, many studies have yet to be done, especially on the comprehension of the different capability of cells to migrate and invade the bone marrow compartment. The complete understanding of the molecular characteristics of T-LBL and T-ALL represents the driving element toward the design of fully successful therapeutic approaches. The second part of the study focused on the genetic characterization of two different groups of T-ALL patients on the basis of MRD response. With the aim to find biological correlates with the outcome for HR and nonHR patients, we performed many analyses, starting from copy number analysis to microRNAs expression profiling. Furthermore, we tried to integrate all the data in order to delineate common characteristics for each group of patients. First of all, the study of copy number revealed the presence of multiple abnormalities in all patients: we found known and unknown lesions, and in some cases we were able to associate them with HR or nonHR patients. The improvement of copy number results was obtained by the study of translocations. Also in this case we found one or more translocation in the majority of patients and we identified that the most recurrent were SIL-TAL1 and TLX3 translocations. Moreover, we tested the Notch1 mutations and, as expected, about 60% of patients were mutated for Notch1, with a tendency for Notch1 mutations to be more frequent in the nonHR group. The second step was the analysis of gene and microRNAs expression. The initial unsupervised analysis between HR and nonHR group failed to distinguish the two groups; but the successively supervised analysis revealed the distribution of MR patients in an equal manner between HR and SR patients. Thus an unsupervised analysis without MR patients showed a specific pattern of expression for each group (SR and HR). The GSEA analysis performed highlighted the enrichment of two specific pathways: the mir-215/192 pathway and the methylation pathway. The results obtained by the expression profile of about 700 microRNAs in the HR and nonHR group and those achieved by the combined analysis of GEP and microRNAs suggested miR-215 and miR-107 as the most differentially expressed and provided some possible target genes of these microRNAs. Moreover, to delineate specific pattern of expression not driven by MRD but by other alterations, we used the data derived from copy number, translocation and mutational analyses to supervise genetic subgroups. Significant results were obtained for Nocth1 mutated vs non-mutated, TLX3 translocated vs non-translocated, PTEN and LEF1 deleted vs non-deleted. We also tried to integrate all the data provided by both genomic and trasncriptomic analyses to understand whether the distribution of MR and the different signature of HR and SR were correlated with specific gene lesions. SIL-TAL1 fusion gene and the deletion of LEF1 and PTEN seemed to be specific for the HR group while the TLX3-translocation seemed to be peculiar for the nonHR group of patients. In conclusion HR and nonHR patients seemed to show some peculiar lesions and patterns of expression that could be justify the different response to therapy. In summary, several high throughput methodologies have been applied to the selected subgroup of patients to study the biological correlates of the different response to therapy. By this work we tried to provide a better characterization of T-ALL and to give a way of interpretation for the different outcome of T-ALL patients.

Introduction: Arrhythmogenic right ventricular cardiomyopathy (ARVC) is an inherited cardiac disease characterized by fibrofatty replacement of myocardial tissue and high incidence of serious ventricular tachyarrhythmias. To date, ten disease-genes have been identified, five of which encoding desmosomal proteins (PKP2, DSP, DSG2, DSC2 and JUP). Aim of the study: The study described in the present thesis aimed at determining the spectrum and prevalence of desmosomal mutations in 80 Italian unrelated index cases. Moreover, the identification of novel disease loci and genes in three ARVC families was attempted by integrating different approaches, including genome-wide linkage study, Copy Number Variations (CNVs) analysis, and exome-sequencing. Methods: Mutation screening of the desmosomal ARVC genes was performed by Denaturing High-Performance Liquid Chromatography (DHPLC) and direct sequencing. In order to map novel loci and, possibly, to identify novel genes, genome-wide linkage study and CNVs analysis was performed in three independent families with recurrence of ARVC, showing no mutations in any of the desmosomal genes. In addition, in Family #2 and in Family #3 a combined strategy of linkage analysis and exome sequencing was adopted to identify novel ARVC genes. Results: Mutation screening of five desmosomal ARVC genes in 80 Italian probands identified single point mutations in 32.5% of cases and multiple mutations in 12.5%. No mutations were identified in the remaining 55% of probands. The genes most frequently involved were PKP2, DSP and DSG2. Among the index cases negative for point mutations in the desmosomal genes, three belong to independent families with recurrence of ARVC. Assuming a further genetic heterogeneity in these families, a genome-wide scan was performed in order to identify novel loci and genes. In Family #1, a CNV involving a desmosomal ARVC gene was identified in all affected family members. In Family #2, a novel ARVC locus was mapped on chromosome 19 by linkage analysis. Finally, in Family #3, a possible novel candidate gene for ARVC was detected by a combined strategy of linkage analysis and exome-sequencing. Discussion: In the present study, genetic screening of five desmosomal ARVC genes in 80 Italian probands identified causative mutations in 45% of cases, mainly involving the “big3” genes PKP2, DSP, and DSG2, according to data reported in literature. Genetic analysis of available family members confirmed the high heterogeneity in the clinical expression of ARVC mutations even among relatives. Mutation screening of desmosomal genes failed to detect causative mutations in more than 50% of index cases, suggesting that additional and still unknown genes could be involved. In this perspective, a genome-wide scan was performed in three large ARVC families, showing no mutations in any of the desmosomal genes. In Family #1, a CNV was identified in a desmosomal ARVC gene, highlighting the importance of complementing the conventional mutation screening in ARVC genes with other approaches able to detect possible structural variations. In Family #2, genome-wide linkage results provided strong evidence for a novel ARVC locus on chromosome 19, highlighting the soundness of this strategy for identifying susceptibility regions in large, highly informative families and providing the basis for the identification of a novel disease gene. Finally, in Family #3, exome-sequencing identified a novel putative candidate gene for ARVC. Identification of novel ARVC genes is of great importance for understanding the molecular pathogenesis of this disease, as well as for increasing the power of genetic screening and developing successful targeted therapies
Introduzione: La Cardiomiopatia Aritmogena del Ventricolo Destro (ARVC) è una malattia ereditaria del muscolo cardiaco caratterizzata dalla progressiva perdita e sostituzione fibro-adiposa dei cardiomiociti che costituiscono la parete libera delventricolo destro. Tale disomogeneità del tessuto cardiaco altera la normale conduzione dell'impulso elettrico, determinando l'insorgenza di aritmie che occasionalmente portano a fibrillazione ventricolare e morte improvvisa per arresto cardiaco, soprattutto nei giovani e negli atleti. Attualmente, sono noti 10 geni implicati nella determinazione genetica dell’ARVC e cinque di questi codificano per proteine costituenti il desmosoma cardiaco: Placofilina-2 (PKP2), Desmoplachina (DSP), Desmogleina-2 (DSG2), Desmocollina-2 (DSC2) e Placoblobina (JUP). Scopo dello studio: Lo studio descritto nella presente tesi mira a valutare la prevalenza e lo spettro di mutazioni nei cinque geni ARVC desmosomali in un gruppo di 80 casi indice Italiani, non imparentati tra loro. Inoltre, in tre grandi famiglie con ricorrenza di casi ARVC e in cui non sono state identificate mutazioni nei geni desmosomali, è stata effettuata un'analisi genome-wide integrando diversi approcci, quali studio di linkage, analisi di Copy Number Variations (CNVs) e sequenziamento dell’esoma. Metodi: Lo screening per la ricerca di mutazioni nei cinque geni ARVC desmosomali ha coinvolto 80 casi indice ed è stato effettuato tramite analisi DHPLC (Denaturing High-Performance Liquid Chromatography) e sequenziamento diretto del DNA.I soggetti appartenenti a ciascuna delle tre famiglie selezionate per l'analisi genome-wide sono stati genotipizzati utilizzando un pannello di marcatori ad alta densità che include più di 370.000 polimorfismi di singolo nucleotide (SNPs) (Illumina HumanCNV370-Duo BeadChip). In ciascuna famiglia è stato effettuato uno studio di linkage ed un’analisi di CNVs. Inoltre, Nella Famiglia #2 e nella #3 l'identificazione del gene malattia è stata tentata integrando i risultati dello studio di linkage con i dati ottenuti dal sequenziamento dell'esoma di due soggetti affetti. Risultati: L'analisi delle sequenze codificanti dei geni PKP2, DSP, DSG2, DSC2 e JUP in 80 casi indice Italiani ha permesso di identificare mutazioni singole nel 32.5% dei casi e mutazioni multiple nel 12.5%. Il 55% dei probandi non è risultato portatore di alcuna mutazione nei geni desmosomali. La maggior parte delle mutazioni ha coinvolto i geni PKP2, DSP, DSG2, confermando i dati riportati in letteratura. Tra i casi indice in cui non sono state identificate mutazioni nei geni desmosomali, tre appartengono a famiglie indipendenti con ricorrenza di casi ARVC. In ognuna di queste famiglie è stata effettuata un'analisi genome-wide allo scopo di identificare nuovi loci e geni malattia. Nella Famiglia #1 è stata identificata una CNV che coinvolge uno dei geni ARVC desmosomali noti e co-segrega con il fenotipo patogeno. Nella Famiglia #2, l’analisi di linkage ha permesso di identificare un nuovo locus ARVC sul cromosoma 19. Infine, nella Famiglia #3 è stato identificato un nuovo potenziale gene candidato per l’ARVC. Discussione: L’analisi genetica delle sequenze codificanti dei cinque geni ARVC desmosomali, in un gruppo di 80 probandi italiani, ha permesso di identificare mutazioni nel 45% dei casi, confermando il prevalente coinvolgimento dei tre geni PKP2, DSP e DSG2, in accordo con i dati riportati in letteratura. L'analisi genetica dei familiari dei probandi ha confermato la penetranza incompleta e l'espressività variabile della malattia, anche all'interno della stessa famiglia. L'assenza di mutazioni in più del 50% dei casi suggerisce il coinvolgimento di altri geni nella determinazione genetica dell'ARVC. In quest' ottica, un'analisi genome-wide è stata effettuata in tre famiglie con ricorrenza di casi ARVC e in cui non sono state individuate mutazioni nei geni desmosomali. Nella Famiglia #1, l’identificazione di una CNV in uno dei geni ARVC desmosomali, presente in tutti i soggetti affetti, sottolinea l'importanza di associare alle metodiche tradizionali utilizzate per lo screening di mutazioni puntiformi approcci che permettano di identificare eventuali variazioni strutturali presenti nel genoma. Nella Famiglia #2, l'analisi di linkage ha fornito una significativa evidenza dell'esistenza di un nuovo locus ARVC sul cromosoma 19, fornendo le basi per l'identificazione di nuovi geni. Infine, nella Famiglia #3, il sequenziamento dell'esoma di due soggetti affetti ha identificato un nuovo gene come un possibile candidato per l'ARVC

The work comprised in this PhD thesis described the development of a novel mathematical and statistical framework to analyse and combine, at genome-wide level, gene expression profile and DNA copy number data obtained by high-throughput oligonucleotide microarray platform. This dual strategy is now considered the most effective to understand the genetic causes underlying neoplastic diseases and identify interesting regions and genes with potential clinical application as novel tumor markers. In this thesis, we applied this combined approach to study the clear cell renal carcinoma (ccRCC) pathology, using firstly a human metastatic cell line as in vitro model and then a collection of clinical tumor tissue samples. Considering the physical position of genes along the genome, high-throughput gene expression data were used to assemble a regional transcriptional activity profile. In the meantime, a genome-wide DNA copy number map was assembled by high-throughput SNP mapping technology, thus identifying recurrent aberrations that might be novel candidate regions characterizing all or subsets of ccRCC samples. To filter the large amount of array-based data and narrow down the hundreds of candidate regions to those whose altered expression level was attributable to underlying chromosomal alterations, regional gene expression data were combined with DNA copy number alteration map at genome-wide level. After confirming a strong association between aneuploidy and regional transcriptional activity profiles, we identified a set of regions showing concomitant DNA alteration and modulated expression level and, within, particularly interesting genes as novel candidate RCC-related markers. Overall, this study demonstrates the efficacy of the combination of DNA and RNA profiles to improve the specificity of analysis and increase the possibility of identifying the genetic causes underlying ccRCC pathology, so highlighting candidate genes that are actively involved in the causation or mainteinance of the malignant phenotype.

Copy number variation (CNV) is an important polymorphism that is associated with a wide range of traits in human, wild and livestock species. In chicken, an important source of animal protein and a developmental model organism, CNV is associated with several phenotypes and evolutionary footprints. However, identification and characterization of CNV inheritance on chicken genome lacks further investigation. We screened CNVs in chicken using two distinct populations with known pedigree. In 826 broilers we identified 25,819 CNVs (4,299 deletions and 21,520 duplications) of which 21,077 were inherited, 201 showed no inheritance and 4,541 were classified as de novo CNVs. In 514 F2 animals (layer and broiler cross) we identified 21,796 CNVs (2,254 deletions and 19,543 duplications) of which 18,230 were inherited, 587 not inherited and 2,979 were classified as de novo CNVs. After a strict filtering step to remove potential false positives and negative CNVs, only 220 (4.84%) and 430 (14.43%) de novo CNVs remained in the broiler and F2 populations, respectively. A total of 33.11% (50 out of 151) of the inherited CNVs identified in ten animals were validated by sequencing data. From the validated CNVs, 64% had more than 80% of their size (bp) validated. A total of 59% and 48.8% were classified as novel CNVs regions (CNVRs) in the broiler and F2, respectively. Considering the Bonferroni-corrected p-values for multiple testing and statistically significant p-values ≤ 0.01, we found two CNV segments significantly associated with breast weight, one with breast weight yield, six with breast meat weight, 18 CNV segments with breast meat yield, four with breast filet weight and two with breast yield. These CNV segments that were significantly associated overlapped with 181 protein-coding genes. The CNVseg 300, that was associated with all traits and encompass six CNVRs, overlapped a total of 26 protein-coding genes. Among these genes, the gene MYL1 (Myosin Light Chain 1) is expressed in the fast skeletal muscle fibers, and the genes MLPH (Melanophilin), PRLH (Prolactin Releasing Hormone) and RAB17 (Member RAS Oncogene Family), that were associated with the lavender phenotype (feather blue-grey color) and regulation of homeothermy and the metabolism. The present study improves our knowledge about CNV in the chicken genome and provides insight in the distribution and of different classes of CNVs, i.e. inherited and de novo CNVs, in two experimental chicken populations. In addition, the genome-wide association analyses were the first performed on broiler population with breast muscle traits, that are important characteristics for poultry production. The GWAS results allow us to understand the probably relationship between some genes and CNVRs that are significantly associated with breast muscle traits.
A variação de número de cópias (CNV) é um polimorfismo importante que está associado a uma ampla gama de características em seres humanos, espécies selvagens e domésticas. Em frango, que é uma importante fonte de proteína e considerado um modelo biológico, CNVs foram associados a vários fenótipos e passos evolutivos. No entanto, nenhum estudo foi realizado para a identificação e caracterização da herança da CNV no genoma da galinha. Identificamos as CNVs no genoma da galinha usando duas populações experimentais e com pedigree conhecido: uma população de frangos de corte e uma F2. Em 826 frangos de corte, identificamos 25.819 CNVs (4.299 deleções e 21.520 duplicações), dos quais 21.077 foram herdados, 201 não foram herdados e 4.541 foram CNVs denominados de novo. Em 514 animais F2, identificamos 21.796 CNVs (2.254 deleções e 19.543 duplicações) das quais 18.230 foram herdadas, 587 não foram herdadas e 2.979 foram de novo CNVs. Após a etapa de filtragem nos de novo CNVs, apenas 220 (4,84%) e 430 (14,43%) permaneceram nas populações de frango de corte e F2, respectivamente. Um total de 33,11% (50 de 151) das CNV identificadas por dados de genotipagem em dez animais foram validados por dados de sequenciamento. Dos validados, 64% tinham mais de 80% do tamanho (pb) validados. Um total de 59% e 48,8% foram classificados como novas regiões de CNVs (CNVRs) nas populações de frango de corte e F2, respectivamente. Considerando os p-values corrigidos por Bonferroni para testes múltiplos e estatisticamente significativos (≤ 0,01), encontramos dois segmentos de CNV significativamente associados ao peso do peito, um ao rendimento de peso de peito, seis ao peso de carne de peito, 18 ao rendimento de carne de peito, quatro ao peso de filé de peito e dois ao rendimento do filé de peito. Esses segmentos de CNV significativamente associados estão sobrepostos com 181 genes codificadores de proteínas. O CNVseg 300, que foi associado a todas as características e abrange seis CNVRs, foram sobrepostos a um total de 26 genes codificadores de proteínas. Entre estes genes, o gene MYL1 (Myosin Light Chain 1) é expresso nas fibras rápidas do músculo esquelético, e os genes MLPH (Melanophilin), PRLH (Prolactin Releasing Hormone) e RAB17 (Member RAS Oncogene Family), que foram anteiromente associados ao fenótipo de cor azul acinzentado de penas e à regulação da homeotermia e do metabolismo. O presente estudo melhora o conhecimento sobre CNVs no genoma de frango, especialmente sobre a distribuição de CNV herdadas, não herdadas e de novo, em duas populações experimentais de frango. Além disso, a associação genômica foi a primeira realizada na população de frangos de corte com características do músculo do peito, que são muito importantes para a avicultura. Os resultados do GWAS nos permitem compreender a provável relação entre alguns genes e CNVRs que foram significativamente associados às características do músculo do peito.

We developed a pipeline to identify the copy number polymorphisms (CNPs) in the Northern Swedish population using whole genome sequencing (WGS) data. Two different methodologies were applied to discover CNPs in more than 1,000 individuals. We also studied the association between the identified CNPs with the expression level of 438 plasma proteins collected in the same population. The identified CNPs were summarized and filtered as a population copy number matrix for 1,021 individuals in 243,987 non-overlapping CNP loci. For the 872 individuals with both WGS and plasma protein biomarkers data, we conducted linear regression analyses with age and sex as covariance. From the analyses, we detected 382 CNP loci, clustered in 30 collapsed copy number variable regions (CNVRs) that were significantly associated with the levels of 17 plasma protein biomarkers (p < 4.68×10-10).

Brain disorders present a significant burden on affected individuals, their families and society at large. Existing diagnostic tests suffer from a lack of genetic biomarkers, particularly for substance use disorders, such as alcohol dependence (AD). Numerous studies have demonstrated that AD has a genetic heritability of 40-60%. The existing genetics literature of AD has primarily focused on linkage analyses in small family cohorts and more recently on genome-wide association analyses (GWAS) in large case-control cohorts, fueled by rapid advances in next generation sequencing (NGS). Numerous AD-associated genomic variations are present at a common frequency in the general population, making these variants of public health significance. However, known AD-associated variants explain only a fraction of the expected heritability. In this dissertation, we demonstrate that systems biology applications that integrate evolutionary genomics, rare variants and structural variation can dissect the genetic architecture of AD and elucidate its heritability. We identified several complex human diseases, including AD and other brain disorders, as potential targets of natural selection forces in diverse world populations. Further evidence of natural selection forces affecting AD was revealed when we identified an association between eye color, a trait under strong selection, and AD. These findings provide strong support for conducting GWAS on brain disorder phenotypes. However, with the ever-increasing abundance of rare genomic variants and large cohorts of multi-ethnic samples, population stratification becomes a serious confounding factor for GWAS. To address this problem, we designed a novel approach to identify ancestry informative single nucleotide polymorphisms (SNPs) for population stratification adjustment in association analyses. Furthermore, to leverage untyped variants from genotyping arrays – particularly rare variants – for GWAS and meta-analysis through rapid imputation, we designed a tool that converts genotype definitions across various array platforms. To further elucidate the genetic heritability of brain disorders, we designed approaches aimed at identifying Copy Number Variations (CNVs) and viral insertions into the human genome. We conducted the first CNV-based whole genome meta-analysis for AD. We also designed an integrated approach to estimate the sensitivity of NGS-based methods of viral insertion detection. For the first time in the literature, we identified herpesvirus in NGS data from an Alzheimer’s disease brain sample. The work in this dissertation represents a three-faceted advance in our understanding of brain disease etiology: 1) evolutionary genomic insights, 2) novel resources and tools to leverage rare variants, and 3) the discovery of disease-associated structural genomic aberrations. Our findings have broad implications on the genetics of complex human disease and hold promise for delivering clinically useful knowledge and resources.

Les variations en nombre de copies (CNV) sont des gains ou pertes d’une séquence d’ADN et peuvent avoir un rôle dans la susceptibilité à certaines maladies. Les CNVs peuvent être détectés par les puces de SNPs de haute résolution en analysant les intensités des allèles avec des algorithmes de détection des CNVs tels que CNV partition, PennCNV et QuantiSNP. Dans cette thèse, nous avons évalué les performances de ces outils pour la détection des CNVs au niveau pangénomique et pour les tests d'association. Nous avons également étudié des stratégies d'association combinant les informations de l'allèle et du nombre de copies pour des SNP situés dans des CNV. Nous avons appliqué ces outils pour mener une étude d’association pan-génomique avec les CNV en utilisant les données de l'étude espagnole du cancer de lavessie (SBC)/EPICURO générées par la puce Illumina 1M.Nos résultats montrent une faible fiabilité et une faible sensibilité des algorithmes de détection des CNV. Dans la région du gène GSTM1 où un CNV très fréquent existe qui est associé au risque de cancer de la vessie, nous avons constaté que les algorithmes de détection des CNV ont de faibles performances. Néanmoins, l’utilisation de la mesure d'intensité des allèles dans les tests d'association peut alors être une alternative intéressante car cela nous a permis de détecter cette association connue. Pour les SNPs situés dans des CNVs, nous avons étudié plusieurs stratégies de tests d'association et nous avons montré que la plus puissante était d’utiliser un modèle avec deux termes correspondant respectivement à la somme et à la différence du nombre de copies des deux allèles. Finalement, en appliquant ces stratégies à l'étude (SBC)/EPICURO, nous avons identifié des CNVs potentiellement associés au risque de cancer de la vessie, ainsi que des SNP dont l'allèle et le nombre de copies pourraient être impliqués dans le risque de cancer de la vessie
Copy number variations (CNVs) are losses or gains of DNA sequences that may play a role in specific disease susceptibility. CNVs can be detected by high-resolution SNP-arrays through the analysis of allele intensities with CNV calling algorithms such as CNVpartition, PennCNV and QuantiSNP. In this thesis, we identified and assessed the performances of available tools for CNV calling and for association testing, at the genome-wide level. We also investigatedassociation strategies that combine information on both the allele and the number of copies for SNPs located in CNV regions. We applied these tools to conduct a genome-wide association study with CNV using data from the Spanish Bladder Cancer (SBC)/EPICURO Study generated by the Illumina 1M SNP-array. Our results showed a low reliability and a low sensitivity of the investigated CNV calling algorithms applied to SNP-array data. The GSTM1 locus shows a very frequent CNV that is associated with bladder cancer (BC) risk. We reported that the calling algorithms performed very poorly in identifying this CNV. We proposed using allele intensity measures (LRR) as a screening step to assess association as it allowed the detection of the GSTM1 CNV association with BC. To combine the allele and the number of copies for SNPs located in CNV regions, we investigated several strategies of association testing and we showed that the more powerfulone used a two-term model with the sum and the difference of the number of copies of both alleles. Finally, by applying these strategies to the (SBC)/EPICURO Study, we identified CNV regions potentially associated with BC risk, as well as SNPs for which both the allele and the number of copies could be involved in BC risk

Obesity is a serious public health crisis and recent estimates of its incidence are the highest in United States history, with 35% and 17% of American adults and children affected, respectively. The clinical definition of adult obesity is operationalized as a body mass index (BMI) greater than 30 kg/m2. Although the prevalence of common obesity has increased dramatically over the past 30 years–largely thought to be due to changes in the environment, such as high calorie diets and sedentary lifestyles—twin and family studies have shown consistently that relative body weight is under considerable genetic influence in both children and adults, with heritability estimates ranging from 40% to 90%. Elucidating the genetic and environmental liability to relative body weight is an important public health endeavor. To further our understanding of the genetics of BMI and common complex obesity, several studies are described that integrate clinical, twin, and genome-wide association (GWAS) methodology in the context of genetic risk scores, clinical risk prediction, development across adolescence into adulthood, and comorbidity with depression symptoms and smoking behavior. First, in two cross-sectional genetic association studies, the utility of genetic risk sum scores (GRSS) were assessed, which summarize the total number of risk alleles, as an alternative form of replication and for potential clinical utility for obesity risk prediction. Next, since there has been only limited research on when during development BMI-associated variants begin influencing BMI, a longitudinal twin study was utilized to assess the effects of adult-validated BMI-SNPs across adolescence into adulthood. In addition, obesity is comorbid with numerous medical conditions including cardiovascular disease, insulin-resistance and some forms of cancer, as well as, various psychiatric disorders including eating disorders, mood disorders, and substance use. The next series of studies aimed to understand phenotypic and genetic associations between BMI/obesity and binge eating disorder (BED), depression symptoms and smoking behavior. Using a clinical sample of overweight and obese women with and without BED, the relationship of BED, food intake and internalizing symptoms of depression and anxiety was examined. Next, twin study methodology was used to investigate if shared genetic and/or environmental liability was responsible for phenotypic associations found between BMI, depression symptoms, and impulsivity. Finally, a genetic association study aimed at investigating whether genetic variants were associated with multiple behaviors, body composition and smoking behavior, or were trait-specific is presented. By utilizing several samples and methodologies and by pursuing methods development, a comprehensive approach is presented that is hoped to represent a more powerful evidence-based strategy to understanding the genetic and environmental determinants of BMI and common complex obesity, along with associated depression symptoms and smoking behavior.

Introdução: Com a sofisticação das técnicas de análise do DNA, a medicina moderna tem à sua disposição boas possibilidades para elucidar quadros clínicos indefinidos em pacientes que possuem microrrearranjos cromossômicos complexos. O desenvolvimento da técnica de MLPA (Multiplex ligation-dependent probe amplification) aliado à tecnologia dos arrays (WGAS - whole genome array screening) possibilitou analisar de uma só vez, diferentes regiões de interesse clínico no genoma humano. Objetivo: O presente trabalho teve como objetivo estudar pacientes com atraso de desenvolvimento neuropsicomotor (ADNPM) associado à malformação congênita (MC) com cariótipo prévio normal ou inconclusivo. Material e métodos: Participaram do estudo 71 pacientes com ADNPM associado à MC que foram analisados utilizando o teste de MLPA com os kits P036 e P064, seguido de WGAS com as diferentes plataformas (Agilent, Affymetrix e Illumina). Resultados: Entre os 33 pacientes com alterações patogênicas e de significado clínico incerto (VOUS) encontramos: 12 pacientes com deleção, 5 com duplicação e 16 com duplicações e deleções (dup/del) concomitantes. Foram 29 pacientes com alterações patogênicas conclusivas, 4 pacientes com CNVs classificadas como VOUS e 15 pacientes tiveram resultado de array normal além dos outros 23 que apresentaram alterações benignas, ou por não apresentarem genes na região alterada, ou por serem genes sem fenótipos descritos, ou ainda, as alterações foram herdadas de genitores normais. Na casuística total foram encontrados 4 pacientes com regiões de perda de heterozigosidade. Conclusões: A utilização de uma estratégia combinada utilizando diferentes kits de MLPA, com capacidade para detectar as principais microalterações genômicas patogênicas conhecidas, associada à aplicação do WGAS possibilitou a detecção de alterações submicroscópicas, bem como a correlação clínica adequada para pacientes não diagnosticados pela citogenética clássica. Dessa forma, nosso estudo sugere um novo modelo para a aplicação combinada desses testes que representa uma alternativa de bom custo-benefício para a triagem genômica e definição diagnóstica dos pacientes com quadros sindrômicos complexos e suas famílias
Introduction: The recent technological advances on DNA-based techniques have established in modern medicine good opportunities to elucidate undefined clinical cases in patients with complex chromosomal microrearrangements. The performance of MLPA (Multiplex ligation-dependent probe amplification) technique together with array technologies (WGAS - whole genome array screening) created the possibility of one single experiment to analyze different regions of interest in the human genome. Objective: Patients with psychomotor delay (PSMD) associated with multiple congenital anomalies who had normal or inconclusive G-band-karyotype (MCA) were studied in order to understand the genotype-phenotype correlations. Material and methods: This study involved 71 patients with psychomotor delay (PSMD) associated with multiple congenital anomalies (MCA) analyzed by MLPA (P036 and P064 kits), followed by WGAS different platforms (Agilent, Affymetrix e Illumina®). Results: Among 33 patients with pathogenic and uncertain (VOUS) copy number variations (CNV) were found: 12 deletions, 5 duplications and 16 concomitant duplication and deletion (dup/del). There were 29 patients with conclusive pathogenic findings, 4 patients with VOUS and 16 patients with normal array, but others 23 patients with benign results, which means there is no gene content in the region involved, or because these genes were not linked to phenotype, or even due to CNVs inherited of healthy parents. From the whole casuistic, 4 individuals presented loss of heterozygosity (LOH) regions. Conclusions: The use of a combined strategy of analysis (MLPA - WGAS) with a high capacity to detect pathogenic CNVs allows unraveling microscopic imbalances, and consequently, offers an adequate clinical correlation for patients not previously diagnosed by classical cytogenetics. In conclusion, this study suggests a new model for the combined application of these techniques, which represents an optimal alternative for a genomic screening and diagnostic establishment in patients with rare complex disorders and their families

Canine osteosarcoma (OSA) is the most common bone tumour in dogs and is characterized by massive genomic instability throughout the cancer genome. Using matched primary tumour and metastasis samples from 8 OSA cases, we identified 2 focally deleted (<0.2 Mb) novel candidate genes, cyclin-dependent kinase inhibitor 2B (CDKN2B) and the membrane associated guanylate kinase, WW and PDZ domain containing 2 (MAGI2), which have not been previously identified in canine OSA and may have roles in driving OSA tumour progression. We also identified highly similar genomic profiles between matched samples, which yielded a small list of candidate regions that may harbour genes that drive metastasis formation. This study identified potential therapeutic targets, prognostic markers and early detectors of metastasis, in addition to providing insight into the OSA tumour progression.

碩士
國立陽明大學
腦科學研究所
103
Autism spectrum disorder (ASD) is an early-onset neurodevelopmental disorder with complex genetic underpinning in its etiology. Copy number variations (CNVs), such as deletions or duplications at the submicroscopic level, are one of the genetic factors that are significantly associated with ASD addressed in recent genome-wide association studies (GWAS). However, the significance of CNV has not been well investigated in non-Caucasian ASD population. To identify the pathogenic CNVs of ASD cohort in Taiwan, we performed a segment-based GWAS of CNV in 340 ASD cases and 1093 healthy controls using Affymetrix SNP array. We were particularly interested in a significantly associated gene, PARK2, which overlapped with a case-specific CNV region. The findings were validated in the initial screen sample set and replicated in another sample set by real-time qPCR. To further understand the pathology, we evaluated the gene expression levels of PARK2 in two probands and the carriers in their family. Our result revealed that CNVs at the PARK2 locus was at significantly higher prevalence in ASD cohorts than in controls (P=0.015). We identified six ASD patients carrying CNVs at 6q26 that spanned different exons of PARK2. The CNVs were categorized into three coverage regions: the A (2 dup, 1 del), B (2 del) and C (1 del) regions corresponded to exon 3-4, exon 5 and exon 6-7, respectively. We observed that the PARK2 expression level was down-regulated in the case with deletion at both the CNV-affected regions (EX3-4 and EX4-5) but not at both the CNV-unaffected regions (EX6-7 and EX9-11). Except for the region EX3-4 of PARK2, we also found the down-regulation of mRNA level at the other 3 regions (EX4-5, EX6-7 and EX9-11) in the case with duplication. The result revealed that the gene function might be disrupted by either exonic deletion or duplication. The case with duplication showed more severe autistic symptoms compared to the case with deletion. Hence, we suggest that duplication on PARK2 may have more impact than deletion on susceptibility of ASD. Remarkably, mitochondrial dysfunction has recently been considered as a part of ASD pathogenesis and mutations in PARK2 are known to cause mitochondrial dysfunction. It is suggested that CNVs affect PARK2 might contribute to the hypothesis of mitochondrial dysfunction in a portion of ASD cases. The future works should focus on characterizing the functional impact of these CNVs on PARK2 gene and investigate the connection between PARK2-involved mitochondrial dysfunction and ASD.

博士
國立陽明大學
生命科學暨基因體科學研究所
96
Background and Motivation In recent studies, copy number variations (CNV) in the human genome have been discovered, some of which have been known to directly affect the levels of gene expressions, accounting for some phenotypic difference. CNV might directly disrupt the gene coding sequence or alter promoter regulation resulting in gene expression differences. Large numbers of severe copy number (CN) abnormalities of chromosomal segments, called chromosomal abnormalities, are usually detected in cancer cells. Chromosomal abnormalities can result in oncogene amplifications, suppressor gene deletions and unregulated gene expressions. If these genes are out of normal control, they can positively or negatively regulate cell growth, proliferation, death, and metastasis. Single nucleotide polymorphism (SNP) microarray is used to genotype genome-wide SNP, and can be used to simultaneously detect CN at high resolution. Currently, there are a few programs available to predict the location of CNVs, however, these methods are subject to a high false-positive rate (FPR). Although FPR is successfully reduced through genomic smoothing, small amplifications and deletions will be lost. Since CNVs have a large impact on gene expressions and phenotypes, effective experiments and methods for detecting CN-altered regions and CNVs are critical and urgent in further genomic studies. Purpose (a) An novel algorithm for SNP microarray in CN analysis: The primary objective of this study is to propose a SNP CN estimation algorithm for Affymetrix 100K and 500K SNP microarrays. Estimation of SNP allelic CN and inference of reliable p-values can be applied in allelic imbalance studies and result in lower false-positive and false-negative rates. Hidden-Markov Model (HMM) method will be applied to detect CN-altered regions based on SNP CN estimation. (b) Appling DNA pooling to CNV studies: CNVs are widely distributed over the human genome, but their coverage remains poorly understood. Due to the large cost of microarray experiments, it is impractical for corresponding research to extensively analyze numerous individuals. DNA pooling has been successfully applied to SNP discovery and validation, so we plan to apply this method to detect common CNV loci. (c) CNV studies in Han Chinese population: We plan to utilize 270 HapMap multi-ethnic individuals to determine the parameters of CNIT algorithm, and apply the parameters to 300 unrelated healthy Han Chinese residing in Taiwan. This large scale CNV survey, especially in our population, is important and will have direct impact on further genomic studies. (d) The correlations between gene expressions and genomic variants: A large amount of expression data, comprising of 47,289 human transcripts for 210 individuals from HapMap, was used to compare expression profiles among three ethnic populations, namely African, Asian, and Caucasian. In addition, large numbers of SNPs and CNVs were used to examine their relative impact on gene expressions. Results (a) Copy number inferring tool (CNIT) algorithm: The CNIT algorithm was constructed based on the considerations of probe selection, mismatch probe, and the coefficient of preferential amplification/hybridization. Compared with other programs, CNIT algorithm had lower FPR and lower variation in CN estimation, and showed better sensitivity in the detection of small amplifications and deletions. (b) Application of DNA pooling in CNV discovery: Equivalent amounts of DNA from two groups of ten and thirty normal unrelated individuals were pooled together for genotyping. The pooled DNA was hybridized to the Affymetrix 100K SNP microarray in duplicate, and the intensity data was analyzed by CNIT with the HMM method. Twelve common CN-variable regions were identified, suggesting that DNA pooling can be applied to discover common CNVs. (c) CNV survey in Taiwanese population: The average size of the CNV regions was 322 kb (ranging from 1.48 kb to 5.68 Mb), and covered a total of 2.47% human genome. There were 449 genes and 5 microRNAs within these 230 CNV regions, and some genes were previously known to be associated with diseases. (d) The impact of SNPs and CNVs on genome-wide gene expressions: Based on linear regression analyses, 1,366 (19.01%) and 604 (8.40%) transcripts correlated with the SNPs and CNVs, respectively. A total of 3,515 transcripts showed significant differences in mean expression level between populations, and genomic variations could account for 1,321 of these differences (37.58%). Conclusion and discussion CNIT algorithm was more accurate and yielded more reliable CN estimates without using a smoothing procedure, and had high power to detect CN changes in either single- or multi-point analyses. CNIT performed well in CNV and cancer genomic studies, including those focusing on leukemia and colorectal cancers, and its basic algorithm should also be applicable to other SNP microarray. DNA pooling can provide the resolution necessary to quickly, easily and economically study CNVs in a population. To our knowledge, this is the first report that has applied DNA pooling to study CNVs. The CNV survey was the pilot study in Taiwan, and a total of 230 CNVs were identified. These CNVs were representative in Han Chinese populations, and should be taken into consideration when further genetic studies are conducted. In addition, based on the genome-wide correlation analyses, CNVs and SNPs were found to have a large impact on the expressions of nearby genes, and should be seriously considered in further genomic studies.

碩士
臺灣大學
分子醫學研究所
98
Copy number variation is a source of genetic diversity in humans. CNVs can cause Mendelian or sporadic traits, be associated with complex disease, also represent benign polymorphic variants between individuals. However, CNVs information of Asian populations remains unexplored. We identified CNV distribution of 1015 individuals in Taiwan populations by using CMDX Bac Array CA2500/ CA3000 and find 438 loci with CNV signals. The CNV frequencies of six regions are over 20%. Most CNVs (92.44%) size are shorter than 1 Mb. Average CNV number in each individual is 4.36. Among CNV regions, 54.79% regions overlap with segmental duplication, 64.49% regions cover gene, 92.47% regions are reported in previous study, only 7.53% regions are novel in this study. In contrast to previous studies, CNVs identified are much less than the platform with high resolution. But it lower disturbance in clinical pathologic diagnosis for this reason. So, Bac array used in this study is more appropriate for prenatal, clinical diagnosis. This study establish CNV map of Taiwan populations for further reference of genetic consulting.

碩士
國立中央大學
系統生物與生物資訊研究所
98
Whole genome copy number variation (CNV) has been noticed and the related studies grew in amount since the completion of Human Genome Project (HGP). Those experiments using mouse as a biological model present a complete microarray data and clear CNV diversities between different strains. Applying two different algorithms, Hidden Markov Model (HMM) & Pair-wised Gaussian Merging (PGM), to determine the mouse genome-wide CNV segment, we found that the results are significantly different on CNV length, CNV location, and the number of CNV segments. We thought the reason might be: The two underlying statistical theories are quite different, leading to the different decision-making patterns of finding CNV segments. The distribution of the length of CNV segment determined by PGM is wider than those determined by HMM. However, after filtering the shorter CNV segments, the total number of results generated by these two algorithms became almost the same. So we can do further study on the data generated by HMM & PGM, such as finding out the CNV segments that only appeared in one of their results and checking the gene symbols or gene annotations. Besides the comparison of the calculating speed and space requirement between these algorithms, we can even applying them on the analysis of mouse whole-genome CNV.

Les recherches menées pour comprendre les troubles du spectre autistique (TSA) et la schizophrénie (SZ) ont communément utilisé une approche dite descendante, partant du diagnostic clinique pour investiguer des phénotypes intermédiaires cérébraux ainsi que des variations génétiques associées. Des études transdiagnostiques récentes ont remis en question ces frontières nosologiques, et suggèrent des mécanismes étiologiques imbriqués. L’approche montante propose de composer des groupes de porteurs d’un même variant génétique afin d’investiguer leur contribution aux conditions neuropsychiatriques (NPs) associées. Les variations du nombre de copies (CNV, perte ou gain d’un fragment d’ADN) figurent parmi les facteurs biologiques les plus associés aux NPs, et sont dès lors des candidats particulièrement appropriés. Les CNVs induisant un risque pour des conditions similaires, nous posons l’hypothèse que des classes entières de CNVs convergent sur des dimensions d’altérations cérébrales qui contribuent aux NPs. L’imagerie fonctionnelle au repos (rs-fMRI) s’est révélée un outil prometteur en psychiatrie, mais presqu’aucune étude n’a été menée pour comprendre l’impact des CNVs sur la connectivité fonctionnelle cérébrale (FC). Nos objectifs étaient de: 1) Caractériser l’effet des CNVs sur la FC; 2) Rechercher la présence des motifs conférés par ces signatures biologiques dans des conditions idiopathiques; 3) Tester si la suppression de gènes intolérants à l’haploinsuffisance réorganise la FC de manière indépendante à leur localisation dans le génome. Nous avons agrégé des données de rs-fMRI chez: 502 porteurs de 8 CNVs associées aux NPs (CNVs-NP), de 4 CNVs sans association établie, ainsi que de porteurs de CNVs-NPs éparses; 756 sujets ayant un diagnostic de TSA, de SZ, ou de trouble déficitaire de l’attention/hyperactivité (TDAH), et 5377 contrôles. Les analyses du connectome entier ont montré un effet de dosage génique positif pour les CNVs 22q11.2 et 1q21.1, et négatif pour le 16p11.2. La taille de l’effet des CNVs sur la FC était corrélée au niveau de risque psychiatrique conféré par le CNV. En accord avec leurs effets sur la cognition, l’effet des délétions sur la FC était plus élevé que celui des duplications. Nous avons identifié des similarités entre les motifs cérébraux conférés par les CNVs-NP, et l’architecture fonctionnelle des individus avec NPs. Le niveau de similarité était associé à la sévérité du CNV, et était plus fort avec la SZ et les TSA qu’avec les TDAH. La comparaison des motifs conférés par les délétions les plus sévères (16p11.2, 22q11.2) à l’échelle fonctionnelle, et d’expression génique, nous a confirmé l’existence présumée de relation entre les mutations elles-mêmes. À l’aide d’une mesure d’intolérance aux mutations (pLI), nous avons pu inclure tous les porteurs de CNVs disponibles, et ainsi identifier un profil d’haploinsuffisance impliquant le thalamus, le cortex antérieur cingulaire, et le réseau somato-moteur, associé à une diminution de mesure d’intelligence générale. Enfin, une analyse d’exploration factorielle nous a permis de confirmer la contribution de ces régions cérébrales à 3 composantes latentes partagées entre les CNVs et les NPs. Nos résultats ouvrent de nouvelles perspectives dans la compréhension des mécanismes polygéniques à l’oeuvre dans les maladies mentales, ainsi que des effets pléiotropiques des CNVs.
Research on Autism Spectrum Disorder (ASD) and schizophrenia (SZ) has mainly adopted a ‘top-down’ approach, starting from psychiatric diagnosis, and moving to intermediate brain phenotypes and underlying genetic factors. Recent cross-disorder studies have raised questions about diagnostic boundaries and pleiotropic mechanisms. By contrast, the recruitment of groups based on the presence of a genetic risk factor allows for the investigation of molecular pathways related to a particular risk for neuropsychiatric conditions (NPs). Copy number variants (CNVs, loss or gain of a DNA segment), which confer high risk for NPs are natural candidates to conduct such bottom-up approaches. Because CNVs have a similar range of adverse effects on NPs, we hypothesized that entire classes of CNVs may converge upon shared connectivity dimensions contributing to mental illness. Resting-state functional MRI (rs-fMRI) studies have provided critical insight into the architecture of brain networks involved in NPs, but so far only a few studies have investigated networks modulated by CNVs. We aimed at 1) Delineating the effects of neuropsychiatric variants on functional connectivity (FC), 2) Investigating whether the alterations associated with CNVs are also found among idiopathic psychiatric populations, 3) Testing whether deletions reorganize FC along general dimensions, irrespective of their localization in the genome. We gathered rsfMRI data on 502 carriers of eight NP-CNVs (high-risk), four CNVs without prior association to NPs as well as carriers of eight scarcer NP-CNVs. We also analyzed 756 subjects with idiopathic ASD, SZ, and attention deficit hyperactivity disorder (ADHD), and 5,377 controls. Connectome-wide analyses showed a positive gene dosage effect for the 22q11.2 and 1q21.1 CNVs, and a negative association for the 16p11.2 CNV. The effect size of CNVs on relative FC (mean-connectivity adjusted) was correlated with the known level of NP-risk conferred by CNVs. Consistent with results on cognition, we also reported that deletions had a larger effect size on FC than duplications. We identified similarities between high-risk CNV profiles and the connectivity architecture of individuals with NPs. The level of similarity was associated with mutation severity and was strongest in SZ, followed by ASD, and ADHD. The similarity was driven by the thalamus, and the posterior cingulate cortex, previously identified as hubs in transdiagnostic psychiatric studies. These results raised questions about shared mechanisms across CNVs. By comparing deletions at the 16p11.2 and 22q11.2 loci, we identified similarities at the connectivity, and at the gene expression level. We extended this work by pooling all deletions available for analysis. We asked if connectivity alterations were associated with the severity of deletions scored using pLI, a measure of intolerance to haploinsufficiency. The haploinsufficiency profile involved the thalamus, anterior cingulate cortex, and somatomotor network and was correlated with lower general intelligence and higher autism severity scores in 3 unselected and disease cohorts. An exploratory factor analysis confirmed the contribution of these regions to three latent components shared across CNVs and NPs. Our results open new avenues for understanding polygenicity in psychiatric conditions, and the pleiotropic effect of CNVs on cognition and on risk for neuropsychiatric disorders.

博士
國防醫學院
生命科學研究所
102
Hypertension is an important public health problem in Taiwan and in the world. Although many large-scale genome-wide association studies (GWAS) have been performed, few study identified replicable and large-impact hypertension loci, not to mention the scanty Chinese studies. Young onset hypertension (YOH) is considered as a more promising target disorder to investigate than the late-onset one due to its stronger genetic component. To map YOH genetic variants, we carried out GWAS to search for hypertension susceptibility loci/genes with whole genome SNPs, CNVs and gene expression profiles. This investigation was consisted of 3 studies. The first part is a three-stage genome-wide association study using SNPs as genetic markers (SNPs-GWAS), combining 1st-stage multilocus GWAS (400 age- and sex-matched pairs), 2nd-stage gene expression analysis, and 3rd-stage multilocus confirmatory study (992 matched pairs). The second part is a two-stage genome-wide CNVs association study (CNVs-GWAS) consisting of an usual first stage GWAS (200 matched pairs) to find potential susceptibility CNV regions and a second stage confirmatory association study with an independent set of samples (199 matched pairs). Gene expression profiles were used to find differentially expressed genes among those implicated in both the first and second stages of the study. The third part is a two-stage genome-wide gene expression association study (GEs-GWAS), consisting of a regular GWAS in the 1st - stage (126 YOHs and 149 controls) to find the differentially expressed genes and a further confirmation with an independent set of samples (127 YOHs and 150 controls). We also integrated the gene expression data and SNP data to find eSNPs. In the SNPs-GWAS, a total of six SNP septets flanking the C1orf135, GSN, LARS, and ACTN4 remained significant in all three stages. Among them, the septet flanking ACTN4 and LARS was also associated with blood pressure/hypertension in two external replication studies: Hong Kong Hypertension Study (HKHS) and WTCCC hypertension study. In the CNVs-GWAS, 11 CNVs regions involving 14 genes were identified. In the GEs-GWAS, 9 genes were significantly associated with hypertension in both the first- and second-stage. Among these genes, ZRANB1, FAM110A, PREP, ANKRD9 and LAM2 were also differentially expressed in an existing database of hypertensive mouse model (GSE19817). Our study identified several previously unknown YOH loci/genes in Han Chinese. Identification of these genes enriches hypertension susceptibility gene list, thereby shedding light on the etiology of hypertension in Han Chinese.

Le Syndrome d’Impatiences Musculaires de l’Éveil (SIME) est une maladie neurologique caractérisée par un besoin urgent de bouger les jambes. C’est également l’une des causes les plus fréquentes d’insomnie. C’est une maladie très répandue, avec une prévalence de presque 15 % dans la population générale. Les maladies multifactorielles comme le SIME sont souvent le résultat de l’évolution d’une composante génétique et d’une composante environnementale. Dans le cadre du SIME, les études d’association génomique ont permis l’identification de 4 variants à effet modéré ou faible. Cependant, ces quatre variants n’expliquent qu’une faible partie de la composante génétique de la maladie, ce qui confirme que plusieurs nouveaux variants sont encore à identifier. Le rôle des déséquilibres génomiques (Copy Number Variations ou CNVs) dans le mécanisme génétique du SIME est à ce jour inconnu. Cependant, les CNVs se sont récemment positionnés comme une source d’intérêt majeur de variation génétique potentiellement responsable des phénotypes. En collaboration avec une équipe de Munich, nous avons réalisé deux études CNVs à échelle génomique (biopuces à SNP et hybridation génomique comparée (CGH)) sur des patients SIME d’ascendance germanique. À l’aide d’une étude cas-contrôle, nous avons pu identifier des régions avec une occurrence de CNVs différentes pour les patients SIME, comparés à différents groupes contrôles. L’une de ces régions est particulièrement intéressante, car elle est concordante à la fois avec des précédentes études familiales ainsi qu’avec les récentes études d’associations génomiques.
Restless Legs syndrome (RLS) is a neurological disorder characterized by the urge to move one’s limbs. It is also one of the most frequent causes of insomnia. The prevalence of RLS is estimated to be around 15% in the general population. Complexes disorders like RLS are often the result of the evolution of genetic and environmental components. For RLS, recent Genome Wide Association Study (GWAS) have identified four variants with mild to moderate effects. However, those four variants explain only a small part of the disease heritability and thus, we expect that many new variants are still to be found. The impact of Copy-Number Variation (CNV) in the genetic mechanism of RLS is still unknown. However, many studies have recently position the CNVs as a significant source of genetic variation potentially responsible of phenotypes. In collaboration with a team from Munich, we conducted two genome-wide CNVs studies (Genome Wide SNP chips and Comparative Genomic Hybridization (CGH)) on RLS patients from Germany. Using cases-controls studies, we identified regions with a different occurrence of CNVs for RLS patients, compared to different groups of controls. One of these regions is particularly interesting, as it has already been identified by both linkage and association studies.