Source of variation | d.f.^{a} | Expected mean square^{b} |
---|---|---|

Replications | r − 1 | |

Parental lines (P) | 1 | |

Z_{1} | n − 1 | |

Marker m | 1 | |

Remainder Z_{1}(m) | n − 2_{m} | |

Z_{2} | n − 1 | |

Marker m | 1 | |

Remainder Z_{2}(m) | n − 2_{m} | |

Error | (2n − 1)(r − 1) |

Sums of squares for pair means

*Z*_{1}and pair differences*Z*_{2}were subdivided into orthogonal contrasts allowing the comparison of genotypic classes at marker*m*.↵

*a**r*, number of replicates;*n*, number of RILs;*n*, number of RILs genotyped at marker_{m}*m*,*n*=_{m}*n*_{m}_{0}+*n*_{m}_{2}with*n*_{m}_{0}and*n*_{m}_{2}referring to the number of RILs with marker genotypes 0 and 2, respectively.↵

*b*and refer to the variance of residual genotypic effects in*Z*_{1}and*Z*_{2}not accounted for by the*m*th marker. error variance.