Procjena genetskih parametara binomne varijable koristeći GLIMMIX SAS proceduru

master's thesis

• Author: Sara Matanović
• Mentor: Miroslav Kapš (mentor)
• Committee member: Miroslav Kapš (predsjednik povjerenstva)
• Committee member: Ino Čurik (član povjerenstva)
• Committee member: Vlatka Čubrić Čurik (član povjerenstva)
• Committee member: Maja Ferenčaković (član povjerenstva)
• Granter: University of Zagreb; Faculty of Agriculture; (Animal Science); Zagreb
• Defense date and country: 2016-09-27, Croatia
• Scientific / art field, discipline and subdiscipline: BIOTECHNICAL SCIENCES; Agronomy; Genetics and Breeding of Plants, Animals and Microorganisms

Abstract

Procijenjene su izvedbe različitih modela za analizu genetskih parametara binomne varijable.
Glavni cilj je bio metodološki prikazati neku binomnu varijablu, usporediti probit i logit
modele međusobno te s običnim linearnim modelom. Podatci su simulirani koristeći SAS 9.3
program. Simulirana su dva seta podataka, oba s 1000 bikova te 100 kćeri po svakome biku, a
razlikovala su se u proporciji binarne varijable koja je iznosila 0,2 za jedan set podatka te 0,5
za drugi set. Podatci su zatim obrađeni procedurama mixed za obični linearni model (model
očeva) i glimmix za uopćeni linearni mješoviti model. Na temelju dobivenih vrijednosti
izračunati su heritabiliteti te su uspoređene vrijednosti varijance očeva i varijance ostatka za
svaki od navedenih modela. Uočene su razlike između običnog linearnog te probit i logit
modela. Također su uočene razlike između probit i logit modela, ali nisu bile značajno velike.

Abstract (english)

The performances of different models for genetic analysis of binomial variable have been
evaluated. The main goal was to methodologically show some binomial variable, compare the
probit and logit models mutually and with a linear model. The data were simulated using SAS
9.3 software. Two sets of data were simulated, both having 1000 bulls with a 100 daughters
for each bull. The data sets differed in the proportion of the binary variable that was 0,2 for
one data set and 0,5 for the other. The data were processed using proc mixed for linear model
(sire model) and proc glimmix for generalized linear mixed model. Based on the obtained
values heritability was calculated for each model. Also, the values of sire and error variances
were compared. Differences between linear model and probit and logit models were noticed.
Differences between probit and logit model were also noticed but were not significantly large.

• Keywords: binomna varijabla; obični linearni model; probit model; logit model; heritabilitet
• Keywords (english): binomial variable; linear model; probit model; logit model; heritability
• Language: croatian
• URN:NBN: urn:nbn:hr:204:096492
• Study programme: Title: Animal Genetics and Breeding; Study programme type: university; Study level: graduate; Academic / professional title: magistar/ magistra inženjer/inženjerka genetike i oplemenjivanja životinja (magistar/ magistra inženjer/inženjerka genetike i oplemenjivanja životinja)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2016-11-29 12:40:16