|Yu, Jianming - CORNELL UNIVERSITY|
|Pressoir, Gael - CORNELL UNIVERSITY|
|Briggs, William - UNIVERSITY OF WISCONSIN|
|Vroh, Bi - CORNELL UNIVERSITY|
|Yamasaki, Masanori - UNIVERSITY OF MISSOURI|
|Doebley, John - UNIVERISTY OF WISCONSIN|
|Gaut, Brandon - UNIVERSITY OF CALIFORNIA|
|Kresovich, Stephen - CORNELL UNIVERSITY|
Submitted to: Nature Genetics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: October 13, 2005
Publication Date: February 15, 2006
Citation: Yu, J., Pressoir, G., Briggs, W., Vroh, B., Yamasaki, M., Doebley, J.F., Mcmullen, M.D., Gaut, B.S., Holland, J.B., Kresovich, S., Buckler Iv, E.S. 2006. A unified mixed-model method for association mapping accounting for multiple levels of relatedness. Nature Genetics. 38:203-208. Interpretive Summary: Association mapping offers scientists a powerful, high-resolution approach to identifying and describing the genes and alleles that control complex traits. However, association tests to date have routinely been plagued by population structure that can produce spurious associations. To reduce these false signals, three major approaches to association mapping that control for population structure have been developed over the last few years. Here, we present a unified, mixed-model method that successfully combines the strengths of all three approaches, and statistically outperforms each individually. In addition to its remarkably good error rates and statistical power, our combined approach is based on molecular marker data rather than extensive pedigree knowledge, making it extremely flexible. As such, our new method should be readily applicable to human, crop, tree, and other animal systems.
Technical Abstract: As population structure can result in spurious associations, it has constrained the use of association studies in human and plant genetics. Association mapping, however, holds great promise if true signals of functional association can be separated from the vast number of false signals generated by population structure. Genomic control and structured association are two methods commonly used to achieve this separation, but recent study suggests that even modest levels of subpopulation structure can influence the dissection of complex traits. We have developed a novel mixed-model approach to simultaneously account for multiple levels of relatedness detected by random genetic markers. Based on data from a maize association mapping project, this approach has excellent Type I and Type II error rates. This method should be readily applicable to a wide range of species and populations, as it estimates population structure based on increasingly available molecular marker data.