| | SLO | ENG | Cookies and privacy

Bigger font | Smaller font

Show document

Title:PRIMERJAVA RAZLIČNIH PRISTOPOV ZA IZRAČUN TVEGANJA ZA KRONIČNO VNETNO ČREVESNO BOLEZEN NA OSNOVI ANALIZE PROFILOV GENETSKIH POLIMORFIZMOV
Authors:Gabor, Kristjan (Author)
Potočnik, Uroš (Mentor) More about this mentor... New window
Štiglic, Gregor (Co-mentor)
Files:.pdf MAG_Gabor_Kristjan_2013.pdf (3,15 MB)
 
Language:Slovenian
Work type:Master's thesis/paper (mb22)
Typology:2.09 - Master's Thesis
Organization:FZV - Faculty of Health Sciences
Abstract:Kronična vnetna črevesna bolezen (KVČB) je bolezen prebavil. Delimo jo na dva glavna podtipa, ulcerozni kolitis (UK) in Crohnovo bolezen (CB). Za KVČB je značilen kroničen, običajno vseživljenjski potek z aktualnimi zagoni ter vmesnimi krajšimi ali daljšimi obdobji odsotnosti kliničnih znakov. Vzrok za nastanek bolezni še ni poznan. Predvidevajo, da je KVČB posledica pretiranega imunskega odziva črevesne sluznice na dejavnike okolja. Pomembno vlogo pri njenem razvoju imajo genetski dejavniki. Je kompleksna bolezen, saj jo povzročajo genske okvare na več mestih, vendar ima vsak gen zelo majhen doprinos k razvoju bolezni. Namen magistrskega dela je izračunati statistične povezave in primerjati različne pristope napovednih modelov na osnovi analize genetskega zapisa, kot so napovedni model s kombinacijo polimorfizmov SNP (angl. single nucleotide polymorphism), napoved tveganja po multiplikativnem modelu in napovedni model, zgrajen s pomočjo strojnega učenja. Napovedni modeli so zgrajeni na modelu bolezni KVČB za posamezen SNP na osnovi rezultatov genotipizacije 33 različnih polimorfizmov SNP oziroma 30 genov pri slovenskih bolnikih s KVČB. Ti napovedni modeli nam omogočajo, da na osnovi genske analize napovemo tveganje bolezni za posameznika. Ugotovili smo, kateri SNP-ji predstavljajo najpomembnejši dejavnik tveganja v slovenski populaciji. Z metodo primerjave kombinacije SNP-jev med dvema skupinama smo najboljšo napoved dosegli pri primerjanju bolnikov z UK s kontrolno skupino zdravih posameznikov. Občutljivost testa je bila 64,9-odstotna in specifičnost 72,13-odstotna. Z napovednim modelom, zgrajenim s pomočjo strojnega učenja v odprtokodnem statističnem programu R, smo pri razlikovanju med skupino bolnikov s CB in kontrolno skupino zdravih posameznikov dosegli 76,38-odstotno občutljivost in 54,55-odstotno specifičnost. Prav tako smo poiskali najbolj informativne kombinacije genov in genetskih profilov, ki z optimalnim razmerjem med občutljivostjo in specifičnostjo napovedujejo tveganje za KVČB. Tako smo s pomočjo strojnega učenja ugotovili, da so za nastanek bolezni najbolj vplivni geni SLC22A5, FCGR3A, NR1/2 in SLC22A4. Ob primerjavi relativnega tveganja (RR) po standardni enačbi in po sistemu podjetja deCODEme, ki ima nekoliko drugačen pristop izračuna RR, smo dobili 12-odstotno neujemanje rezultatov oziroma napovedi relativnega tveganja za določene posamezne genotipske kombinacije.
Keywords:kronična vnetna črevesna bolezen, genetika, DNA, polimorfizmi SNP, napovedni modeli, strojno učenje, bioinformatika
Year of publishing:2013
Publisher:[K. Gabor]
Source:Maribor
UDC:616.34-002:575(043.2)
COBISS_ID:1921956 Link is opened in a new window
NUK URN:URN:SI:UM:DK:HBZ7MJYE
Views:1962
Downloads:154
Metadata:XML RDF-CHPDL DC-XML DC-RDF
Categories:FZV
:
  
Average score:(0 votes)
Your score:Voting is allowed only for logged in users.
Share:AddThis
AddThis uses cookies that require your consent. Edit consent...

Hover the mouse pointer over a document title to show the abstract or click on the title to get all document metadata.

Secondary language

Language:English
Title:A comparison of different approaches for calculating a risk for inflammatory bowel disease based on genetic polymorphism profile analysis
Abstract:Inflammatory bowel disease (IBD) is a chronic disease of gastrointestinal tract. It is divided into 2 subtypes, Ulcerative colitis (UC) and Crohn's disease (CD). IBD is characterized as chronical, usually lifelong course with acute recurring flare-ups and intermediate shorter or longer periods without clinical signs (remission). Cause of the disease is still unknown. Scientists anticipate that IBD is a result of excessive immune response of intestinal mucosa to environmental factors. Important role in it's development also has genetic factors. The disease is complex, caused by genetic defects on several places but every gen contributes only a little to development of the IBD. Purpose of this masters is to calculate statistical relationships and compare different approaches of predictive models based on genetic code analysis, such as predictive model in combination with SNP (single nucleotide polymorphism), prediction of risk following multiplicative model and predicive model, made by machine learning. Predictive models are build based on model of IBD for single SNP, resulting genotyping of 33 various SNPs or in slovenian patients with IBD of 30 various SNPs. This predictive models allow us to predict the disease risk for individuals based on genetic analysis. We have discovered, which SNPs present the most important risk factor in slovenian population. Using SNPs combination method between two groups, we have achieved the best prediction by comparing patients with UC to control group of healthy individuals. Sensitivity of the test was 64,9 percent and specificity 72,13 percent. Using predictive model, made by machine learning in opencoded statistical programme R, we have attained 76,38 percent sensitivity and 54,55 percent specificity in comparing group of patients with CD to control group of healthy individuals. Moreover, we have found the most informative combinations of genes and genetic profiles, that in optimal relation between sensitivity and specificity predict risk for IBD. The most influental genes on developing the disease are SLC22a5, FCGR3A, NR1/2 and SLC22A4, according to machine learning. Comparing relative risk (RR) using standard equation to relative risk when using company deCODEme system, we have obtained 12 percent of mismatch results or predictions of relative risk for individual genotypic combinations.
Keywords:Inflammatory bowel disease, genetics, DNA, SNPs, predictive models, machine learning, bioinformatics


Comments

Leave comment

You have to log in to leave a comment.

Comments (0)
0 - 0 / 0
 
There are no comments!

Back
Logos of partners University of Maribor University of Ljubljana University of Primorska University of Nova Gorica