Optimum allocation of test resources and relative efficiency of alternative procedures of within‐family selection in hybrid breeding

Selection within families can be conducted as family deviation (FDS) or strict within‐family selection (WFS). Our objectives were to (i) investigate two breeding schemes combining selection among families with FDS or WFS and (ii) compare the optimum allocation of test resources for these breeding schemes. We focused on selection among S₁ families and doubled haploid (DH) lines within S₁ families and used Monte Carlo simulations to determine the selection gain (Δ?), its standard deviation (SD_Δ?), and the average coefficient of coancestry among the selected DH lines (). For breeding schemes focusing only on within‐family selection, as employed in animal breeding, the maximum Δ?, its SD_Δ?, and were larger for FDS than for WFS. However, for breeding schemes combining among‐ and within‐family selection, as employed commonly in plant breeding programmes, the maximum Δ?, SD_Δ?, and were almost equal for FDS and WFS. Furthermore, the optimum allocation of test resources was similar for FDS and WFS. We conclude that FDS and WFS are equally suited for short‐ and long‐term success in breeding schemes where among‐family selection is followed by within‐family selection.