RT Dissertation/Thesis
T1 Molecular genetic analysis of modified recurrent full-sib selection in two European F 2 flint maize populations
A1 Falke,Karen Christin
WP 2008/02/19
AB The continuous improvement of breeding material is a main goal in plant breeding. However, many breeding methods result in a reduction of the genetic variation of the breeding material. The primary objective of recurrent selection (RS) methods is to assure a long-term selection response for the target traits while maintaining the genetic variability of the germplasm for continued selection. The main goal of this thesis research was the molecular evaluation of two European F2 flint maize populations under modified recurrent full-sib selection. In detail, the objectives were to (1) investigate linkage mapping with intermating populations, (2) verify the decay of parental linkage disequilibrium (LD) present in the F2 base population through three generations of intermating, (3) identify quantitative trait loci (QTL) of traits under selection, (4) separate the effects of random genetic drift and selection on allele frequency changes during the selection procedure in QTL regions, (5) determine the extent of LD build-up by selection, and (6) analyze the effects of LD on the selection response.  Four homozygous inbred lines (A, B, C, and D) were crossed to developed 380 F2:3 lines of population A × B and 140 F3:4 lines of population C × D. The F2 generations of both populations were intermated for three generations by chain crossing to produce F2Syn3 populations. Starting from the F2Syn3 populations, four (A × B) and seven (C × D) cycles of modified recurrent full-sib selection were conducted using a pseudo-factorial mating design. The selection procedure was based on a selection index. For the marker assays, a total of 104 (A × B) and 101 (C × D) SSRs were employed to genotype random subsets of 146 F2:3 lines in A × B and 110 F3:4 lines C × D, 148 F2Syn3 plants and the parents of the 36 selected full-sib families in each cycle of both RS programs. Genetic linkage maps were constructed for population F2:3 (A × B) and F3:4 (C × D) and of both intermated populations F2Syn3. In contrast to earlier studies, mapping methods developed specifically for intermated populations were applied and, thus, the estimated recombination frequencies referred to a single meiosis. Consequently, the map expansion reported in earlier studies was neither expected nor observed. To verify the expected higher mapping resolution of intermated compared with the respective base population, the precision of estimates of the recombination frequencies r was quantified with the average information per individual ir relative to an F2 individual. For infinite population sizes of intermated populations ir > 1 when r < 0.142, however, with a population size of N = 240, as used in our mapping study, ir >1 only when r < 0.04. Therefore, we conclude that random genetic drift has a sizeable effect on ir and, thus, can overrule the advantages of intermating mapping populations. Three generations of intermating were primarily conducted to reduce the initial parental LD between linked loci in the F2 populations and its negative influence on the selection response. Our results demonstrated that the observed decay of LD agreed well with the theoretical expectations. In our modified recurrent full-sib selection program, a comparatively high selection response was reached. An evaluation at the molecular level revealed that further selection response can be expected, because neither fixation nor extinction of alleles was observed at the investigated marker loci. However, using SSR markers we also detected migration effects since the first selection cycles.  Selection and random genetic drift are main forces affecting changes in allele frequencies and, therefore, the selection response of RS programs. In our study, we analyzed changes in allele frequencies employing a test that allows the separation of selection and random genetic drift. Significant allele frequency changes due to selection were observed for 23% (C0 vs. C1) and 20% (C0 vs. C4) of all loci in population A × B and for 6% (C0 vs. C1) and 13% (C0 vs. C7) in population C × D. In the base populations F2:3 (A × B) and F3:4 (C × D), several QTL for selection index and its components were detected. At some of them, loci displayed significant allele frequency changes due to selection. Selection is expected to increase the frequency of favorable alleles and simultaneously build up a negative LD between them, which causes a reduction in   and, hence, a decline in the selection response. However, the development of LD in our study displayed an erratic change over the selection cycles with only slight increases in positive and negative LD. The reduction in   due to a build-up of negative LD is expected only if negative LD is observed at many marker loci. Therefore, we concluded that LD due to selection did not limit the selection response in our RS programs. 
K1 Mais
K1 Rekurrente Selektion
K1 Genkartierung
K1 Genetische Variabilität
K1 Markierungsgen
PP Hohenheim
PB Kommunikations-, Informations- und Medienzentrum der Universität Hohenheim
UL http://opus.uni-hohenheim.de/volltexte/2008/209