Changes of Gene Frequencies in Synthetic Corn Populations by Two Methods of Recurrent Selection and Pedigree Selection

Abstract

Forty-five F1C0 corn hybrids were derived from diallel series of 10 selected inbreds from different sources of germplasm. AgSyn1C0 was synthesized from bulked balanced seeds of 45 F1C0. The remnant seeds were tested for their yielding ability in randomized complete block design with 4 replications simultaneously, selected plants in border rows were selfed to obtain S1C0. Forty-five S1C0 from diallel set plus 4 additional S1C0 were planted in R-49 grouped replicated honeycomb design. Top-10 F1C0 (diallel selection) and top-10 S1C0 (S1 selection) were selected from each cycle. Their pedigrees were recorded and frequencies of original lines presented in pedigree of selected lines were used to calculate gene frequencies in derived populations. Pedigree selection was also conducted from each selected S1 set from each cycle. Two cycles of selection were done to obtain AgSyn1C1, AgSyn1C2. Two selected lines from pedigree selection were added to AgSyn1C2 to obtain AgSyn1C2-Sem. Changes of gene frequency patterns by the 3 methods of selection were very similar but diallel selection and pedigree selection retained more genetic diversity than S1 selection. Therefore, there was no clear advantage of recurrent selection over the pedigree selection. However, recurrent selection enhances gene recombination and accumulation of desirable genes in the population while pedigree selection is suitable for accumulation of few desirable genes in single plant. The results suggested that recurrent selection should be a supporting program to provide desirable germplasm to pedigree program. Diallel selection not only retained more genetic diversity but needs only 2 planting seasons to complete a cycle instead of 3 planting seasons as most recurrent selections now generally in use. Therefore, diallel selection should be more effective than S1 selection for the improvement of population per se from which good inbreds could be derived. The results also suggested that genetic diversity of first few cycles of selection drastically dropped when combining ability of the original lines were widely different.