Optimising the construction of a substitution library in Arabidopsis thaliana using computer simulations

The feasibility of constructing an overlapping part-chromosome substitution library between the Landsberg erecta (Ler) and Columbia (Col) ecotypes of Arabidopsis thaliana was investigated using computer simulations. Such a library can be used effectively in the fine mapping of quantitative trait loci (QTLs) in this species, with donor tracts of Col inserted into the recurrent background of Ler and vice versa. The study was based on the known RFLP profiles of Ler and Col ecotypes, and of some selected recombinant inbred lines (RILs) that have already been extracted from their cross by self pollination. It was shown that homozygous substitution lines can be produced for a large segment of the genome ( > 80%) by crossing just 11 RILs with their respective recurrent/recipient parents and selecting the desired recombinants from the F-2 or Bc(1) generation or their selfed progenies. In the case of four RILs, however, at least two rounds of backcrossing were deemed necessary to remove the unwanted donor tracts from the background genotype prior to selfing/selection. The simulations also provided valuable information on the scale of the breeding programme and show that a resource of up to 59 substitutions can be produced within a short period of 4-5 generations.