20.500.12556/DKUM-82708 Primerjava števila in kvalitete imputacije genotipov polimorfizmov posameznega nukleotida pridobljenih s HRC in TOPMed imputacijskim serverjem Comparison of the number and quality of single nucleotide polymorphisms genotypes imputation acquired with HRC and TOPMed imputation servers Uvod: Imputacija genotipskih podatkov je postala že vsakdanje orodje za uporabo na področju človeške genetike, ki predvideva in vstavlja manjkajoče genetske podatke med posameznimi biooznačevalci glede na različice, ki so že določene na referenčnem panelu predhodno sekvenciranih posameznikov. Učinkovitost trenutnih metod imputacije je privedla do pojava brezplačnih spletnih storitev imputacije genotipa, ki bistveno povečajo statistično moč ter so bistvenega pomena pri metaanalizah gensko povezanih študij. Metode: Za pripravo surovih genotipskih podatkov na imputacijo smo uporabili program PLINK, programski jezik R in operacijski sistem Linux. Datoteke smo naložili na Michigan (HRC za ang. Haplotype Reference Consortium) in TOPMed imputacijska strežnika, jih po imputaciji ustrezno združili ter prešteli filtrirane (Rsq > 0,3) in nefiltrirane (Rsq > 0,0) različice. Rezultate smo statistično primerjali in dokazali, kateri imputacijski strežnik je učinkovitejši. Rezultati: Na podlagi Wilcoxonovega testa predznačenih rangov pri številu nefiltriranih različic in statistične primerjave filtriranih različic (Rsq > 0,3) s pomočjo parnega T-testa smo na podlagi obeh testov ugotovili, da lahko s strežnikom TopMED panel statistično značilno imputiramo večje število manjkajočih polimorfizmov kot s strežnikom HRC. Razprava in sklep: Z uporabo dveh imputacijskih strežnikov smo dokazali, da pride do bistvenih razlik pri številu imputiranih različic in je za kvaliteto študije pomembno, katerega se raziskovalec odloči uporabiti v praksi. Introduction: The imputation of genotype sequence data has become an everyday tool for use in human genetics, anticipating and inserting missing genetic data between individual biomarkers according to variants already identified in the pre-sequenced reference panels of individuals. The effectiveness of current imputation methods has led to the emergence of free online genotype imputation services, which significantly increase statistical power and are essential in meta-analyzes of genetically related studies. Methods: Software tool PLINK, R programming language and Linux operating system were used to prepare raw genotype data for imputation. Files were uploaded to Michigan and TOPMed imputation servers, merged after imputation, and filtered (Rsq> 0.3) and unfiltered (Rsq> 0.0) versions were counted. We statistically compared the results and proved which imputation server is more suitable. Results: Based on the Wilcoxon test of predicted ranks in the number of unfiltered versions and statistical comparison of filtered versions (Rsq> 0.3) using a pair T-test, we found based on both tests that the TopMED panel imputes statistically significantly more compared to HRC server. Discussion and conclusion: Using two imputation servers, we have proven that there are significant differences in the number of imputed versions and it is important for the quality of the study which researcher decides to use in practice. genotip imputacija TOPMed HRC imputacijski strežnik genotype imputation TOPMed HRC imputation servers true true false [A. Softić] Slovenski jezik Angleški jezik Magistrsko delo/naloga 2022-08-30 14:50:36 2022-12-19 14:07:59 2023-01-23 15:36:51 0000-00-00 00:00:00 2022 0 Maribor 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 575.112(043.2) 131545603 MAG_Softic_Almir_2022.pdf MAG_Softic_Almir_2022.pdf 1 9117F1E031D94A924154DE5CAAB48F05 0410f35490f1c3f4981a44d791dfe4c0a3b4e699ed659c5b3dac598c80f9908d 24dad1af-2865-11ed-9649-00155d000105 https://dk.um.si/Dokument.php?lang=slv&id=161224 Fakulteta za zdravstvene vede 0 0 0