Evaluation of short-day onion doubled haploid lines

Molecular marker analysis of seven putative onion (Allium cepa) doubled haploid (DH) lines developed at Texas A&M University was conducted to verify genetic homozygosity. Analysis was also conducted on five equivalent conventional inbred lines, breeding lines developed from the same parental crosses as the DH lines, and the original parent lines. The markers have revealed polymorphisms within the parental lines and the conventional inbreds, but not in the DH lines. We can conclude therefore that these seven lines are true DH lines. Performance of these DH lines was tested in two field locations and compared to commercial check lines. Bulbs from the various crosses were evaluated for eight bulb traits: diameter, height, centers/bulb, ring thickness, number of rings/bulb, bulb weight, soluble solids content, and pungency. Some crosses were detected that yielded significantly greater bulb weight than the check lines. However, these lines also had significantly greater numbers of centers per bulb. To test how these lines would perform in a breeding program, two full diallel analyses were conducted according to Griffing?s Model I, Method 1. The first consisted of a four parent diallel cross using two red DH lines and two yellow DH lines. Bulbs from the various crosses were evaluated for the same eight bulb traits mentioned above. Significant variation was detected for genotypic, general combining ability (GCA), specific combining ability (SCA), reciprocal (REC), maternal (MAT), and nonmaternal (NMAT) effects for all traits except number of rings/bulb, soluble solids content, and pungency. Significant environmental effects were only detected with number of centers per bulb. The second diallel analysis, a four parent diallel with two DH lines and two inbred lines from the breeding program, showed significant variation for the same effects for all traits except soluble solids content. Generally, GCA effects were more important than SCA effects in explaining the variation observed between crosses. For all traits GCA and SCA were always larger than the reciprocal effects (divided into maternal and nonmaternal components).