Prediction of sample size to maintain genetic variation in doubled-haploidpopulations following marker selection

Marker selection (MS) and doubled-haploid (DH) technologies have the potent ial to reduce the time taken to breed new cereal cultivars. However, a limi ting factor is the potential increased genetic drift. The aim of this study was to design and test a genetic model for predicting the sample sizes nee ded to maintain genetic variation among DH plants following marker selectio n. The model estimates the amount of the genome that is fixed during the pr oduction of DH populations of a given size using a given number of markers. To test the model. doubled-haploids were produced from wheat plants select ed for three PCR-based markers. When the genetic variation of the DH popula tion (108 plants), produced from 15 selected F-2 plants homozygous at three loci, was compared to the genetic variation of an unselected F-3 populatio n (200 plants), five of the six measured quantitative traits were identical and normally distributed. This model should prove to be a valid breeding t ool, allowing a breeder to apply MS to a breeding programme and estimate th e minimum DH population sizes required for minimal loss of genetic variatio n through genetic drift.