Naslov
Selecting Parents for White Lupin Breeding Programme

Autori
Gunjača, Jerko ; Šatović, Zlatko ; Pecina, Marija ; Kolak, Ivan

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Izvornik
Proceedings of XXIInd International Biometric Conference in parallel with Australian Statistical Conference 2004 / Basford, Kaye ; Bartlett, Neville - Cairns : IBC/ASC 2004, 2004, 483-483

Skup
XXIInd International Biometric Conference (IBC) in parallel with Australian Statistical Conference (ASC) 2004

Mjesto i datum
Cairns, Australija, 11.07.2004. - 16.07.2004

Vrsta sudjelovanja
Poster

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
cluster analysis; PCA; plant breeding; selection; white lupine

Sažetak
White lupin (Lupinus albus L.) is herbaceous annual plant of the legume family (Fabaceae) which has been grown for over 3000 years. As a human food source white lupin has been largely neglected. The major interest in white lupin arises from its potential to contribute to high profitability of farming on very poor sandy and acidic soils in marginal agricultural areas. Its production in Croatia is based entirely on landraces adapted to local conditions. Therefore, superior genotypes can be used as parents in a breeding programme. In order to detect the most prospective parent combinations, 18 traits of 17 accessions in a 5 replicate randomized complete block design in two consecutive years were scored. Traits can be divided into four groups: whole plant, stem, root and pod/grain traits. In order to classify accessions cluster analysis was used. Straightforward clustering based on these traits would be strongly affected by their high multicollinearity. One of the possible solutions to this problem is the application of principal component analysis prior to clustering. To choose the appropriate breeding strategy it was necessary to identify which original traits the final clusters differ in. Therefore, the analysis of variance for all of the traits was carried out and cluster separation was investigated. According to the results, accessions can be divided into three distinctive groups. Analysis of variance for original traits detected groups which differ significantly regarding a certain trait, and crossing the accessions from different groups can improve that trait.

Izvorni jezik
Engleski

Znanstvena područja
Poljoprivreda (agronomija)

POVEZANOST RADA