|
|
|
|
Genotype imputation using support vector machine in parent-offspring
trios
|
Abbas Mikhchi1, Mahmood Honarvar2,
Nasser Emam Jomeh Kashan1*, Saeed Zerehdaran3 and
Mehdi Aminafshar1 |
1Department of Animal Science, Science and Research
Branch, Islamic Azad University, Tehran, Iran; 2Department of
Animal Science, Shahr-e-Qods Branch, Islamic Azad University, Tehran,
Iran; 3Department of Animal Science, Ferdowsi University of
Mashhad, Mashhad, Iran
|
|
Abstract |
An important problem in
genomic selection in livestock is the cost of genotyping. Genotype
imputation is a process of predicting unknown genotypes or un-typed
Single nucleotide polymorphism (SNP), which uses reference population to
predict missing genotypes for animal genetic variations. Support vector
machines are algorithms based machine learning methods. We compared the
Support Vector Machines (SVMs) and Beagle software for Genotype
imputation in parent-offspring trios in term of imputation accuracy and
computation of time. The methods employed uses simulated data (1000
trios with 10k SNPs) to impute the missing SNPs in parent-offspring
trios. The genome consists of 5 chromosomes and each chromosome was set
as 100 CM length. For simulated dataset five versions: NA10, NA30, NA50,
NA70 and NA 90, were created (10, 30, 50, 70 and 90 percent of offspring
genotypes are missing). Our results show that in all versions of
simulated dataset Beagle outperformed SVM in term of imputation accuracy
and computation of time. The Beagle requires almost no tuning and can
easily handle missing predictor genotypes. We conclude to use of SVM in
larger Sample size (i.e 10000) for imputation of parent-offspring
trios.
|
Keywords:
Genotype imputation; trios; support vector machine; machine learning
methods |
|
To cite this article:
Mikhchi
A, M Honarvar, NEJ Kashan, S Zerehdaran and M Aminafshar, 2015.
Genotype imputation using support vector machine in
parent-offspring trios.
Res.
Opin.
Anim. Vet. Sci., 5(10): 416-419. |
|
|