Lífupplýsingafræðileg og sameindalíffræðileg greining á eiginleikum DNA-metýlunar í erfðamengi mannsins

Epigenetics is the study of DNA related information heritable through meiosis and mitosis that does not include the DNA code itself. DNA methylation is currently the most studied epigenetic mark. It is a part of the inactivation of the X chromosome, defense against transposable elements and the control of tissue-specific gene expression and the expression of imprinted genes. Changes in DNA methylation are thought to be involved in the pathogenesis of many common diseases, including cancer. The aim of the Ph.D. project was to apply bioinformatic and biological methods to further the understanding of the properties of DNA methylation in the human genome. To assist interpreting results from global methylation assays, a bioinformatics analysis of the properties of methylation-sensitive restriction endonucleases suitable for such measurements was performed. Intra-individual changes in DNA methylation over time were demonstrated in longitudinal samples from two populations from Iceland and USA. Global methylation changed by more than 10% for 29% of the individuals in the Icelandic population (P<0.001). The change was bi-directional; DNA methylation decreased for a part of the population but increased for another part of the population. The global methylation similarly changed in the USA population, and there was a familial clustering of conservation of methylation (h2=0.99, P<0.001). Since sampling the human germline is difficult, a genome-wide bioinformatic surrogate marker for germline methylation utilizing methylation-associated single base pair polymorphism (mSNP) was developed. It was used to demonstrate a positive correlation between germline methylation and homologous recombination (r=0.622, P<10-15) that remained significant after correcting for confounding variables (r=0.172, P<10-15). The marker was then used to explore the relationship between germline methylation and subfamilies of transposable elements in the human genome. After correcting for confounding variables, a negative correlation was found between the mSNP marker and Alu subfamily (r=-0.14, -0.16, -0.16, -0.20 for 125, 250, 500 and 1000 kb genome windows) The correlation pattern between the mSNP marker and the L1 subfamily varied with window size (r=-0.01, -0.01, -0.01, -0.17 for 125, 250, 500 and 1000 kb genome windows). Finally methods of systems biology were used to study the metabolic effects of different expression level of imprinted genes. The greatest perturbation in the metabolic reconstruction occurred when differential expression of the ATP10A gene was simulated. The simulated effects of differential expression on metabolism did not support Haig's parental intergenome conflict theory, since only 50% of the genes followed its predictions (P=1.0). The results of the thesis support one prerequisite of an epigenetic model of common disease pathogenesis, i.e. that epigenetic marks change with time (Paper II). The results of the project suggest that DNA methylation of the germline is associated with homologous recombination (Paper III). The results indicate that DNA methylation is unlikely a part of a defense system against Alu elements, although it might participate in a defense system against L1 elements (Paper IV). The systems biology analysis of the metabolic effect of imprinted genes are not supportive of Haig's theorem (Paper V). Methods developed in this project can be used to interpret global DNA methylation analysis measurements with restriction endonucleases, to map the DNA methylation of the human germline and test dosage sensitivity of human metabolic genes.Utangenaerfðir (e. epigenetics) fjalla um upplýsingar tengdar erfðaefninu sem erfast við frumuskiptingu án þess að vera hluti af DNA röðinni sjálfri. Metýlun á DNA er mest rannsakaða utangenamerkið. Hún kemur við sögu í óvirkjun X litningsins, vörnum gegn stökklum, stýringu á vefjasérhæfðri tjáningu gena og tjáningu genagreyptra gena. Breytt DNA-metýlun er talin hluti af meingerð margra algengra sjúkdóma, þar á meðal krabbameins. Markmið doktorsverkefnisis var að beita lífupplýsinga- og líffræðilegum aðferðum til að auka skilning á dreifingu og hlutverki DNA-metýlunar í erfðamengi mannsins. Til að auðvelda túlkun á mælingum á heildarmetýlun erfðaefnis var gerð lífupplýsingafræðileg greining á eiginleikum skerðiensíma sem unnt er að nota til slíkra mælinga. Sýnt var fram á að heildarmetýlun erfðaefnis og metýlun í stýriröðum ýmissa gena breyttist með tíma í tveimur mismunandi langsniðsþýðum frá Íslandi og Bandaríkjunum. Heildarmetýlun breyttist um meira en 10% hjá 29% einstaklinga í íslenska þýðinu (P<0.001). Breytingin var í báðar áttir, DNA-metýlun jókst hjá hluta en minnkaði hjá hluta einstaklinganna milli mælipunktanna tveggja. Í þýðinu frá Bandaríkjunum varð sambærileg breyting á heildarmetýlun, og reyndist varðveisla metýlmynstursins fjölskyldulægur eiginleiki (h2=0.99, P<0.001). Þar sem erfitt er að afla sýna úr kímlínu mannsins, var þróað lífupplýsingafræðilegt merki fyrir metýlun kímlínunnar sem byggir á kortlagningu metýltengdra eins basapara erfðabreytileika (mSNP). Merkið var nýtt til að sýna fram á jákvæða fylgni milli DNA-metýlunar kímlínunnar og endurröðunar erfðaefnis (r=0.622, P<10-15) sem hélst þó leiðrétt væri fyrir bjagandi breytum (r=0.172, P