Open AccessMarker Imputation in Barley Association Studies
Author(s) -
Jannink JeanLuc,
Iwata Hiroyoshi,
Bhat Prasanna R.,
Chao Shiaoman,
Wenzl Peter,
Muehlbauer Gary J.
Publication year - 2009
Publication title -
the plant genome
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.403
H-Index - 41
ISSN - 1940-3372
DOI - 10.3835/plantgenome2008.09.0006
Subject(s) - imputation (statistics) , biology , single nucleotide polymorphism , snp , genetic marker , genotyping , missing data , computational biology , genetics , association mapping , computer science , genotype , gene , machine learning
Association mapping requires high marker density, potentially leading to many missing marker data and to high genotyping costs. In human genetics, methods exist to impute missing marker data and whole markers typed in a reference panel but not in the experimental dataset. We sought to determine if an imputation method developed for human data would function effectively in a barley ( Hordeum vulgare L.) panel. The panel contained 98 lines, 2517 single nucleotide polymorphism (SNP) markers, and 716 Diversity Arrays Technology (DArT) markers. Averaged over markers, masked scores were correctly imputed 97.1% of the time. We chose 610 and 273 tag markers in two‐ and six‐row barley subpopulations, respectively. Despite this low number of tags, imputation accuracy was such that for about 80% of non‐tag markers, the prediction r 2 between imputed and true scores was 0.8 or higher. When DArT markers were used as tags, SNP markers were imputed with similar accuracy, suggesting that the method can convert association information from one marker system (e.g., DArT) to another marker system (e.g., SNP). We believe marker imputation methods will have an important future in association studies as a component of tagging methods and in reducing problems due to missing data.