中国农业科学院 作物科学研究所
Institute of Crop Science, Chinese Academy of Agricultural Sciences, Beijing 100081, China

 

 

Marker density issues in QTL mapping using the maize genome

 

1. Introduction

There is no argument that the large amount of markers is becoming available. QTL mapping in biparental populations depends on the crossover between linked markers, and linked marker and QTL. Most application studies on QTL mapping are still using populations with sizes of hundreds, and the number of markers ranges from 100 to 300. In such populations, it's hard to say that 1 cM –distributed markers can more precisely detect QTL compared with 10 cM -distributed markers. Say, if there are 10 closely linked markers in a chromosomal region with a marker interval of 1 cM or less, these markers are very likely identical in a limited-size mapping population. So nine of them won't give us any added values in QTL mapping.

To better answer this question, we simulated two marker densities ( 2 cM vs 10 cM ) and two population sizes. The gain of using 2 cM distributed marker is marginal comparing with 10 cM distributed markers. On the other side, the increase in population size can greatly increase the detection power and reduce the false discovery rate (FDR). When the population size is high, all QTL may be detected with high powers, once again leaving little space for using more densely distributed markers. However, we are not denying the usefulness of dense markers in QTL fine mapping in some secondary populations.

 

2. Maize chromosomes and length (cM)

-Ch1                 301.7

-Ch2                 228.7

-Ch3                 289.9

-Ch4                 249.9

-Ch5                 241.2

-Ch6                 217.0

-Ch7                 149.8

-Ch8                 238.2

-Ch9                 167.3

-Ch10               167.2

Total               2250.9

 

3. Input files for Windows QTL IciMapping for generating simulated DH populations, and simulated populations for Windows QTL Cartographer, Empirical Bayesian Model, and Windows QTL IciMapping

3.1 Marker interval length = 2 cM , 224 RIL lines (Interval02PS224.qtl)

3.1.1 100 simulated RIL populations for Windows QTL Cartographer ( Interval02PS224 -CIM)

3.1.2 100 simulated RIL populations for Empirical Bayesian Model ( Interval02PS224 -BAYES)

3.1.3 100 simulated RIL populations for Windows QTL IciMapping ( Interval02PS224 -ICIM)

 

3.2 Marker interval length = 2 cM , 500 RIL lines (Interval02PS500.qtl)

3.2.1 100 simulated RIL populations for Windows QTL Cartographer ( Interval02PS500 -CIM)

3.2.2 100 simulated RIL populations for Empirical Bayesian Model ( Interval02PS500 -BAYES)

3.2.3 100 simulated RIL populations for Windows QTL IciMapping ( Interval02PS500 -ICIM)

 

3.3 Marker interval length = 10 cM , 224 RIL lines (Interval10PS224.qtl)

3.3.1 100 simulated RIL populations for Windows QTL Cartographer ( Interval10PS224 -CIM)

3.3.2 100 simulated RIL populations for Empirical Bayesian Model ( Interval10PS224 -BAYES)

3.3.3 100 simulated RIL populations for Windows QTL IciMapping ( Interval10PS224 -ICIM)

 

3.4 Marker interval length = 10 cM , 500 DH lines (Interval10PS500.qtl)

3.4.1 100 simulated RIL populations for Windows QTL Cartographer ( Interval10PS500 -CIM)

3.4.2 100 simulated RIL populations for Empirical Bayesian Model ( Interval10PS500 -BAYES)

3.4.3 100 simulated RIL populations for Windows QTL IciMapping ( Interval10PS500 -ICIM)