Advances in microarray-based copy number variation discovery and phenotypic associations
Date
2024-05-10
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Bioloogiatunnist teame, et igal inimesel on kogu DNAst kaks koopiat – üks emalt, teine isalt. Uurides genoomi lähemalt, võib aga mitmete DNA piirkondade puhul näha sellest reeglist kõrvalekaldeid. Neid põhjustavad muutused inimese genoomis, mille tulemusena on mingi DNA piirkond kadunud või mitmekordistunud, ning mida nimetatakse DNA koopiaarvu variatsioonideks (ingl. k. copy number variation ehk CNV). Algselt seostati CNVsid eelkõige raskekujuliste arenguhäiretega, kuid nüüdseks on teada, et CNV piirkondi leidub kõigi inimeste DNAs. Sellest hoolimata on CNVde mõju inimese terviseriskidele hinnatud märkimisväärseks. Et nende efekti paremini kaardistada, oleme oma töös keskendunud CNVde analüüsimetoodika arendamisele ning CNVde ja haiguste ning muude meditsiiniliselt oluliste tunnuste vaheliste seoste tuvastamisele.
CNV piirkondade määramine on arvutuslikult keerukas ning valepositiivsete leidude osakaal on suur. Et hinnata iga leitud CNV piirkonna usaldusväärsust, lõime kvaliteedimudeli, mille kasutamine suurendas oluliselt CNVde ning huvipakkuvate tunnuste vaheliste seoste tuvastamise võimekust. Ülejäänud töös keskendusimegi selliste seoste leidmisele. Koostöös erinevate andmekogude ja biopankadega tuvastasime kokku mitusada seost CNVde ning mitmete haiguste ja muude tunnuste vahel. Nägime, et tihti on üks CNV seotud mitmete väga erinevate haiguste või tunnustega. Ühte sellist, inimese 16. kromosoomis paiknevat CNVd, mida oli varasemalt seostatud autismi, skisofreenia, pea ümbermõõdu, ja kehamassiindeksiga, seostasime lausa 38 erineva tunnusega, sealhulgas puberteedi alguse vanusega. Piirkonda lähemalt uurides leidsime sealt kaks geeni, millele ennustasime põhjuslikku mõju puberteedieale. Kokkuvõttes parendab meie töö CNVde määramise metoodikat ning kaardistab uusi haiguste ja muude tunnustega seotud CNVsid ning nendes paiknevaid põhjuslikke geene. Nende analüüsidega astume sammu lähemale huvipakkuvate tunnuste geneetilise tagapõhja täielikule määramisele.
From the biology class we know that every person has one full copy of each DNA molecule from either parent. On closer inspection, however, we observe that several genomic regions deviate from this rule. These regions have undergone additional loss or gain of genomic content, which are called DNA copy number variations (CNVs). They were first discovered in the context of developmental delay and neuropsychiatric disorders, where they exhibited severe pathogenic effects. The current knowledge indicates, however, that CNVs are abundant in the genomes of all individuals. Still, the impact of CNVs on health is considered substantial. In our work we aim to map the effect of CNVs on human traits, by improving CNV analysis methodology and associating CNVs with various diseases and medically relevant phenotypes. CNV detection is challenging and often results in a large proportion of false positive calls. To differentiate true CNVs from false, we have developed a statistical model to estimate CNV quality. We show that our model improves the ability to detect associations between CNVs and traits of interest. The discovery and explorations of such associations comprise the second part of our work. In collaboration with various biobanks and clinical cohorts, we have detected several hundreds of CNV-phenotype associations. We observed that one CNV region can often be linked to multiple seemingly unrelated diseases and traits. For example, we linked a large CNV region in chromosome 16, previously associated with autism, schizophrenia, head circumference and body mass index, to 38 different disease risks and continuous traits. Among them was an association with pubertal timing, which we studied in depth. We found two genes from within the CNV region that we could causally link to pubertal timing. In summary, our work improves CNV detection methodology and maps several CNV regions to phenotypes. With these analyses we take a step closer to fully dissecting the genetics underlying human health.
From the biology class we know that every person has one full copy of each DNA molecule from either parent. On closer inspection, however, we observe that several genomic regions deviate from this rule. These regions have undergone additional loss or gain of genomic content, which are called DNA copy number variations (CNVs). They were first discovered in the context of developmental delay and neuropsychiatric disorders, where they exhibited severe pathogenic effects. The current knowledge indicates, however, that CNVs are abundant in the genomes of all individuals. Still, the impact of CNVs on health is considered substantial. In our work we aim to map the effect of CNVs on human traits, by improving CNV analysis methodology and associating CNVs with various diseases and medically relevant phenotypes. CNV detection is challenging and often results in a large proportion of false positive calls. To differentiate true CNVs from false, we have developed a statistical model to estimate CNV quality. We show that our model improves the ability to detect associations between CNVs and traits of interest. The discovery and explorations of such associations comprise the second part of our work. In collaboration with various biobanks and clinical cohorts, we have detected several hundreds of CNV-phenotype associations. We observed that one CNV region can often be linked to multiple seemingly unrelated diseases and traits. For example, we linked a large CNV region in chromosome 16, previously associated with autism, schizophrenia, head circumference and body mass index, to 38 different disease risks and continuous traits. Among them was an association with pubertal timing, which we studied in depth. We found two genes from within the CNV region that we could causally link to pubertal timing. In summary, our work improves CNV detection methodology and maps several CNV regions to phenotypes. With these analyses we take a step closer to fully dissecting the genetics underlying human health.
Description
Väitekirja elektrooniline versioon ei sisalda publikatsioone