摘要
Consensus quantitative trait loci (QTL) in meta-analysis of multiple independent QTL mapping experiments provides a strong foundation for marker-assisted selection and gene cloning.However,meta-analysis suffers from the lack of available genomic information and the results vary when different reference linkage maps are used.Here,to overcome these limitations,we propose a linkage-group-based QTL synthesis analysis approach that we have named linkage graph analysis.First,a graph model is constructed from derived linkage groups.Next,an unsupervised classification approach is used to obtain marker intervals with co-segregating patterns among multiple genomes.Finally,a frequent itemset mining technique is used to identify the markers (or intervals) closely linked to the QTL.The proposed method was validated by one Monte Carlo simulation study and by real data analysis of cotton genomes.Two major advantages of the new method are:(i) A reference linkage group is not required;(ii) the effect of the initial QTL is reduced because false QTLs can be detected and excluded from the dataset.The ability to reliably identify the markers associated with a true QTL is valuable in crop breeding.
Consensus quantitative trait loci （QTL） in meta-analysis of multiple independent QTL mapping experiments provides a strong foundation for marker-assisted selection and gene cloning. However, meta-analysis suffers from the lack of available genomic information and the results vary when different reference linkage maps are used. Here, to overcome these limitations, we propose a linkage-group-based QTL synthesis analysis approach that we have named linkage graph analysis. First, a graph model is constructed from derived linkage groups. Next, an unsupervised classification approach is used to obtain marker intervals with co-segregating patterns among multiple genomes. Finally, a frequent itemset mining technique is used to identify the markers （or intervals） closely linked to the QTL. The proposed method was validated by one Monte Carlo simulation study and by real data analysis of cotton genomes. Two major advantages of the new method are： （i） A reference linkage group is not required; （ii） the effect of the initial QTL is reduced because false QTLs can be detected and excluded from the dataset. The ability to reliably identify the markers associated with a true QTL is valuable in crop breeding.
基金
supported by the National Key Basic Research Program of China(2011CB109300)
the National Natural Science Foundation of China(30971848)
the Fundamental Research Funds for the Central Universities(KYT201002)
Program for New Century Excellent Talents in University(NCET-05-0489)
the Program of Introducing Talents of Discipline to Universities(B08025)
