The Illinois Long-Term Selection Experiment, Related Studies, and Perspectives

Abstract

The Illinois Long-Term Selection Experiment was started in 1896 and represents the longest ongoing directional selection program in plants. The experiment was initiated to determine whether selection could alter levels of protein and oil in the maize kernel. The experiment readily demonstrated in the first several cycles that selection resulted in significant changes. Yet the experiment continues to this day, as gain from selection for high oil and for high protein is still progressing. The Illinois Long Term Selection Experiment (ILTSE) now addresses questions on the nature of gains from selection and limits to selection. This chapter focuses on insights developed through molecular marker studies on ILTSE. These studies involved the long term selection strains, mapping populations derived from the strains, and related mapping populations that underwent random mating. Studies on the oil and protein strains revealed trends in molecular markers consistent with response to selection, and that more heterozygosity remains in the strains than predicted earlier. The heterozygosity provides allelic variation for further selection, and also allows for intragenic recombination which may create new alleles. Studies involving populations derived from crosses of Illinois High Protein (IHP) × Illinois Low Protein (ILP), and Illinois High Oil (IHO) × Illinois Low Oil (ILO) detected QTL for kernel composition. Most of the markers associated with trends in response to selection were also associated with QTL. However, a relatively small number of QTL were detected in mapping populations, with six QTL explaining 65% of variation for protein, and six QTL explaining 52% of variation for oil, in the respective populations. These results were inconsistent with earlier predictions of 173 effective factors controlling protein concentration and 69 controlling oil. Subsequent studies involving random mating of related mapping populations, in order to break up large linkage blocks, revealed many more QTL, in a range more consistent with earlier predictions. For example, 50 QTL were associated with 50% of the variation for oil in a random mated population derived from IHO × ILO. Collectively, the large number of loci detected controlling oil and protein, and the much higher levels of heterozygosity maintained in the strains than predicted, help to explain some of the basis for long term gains in selection. Opportunities for use of the selection strains and derived genetic materials in the era of genomics are discussed.